Yuechen Jiang, Zhiwei Liu, Yupeng Cao, Yueru He, Ziyang Xu, Chen Xu, Zhiyang Deng, Prayag Tiwari, Xi Chen, Alejandro Lopez-Lira, Jimin Huang, Junichi Tsujii, Sophia Ananiadou. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Xueqing Peng, Lingfei Qian, Yan Wang, Ruoyu Xiang, Yueru He, Yang Ren, Mingyang Jiang, Vincent Jim Zhang, Yuqing Guo, Jeff Zhao, Huan He, Yi Han, Yun Feng, Yuechen Jiang, Yupeng Cao, Haohang Li, Yangyang Yu, Xiaoyu Wang, Penglei Gao, Shengyuan Lin, Keyi Wang, Shanshan Yang, Yilun Zhao, Zhiwei Liu, Peng Lu, Jerry Huang, Suyuchen Wang, Triantafillos Papadopoulos, Polydoros Giannouris, Efstathia Soufleri, Nuo Chen, Zhiyang Deng, Heming Fu, Yijia Zhao, Mingquan Lin, Meikang Qiu, Kaleb E Smith, Arman Cohan, Xiao-Yang Liu, Jimin Huang, Guojun Xiong, Alejandro Lopez-Lira, Xi Chen, Junichi Tsujii, Jian-Yun Nie, Sophia Ananiadou, Qianqian Xie. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

We propose ELAINE (EngLish-jApanese-chINesE)-medLLM, a trilingual (English, Japanese, Chinese) large language model adapted for the bio-medical domain based on Llama-3-8B. The training dataset was carefully curated in terms of volume and diversity to adapt to the biomedical domain and endow trilingual capability while preserving the knowledge and abilities of the base model. The training follows 2-stage paths: continued pre-training and supervised fine-tuning (SFT). Our results demonstrate that ELAINE-medLLM exhibits superior trilingual capabilities compared to existing bilingual or multilingual medical LLMs without severely sacrificing the base model’s capability.

As an engineering field, research on natural language processing (NLP) is much more constrained by currently available resources and technologies, compared with theoretical work on computational linguistics (CL). In today’s technology-driven society, it is almost impossible to imagine the degree to which computational resources, the capacity of secondary and main storage, and software technologies were restricted when I embarked upon my research career 50 years ago. While these restrictions inevitably shaped my early research into NLP, my subsequent work evolved, according to the significant progress made in associated technologies and related academic fields, particularly CL.Figure 1 shows the research topics in which I have been engaged. My initial NLP research was concerned with a question answering system, which I worked on during my M.Eng and D.Eng degrees. The research focused on reasoning and language understanding, which I soon found was too ambitious and ill-defined. After receiving my D.Eng., I changed my direction of research, and began to be engaged in processing forms of language expressions, with less commitment to language understanding, machine translation (MT), and parsing. However, I returned to research into reasoning and language understanding in the later stage of my career, with clearer definitions of tasks and relevant knowledge, and equipped with access to more advanced supporting technologies.In this article, I begin by briefly describing my views on mutual relationships among disciplines related to CL and NLP, and then move on to discussing my own research.Language is a complex topic to study, infinitely harder than I first imagined when I began to work in the field of NLP.There is a whole discipline on the study of language—namely, linguistics. Linguistics is concerned not only with language per se, but must also deal with how humans model the world.1 The study of semantics, for example, must relate language expressions to their meanings, which reside in the mental models possessed by humans.Apart from linguistics, there are two fields of science that are concerned with language, that is, brain science and psychology. These are concerned with how humans process language. Then, there are two disciplines in which we are involved—namely, CL and NLP.Figure 2 is a schematic view of these research disciplines. Both of the lower disciplines are concerned with processing language, that is, how language is processed in our minds or our brains, and how computer systems should be designed to process language efficiently and effectively.The top discipline, linguistics, on the other hand, is concerned with rules that are followed by languages. That is to say, linguists study language as a system. This schematic view is certainly oversimplified, and there are subject fields in which these disciplines overlap. Psycholinguistics, for example, is a subfield of linguistics which is concerned with how the human mind processes language. A broader definition of CL may include NLP as its subfield.In this article, for the sake of discussion, I adopt narrower definitions of linguistics and CL. In this narrower definition, linguistics is concerned with the rules followed by languages as a system, whereas CL, as a subfield of linguistics, is concerned with the formal or computational description of rules that languages follow.2CL, which focuses on formal/computational description of languages as a system, is expected to bridge broader fields of linguistics with the lower disciplines, which are concerned with processing of language.Given my involvement in NLP, I would like to address the question of whether the narrowly defined CL is relevant to NLP. The simple answer is yes. However, the answer is not so straightforward, and requires us to examine the degree to which the representations used to describe language as a system are relevant to the representations used for processing language.Although my colleagues and I have been engaged in diverse research areas, I pick up only on a subset of these, to illustrate how I view the relationships between NLP and CL. Due to the nature of the article, I ignore technical details and focus instead on the motivation of the research and the lessons which I have learned through research.Background and Motivation. Following the ALPAC report Pierce et al. (1966), research into MT had been largely abandoned by academia, with the exception of a small number of institutes (notably, GETA at Grenoble, France, and Kyoto University, Japan). There were only a handful of commercial MT systems, being used for limited purposes. These commercial systems were legacy systems that had been developed over years and had become complicated collections of ad hoc programs. They had become too convoluted to allow for changes and improvements. To re-initiate MT research in academia, we had to have more systematic and disciplined design methodologies.On the other hand, theoretical linguistics, initiated by Noam Chomsky (Chomsky 1957, 1965) had attracted linguists with a mathematical orientation, who were interested in formal frameworks of describing rules followed by language. Those linguists with interests in formal ways of describing rules were the first generation of computational linguists.Although computational linguists did not necessarily follow the Chomskyan way of thinking, they shared the general view of treating language as a system of rules. They had developed formal ways of describing rules of language and showed that these rules consisted of different layers, such as morphology, syntax, and semantics, and that each layer required different formal frameworks with different computational powers. Their work had also motivated work on how one could process language by computerizing its rules of language. This work constituted the beginning of NLP research, and resulted in the development of parsing algorithms for context-free language, finite-state machines, and so forth.3 It was natural to use this work as the basis for designing the second generation of MT systems, which was initiated by an MT project (MU project, 1082-1986) led by Prof. M. Nagao (Nagao, Tsujii, and Nakamura 1985).Research Contributions. When I began research into MT in the late 1970s, there was a common view largely shared by the community, which had been advocated by the group of GETA, in France. The view was called the transfer approach of MT (Boitet 1987).The transfer approach viewed translation as a process consisting of three phases: analysis, transfer, and generation. According to linguists, a language is a system of rules. The analysis and generation phases were monolingual phases that were concerned with a set of rules for a single language, the analysis phase using the rules of the source language and the generation phase using the rules of the target language. Only the transfer phase was a bilingual phase.Another view shared by the community was an abstraction hierarchy of representation, called the triangle of translation. For example, Figure 3(a)4 shows the hierarchy of representation used in the Eurotra project, with their definition of each level (Figure 3(b)).By climbing up such a hierarchy, the differences among languages would become increasingly small, so that the mapping (i.e., the transfer phase) from one language to another would become as simple as possible. Independently of the target language, the goal of the analysis phase was to climb up the hierarchy, while the aim of the generation phase was to climb down the hierarchy to generate surface expressions in the target language. Both phases are concerned only with rules of single languages.In the extreme view, the top of the hierarchy was taken as the language-independent representation of meaning. Proponents of the interlingual approach claimed that, if the analysis phase reached this level, then no transfer phase would be required. Rather, translation would consist only of the two monolingual phases (i.e., the analysis and generation phases).However, in Tsujii (1986), I claimed, and still maintain, that this was a mistaken view about the nature of translation. In particular, this view assumed that a translation pair (consisting of the source and target sentences) encodes the same “information”. This assumption does not hold, in particular, for a language pair such as Japanese and English, that belong to very different language families. Although a good translation should preserve the information conveyed by the source sentence as much as possible in the target sentence, translation may lose some information or add extra information.5Furthermore, the goal of translation may not be to preserve information but to convey the same pragmatic effects to readers of the translation.More seriously, the abstract level of representation such as Interface Structure6 in Eurotra focused only on the propositional content encoded in language, and tended to abstract away other aspects of information, such as the speaker’s empathy, distinction of old/new information, emphasis, and so on.To climb up the hierarchy led to loss of information in lower levels of representation. In Tsujii (1986), instead of mapping at the abstract level, I proposed “transfer based on a bundle of features of all the levels”, in which the transfer would refer to all levels of representation in the source language to produce a corresponding representation in the target language (Figure 4). Because different levels of representation require different geometrical structures (i.e., different the of this had to for development of a mathematical of representation with which levels (i.e., to be with their mutual relationships the we to the transfer phase was transfer and Tsujii which was by the of in CL. According to the views of linguists at the a language is an set of expressions in is defined by a set of rules. this number of one generate infinitely of the language. claimed that the of a was by the of its using the rules that the translation the same to translation. That is, the translation of a was by the of its In this of infinitely of the source language could be the translation the translation of a sentence would be by a of a source The translation of a would then be by the of its That is, translation would be in a up from of translation to the mapping of a from the source to the target would be by the of the the for the how to a to the In the project, we called this transfer and Tsujii (Figure with the MT systems, which source expressions with target in an and ad hoc the of transfer in the project was defined and and development of the MT systems from research into CL, more defined and design than MT The project and MT systems the of these design we could not have these in such a of the differences between the of the two disciplines also CL to focus on aspects of language as morphology, syntax, semantics, MT systems must be to all aspects of information conveyed by language. As climbing up a hierarchy that focuses on propositional content does not in good more between CL and NLP is the of of is the single significant in NLP it requires the in which expressions to be to be In other it requires understanding of of are in Figure The Japanese a of in and would be into have and so on the for with transfer, it requires to be (i.e., the in which a to be The nature of made the process of transfer was also a in the analysis which I in the general of CL or linguistics is that it to view language as an system and the of understanding, which requires to or However, NLP require an understanding or of language expressions in of and which may other such as and so I this in the on the of research.Background and Motivation. the I was engaged in MT research, in CL, its early in theoretical linguistics by Chomsky assumed that of of rules the two levels of that is, and surface A way of was also shared by the MT They assumed that climbing up the hierarchy would of which from the representation at one level to another representation at the Because each level of the hierarchy required its own geometrical it was not possible to have a representation, in which representations of all the levels view was changed by the of that used to allow of from one level to it mutual relationships among different levels of representation in a This view was in with our of transfer, which used a bundle of features of different levels for some at the the of That is, structures of all the levels are constrained by the of a and these are encoded in This was also in with our significant development in CL at the same a number of the and the had linguistics and to have significant on research into CL and NLP et al. the NLP of view, the of led to the development of (i.e., for research that would these two to the analysis is, parsing based on Contributions. It is claimed that of In the analysis phase of the up the lower levels of processing could not refer to in levels of representation. This was the main of the of at the early of climbing up the analysis could not refer to that in analysis would the other hand, the could describe at all levels in a single it was possible to refer to at all to down the set of possible in the was still This was we not have ways of and pragmatic linguists were interested in formal ways for and levels of representation, but not so much in how are to be To or pragmatic one may have to refer to the mental models of the (i.e., how humans the or structures single and so These of the of CL research at the main focus is on it is whether or be used as They may be more concerned with the of an than the which an should example, the in for parsing using of and how to the of for models were one of the for and the of an However, models for such as and context-free had to be changed for more complex for such as had to be for were in of describing in a the which was a for treating was very To NLP systems, we had to technologies and processing for at the of to study how we could a into and representations for parsing. The research of an abstract machine for processing of structures and development of a et al. et al. into a more representation, such as and Tsujii et al. and from of structures and models and Tsujii of parsing based on of and their and Tsujii et al. representation of our parsing model is in Figure a sentence, its representation of all the levels was at the stage by using the in the first two The first phase was a that would to in a were from the and a set of were to a in its A would the of for the of The was a which could be in a very and Tsujii This that the surface (i.e., of was used for of second phase was A which also was from the was used to whether of by the first phase could a The did not to whether a could or The second phase of would that could not phase not only the representation of all the but it also extra in the Because the first two phases only use in the the phase would by the first two phases if they to these extra In this the system would to the phases to the of these research a based on As in CL were for the systematic development of NLP changed the view of parsing as up the the systematic of of of into and so These also for in NLP such as of and so which are for treating the other hand, of CL to NLP did not this would in had to into more which required significant and on the NLP For example, we had to the into such as and so It is that, while the was to the hierarchy each stage in the was defined and related to each other through the single also that in the fields of computer changed was possible to in NLP. For example, the design of an abstract machine and its for in et al. systems for of and Tsujii Tsujii, and and so would be impossible in the broader fields of computer and much computational et al. the other hand, the in NLP. models in of the with the that we for the the NLP community with systematic ways of with of different models also which made systematic development of NLP systems possible. However, the in parsing still are the in of the number of was the sentence level, the That is, a sentence in which all are very Because of are (i.e., of or of in such as of and Tsujii using there are to a understanding is taken into This to the research language and and Motivation. I was interested in the topic of how to relate language with at the very beginning of my the my led to that a of could be used as a and was engaged in research of a system based on a and Tsujii However, resources such as a of processing of computer systems, and NLP technologies, such as were not available at the soon that the research would a whole of research topics in such as representation of common human ways of and so the topics had to deal with and of or the models that humans have may from one to I that the research target was through research in MT and parsing in the later of my career, I to that NLP research is if it how is in and that NLP are all related to of understanding and the same NLP as an engineering field, I it to be to have a definition of or information with which language is to be I would like to too much of research into and reasoning and to our research focus to the between language and As a research I to focus on the as the There were two for the was that colleagues at the two with which I was that, in to it had become increasingly for to of information in a of in diverse subject fields such as and In to the of they also had diverse that had to be with each In other they had a of shared by that was to be with information in other was that there were colleagues at the of who were interested in According to the on information in a by the group and research into at the of on between and and et al. the had been a natural of The was that information in a and were defined by the target and not by NLP had and to their own in the target Contributions. Although there had been a of research into information and for the there had been no to NLP to in the To address the at the of and the of a research in this this was a research we first had to tasks to to resources, and to not only NLP but also in the target the involvement of NLP and we found that a in the also began to be interested in research In to we a number of research in with colleagues the which led to of a at The first in with the The and shared It been and become one of the in NLP The research field of of NLP to is to et al. et al. and science et al. by the two include the et al. and Tsujii a of with their and Tsujii the et al. the et al. a design for information et al. an system based on Tsujii, and a system for et al. is one of the used in the To information in and how it was encoded in language, we not only from the of view but also from the of of and and and were to the et al. were with of the target which had been by the target to information in diverse in we developed a with which is used by the NLP In with we defined a set of NLP tasks et al. and Tsujii and Tsujii et al. and developed a set of et al. et al. et al. for which were to be into to of of et al. et al. a of this we between such as and and such as and and that (Figure The mapping between structures and the defined by was more complex than the mapping assumed by computational soon that, as an NLP information was very different from In particular, there were differences between the information that the to convey and in and the information that the readers (i.e., a group of to and from of the information that the to the would the information that they were interested of as an NLP made the mapping from language to information very different from the transfer phase in which to the same information in the source and target languages. While the of from structures (i.e., and their structures would only for the of and and these were to be with other The mapping to the transfer based on a bundle of the transfer in the level of representation, we first used the to climb up the hierarchy to the we called from which we to a set of to (Figure assumed that, based on surface of may be this would at a level of is, the same approach to simple transfer at the abstract Although this approach it soon reached its increasingly and we that this was of the nature of and to the approach based on a bundle of features (Figure et al. This to the research, which a language model In this information is assumed to be in the language The information is not in systems (Figure and et al. the other hand, our in the tasks and of In this it that structures more significant example, about an from different each As for the or of are fields such as and about an be made or with different of To the degree of of a we have to examine the of in which the is et al. and on the of an in a sentence, it may be as a and so The in which information by a be to and an research of an is the of in which an NLP system is used to a set of different from different to a of et al. In this a system must be to whether two in different be as the same in a To the system must be to the in which two by the of the This may also require structures to be taken into on the we forms of (i.e., by the target and diverse which include The of information in with these resources to research topics on the between language and of the target Because such as have views on which of information their knowledge, we the and of that may have research in the general with which NLP technologies such as parsing have been less than we research into the of information to the of is the nature of human The CL and NLP to reasoning as a of process based on However, the reasoning that the in the may not be so in example, we found that the reasoning by on is based on between For example, they that a A is to be in an by a that a which is to is in the same of The between A and is based on the between their Because such among are in their in language models on a of would be to their (i.e., a of also to such it may be to use of information, such as of collections of and so to such among and to reasoning based on have the progress and significant changes that models and have to the field of NLP. This is a in which in broader fields of computer up to and the NLP changes by and are and have had a not only on NLP but also and other of It is a the of diverse NLP I it the NLP by NLP with the processing of other information and with and so fields such as and reasoning based on used different and processing They the same basis of and It is much to forms of that NLP in and NLP with are more than we The research of the with which I are on these et al. et al. and the NLP based on language models is increasingly from other research disciplines that the study of language. The nature of and also the way of NLP systems the are very I that the may to by MT could the by computational but and systematic analysis of this to and the of the through analysis of parsing when the of analysis based on by other disciplines, and of NLP systems and the other hand, CL to language as a system or focus on study on aspects of that language in NLP systems, with NLP CL would be to the of their and to a theoretical basis for of NLP is the to NLP and was to the field of NLP by my who was a of the away It is that I could not my and with my career of almost 50 I have research into NLP at institutes Kyoto of the of and I the on of the research and who worked with at these my research could not have in the way that it I their

Sentence pair modelling is defined as the task of identifying the semantic interaction between a sentence pair, i.e., paraphrase and textual entailment identification and semantic similarity measurement. It constitutes a set of crucial tasks for research in the area of natural language understanding. Sentence representation learning is a fundamental technology for sentence pair modelling, where the development of the BERT model realised a breakthrough. We have recently proposed transfer fine-tuning using phrasal paraphrases to allow BERT’s representations to be suitable for semantic equivalence assessment between sentences while maintaining the model size. Herein, we reveal that transfer fine-tuning with simplified feature generation allows us to generate representations that are widely effective across different types of sentence pair modelling tasks. Detailed analysis confirms that our transfer fine-tuning helps the BERT model converge more quickly with a smaller corpus for fine-tuning.

Due to the compelling improvements brought by BERT, many recent representation models adopted the Transformer architecture as their main building block, consequently inheriting the wordpiece tokenization system despite it not being intrinsically linked to the notion of Transformers.While this system is thought to achieve a good balance between the flexibility of characters and the efficiency of full words, using predefined wordpiece vocabularies from the general domain is not always suitable, especially when building models for specialized domains (e.g., the medical domain).Moreover, adopting a wordpiece tokenization shifts the focus from the word level to the subword level, making the models conceptually more complex and arguably less convenient in practice.For these reasons, we propose CharacterBERT, a new variant of BERT that drops the wordpiece system altogether and uses a Character-CNN module instead to represent entire words by consulting their characters.We show that this new model improves the performance of BERT on a variety of medical domain tasks while at the same time producing robust, word-level, and open-vocabulary representations.

Phrase alignment is the basis for modelling sentence pair interactions, such as paraphrase and textual entailment recognition.Most phrase alignments are compositional processes such that an alignment of a phrase pair is constructed based on the alignments of their child phrases.Nonetheless, studies have revealed that non-compositional alignments involving long-distance phrase reordering are prevalent in practice.We address the phrase alignment problem by combining an unordered tree mapping algorithm and phrase representation modelling that explicitly embeds the similarity distribution in the sentences onto powerful contextualized representations.Experimental results demonstrate that our method effectively handles compositional and non-compositional global phrase alignments.Our method significantly outperforms that used in a previous study and achieves a performance competitive with that of experienced human annotators.

Yuki Arase, Jun’ichi Tsujii. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

*: Background Consisting of dictated free-text documents such as discharge summaries, medical narratives are widely used in medical natural language processing. Relationships between anatomical entities and human body parts are crucial for building medical text mining applications. To achieve this, we establish a mapping system consisting of a Wikipedia-based scoring algorithm and a named entity normalization method (NEN). The mapping system makes full use of information available on Wikipedia, which is a comprehensive Internet medical knowledge base. We also built a new ontology, Tree of Human Body Parts (THBP), from core anatomical parts by referring to anatomical experts and Unified Medical Language Systems (UMLS) to make the mapping system efficacious for clinical treatments. *: Result The gold standard is derived from 50 discharge summaries from our previous work, in which 2,224 anatomical entities are included. The F1-measure of the baseline system is 70.20%, while our algorithm based on Wikipedia achieves 86.67% with the assistance of NEN. *: Conclusions We construct a framework to map anatomical entities to THBP ontology using normalization and a scoring algorithm based on Wikipedia. The proposed framework is proven to be much more effective and efficient than the main baseline system.

Motoki Sato, Austin J. Brockmeier, Georgios Kontonatsios, Tingting Mu, John Y. Goulermas, Jun’ichi Tsujii, Sophia Ananiadou. Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers. 2017.

We propose an efficient method to conduct phrase alignment on parse forests for paraphrase detection.Unlike previous studies, our method identifies syntactic paraphrases under linguistically motivated grammar.In addition, it allows phrases to non-compositionally align to handle paraphrases with non-homographic phrase correspondences.A dataset that provides gold parse trees and their phrase alignments is created.The experimental results confirm that the proposed method conducts highly accurate phrase alignment compared to human performance.

This work is an attempt to automatically obtain numerical attributes of physical objects. We propose representing each physical object as a feature vector and representing sizes as linear functions of feature vectors. We train the function in the framework of the combined regression and ranking with many types of fragmentary clues including absolute clues (e.g., A is 30cm long) and relative clues (e.g., A is larger than B).

BACKGROUND: Since their introduction in 2009, the BioNLP Shared Task events have been instrumental in advancing the development of methods and resources for the automatic extraction of information from the biomedical literature. In this paper, we present the Cancer Genetics (CG) and Pathway Curation (PC) tasks, two event extraction tasks introduced in the BioNLP Shared Task 2013. The CG task focuses on cancer, emphasizing the extraction of physiological and pathological processes at various levels of biological organization, and the PC task targets reactions relevant to the development of biomolecular pathway models, defining its extraction targets on the basis of established pathway representations and ontologies. RESULTS: Six groups participated in the CG task and two groups in the PC task, together applying a wide range of extraction approaches including both established state-of-the-art systems and newly introduced extraction methods. The best-performing systems achieved F-scores of 55% on the CG task and 53% on the PC task, demonstrating a level of performance comparable to the best results achieved in similar previously proposed tasks. CONCLUSIONS: The results indicate that existing event extraction technology can generalize to meet the novel challenges represented by the CG and PC task settings, suggesting that extraction methods are capable of supporting the construction of knowledge bases on the molecular mechanisms of cancer and the curation of biomolecular pathway models. The CG and PC tasks continue as open challenges for all interested parties, with data, tools and resources available from the shared task homepage.

BACKGROUND: Electronic medical record (EMR) systems have become widely used throughout the world to improve the quality of healthcare and the efficiency of hospital services. A bilingual medical lexicon of Chinese and English is needed to meet the demand for the multi-lingual and multi-national treatment. We make efforts to extract a bilingual lexicon from English and Chinese discharge summaries with a small seed lexicon. The lexical terms can be classified into two categories: single-word terms (SWTs) and multi-word terms (MWTs). For SWTs, we use a label propagation (LP; context-based) method to extract candidates of translation pairs. For MWTs, which are pervasive in the medical domain, we propose a term alignment method, which firstly obtains translation candidates for each component word of a Chinese MWT, and then generates their combinations, from which the system selects a set of plausible translation candidates. RESULTS: We compare our LP method with a baseline method based on simple context-similarity. The LP based method outperforms the baseline with the accuracies: 4.44% Acc1, 24.44% Acc10, and 62.22% Acc100, where AccN means the top N accuracy. The accuracy of the LP method drops to 5.41% Acc10 and 8.11% Acc20 for MWTs. Our experiments show that the method based on term alignment improves the performance for MWTs to 16.22% Acc10 and 27.03% Acc20. CONCLUSIONS: We constructed a framework for building an English-Chinese term dictionary from discharge summaries in the two languages. Our experiments have shown that the LP-based method augmented with the term alignment method will contribute to reduction of manual work required to compile a bilingual sydictionary of clinical terms.

Georgios Kontonatsios, Ioannis Korkontzelos, Jun’ichi Tsujii, Sophia Ananiadou. Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, volume 2: Short Papers. 2014.

Automatically compiling bilingual dictio-naries of technical terms from comparable corpora is a challenging problem, yet with many potential applications. In this paper, we exploit two independent observations about term translations: (a) terms are of-ten formed by corresponding sub-lexical units across languages and (b) a term and its translation tend to appear in similar lex-ical context. Based on the first observa-tion, we develop a new character n-gram compositional method, a logistic regres-sion classifier, for learning a string similar-ity measure of term translations. Accord-ing to the second observation, we use an existing context-based approach. For eval-uation, we investigate the performance of compositional and context-based methods on: (a) similar and unrelated languages, (b) corpora of different degree of compa-rability and (c) the translation of frequent and rare terms. Finally, we combine the two translation clues, namely string and contextual similarity, in a linear model and we show substantial improvements over the two translation signals. 1

Recent research has shown the effectiveness of rich feature representation for tasks in natural language processing (NLP). However, exceedingly large number of features do not always improve classification performance. They may contain redundant information, lead to noisy feature presentations, and also render the learning algorithms intractable. In this paper, we propose a supervised embedding framework that modifies the relative positions between instances to increase the compatibility between the input features and the output labels and meanwhile preserves the local distribution of the original data in the embedded space. The proposed framework attempts to support flexible balance between the preservation of intrinsic geometry and the enhancement of class separability for both interclass and intraclass instances. It takes into account characteristics of linguistic features by using an inner product‐based optimization template. (Dis)similarity features, also known as empirical kernel mapping, is employed to enable computationally tractable processing of extremely high‐dimensional input, and also to handle nonlinearities in embedding generation when necessary. Evaluated on two NLP tasks with six data sets, the proposed framework provides better classification performance than the support vector machine without using any dimensionality reduction technique. It also generates embeddings with better class discriminability as compared to many existing embedding algorithms.

OBJECTIVE: To create an end-to-end system to identify temporal relation in discharge summaries for the 2012 i2b2 challenge. The challenge includes event extraction, timex extraction, and temporal relation identification. DESIGN: An end-to-end temporal relation system was developed. It includes three subsystems: an event extraction system (conditional random fields (CRF) name entity extraction and their corresponding attribute classifiers), a temporal extraction system (CRF name entity extraction, their corresponding attribute classifiers, and context-free grammar based normalization system), and a temporal relation system (10 multi-support vector machine (SVM) classifiers and a Markov logic networks inference system) using labeled sequential pattern mining, syntactic structures based on parse trees, and results from a coordination classifier. Micro-averaged precision (P), recall (R), averaged P&R (P&R), and F measure (F) were used to evaluate results. RESULTS: For event extraction, the system achieved 0.9415 (P), 0.8930 (R), 0.9166 (P&R), and 0.9166 (F). The accuracies of their type, polarity, and modality were 0.8574, 0.8585, and 0.8560, respectively. For timex extraction, the system achieved 0.8818, 0.9489, 0.9141, and 0.9141, respectively. The accuracies of their type, value, and modifier were 0.8929, 0.7170, and 0.8907, respectively. For temporal relation, the system achieved 0.6589, 0.7129, 0.6767, and 0.6849, respectively. For end-to-end temporal relation, it achieved 0.5904, 0.5944, 0.5921, and 0.5924, respectively. With the F measure used for evaluation, we were ranked first out of 14 competing teams (event extraction), first out of 14 teams (timex extraction), third out of 12 teams (temporal relation), and second out of seven teams (end-to-end temporal relation). CONCLUSIONS: The system achieved encouraging results, demonstrating the feasibility of the tasks defined by the i2b2 organizers. The experiment result demonstrates that both global and local information is useful in the 2012 challenge.

The present article describes a robust approach for abbreviating terms. First, in order to incorporate non-local information into abbreviation generation tasks, we present both implicit and explicit solutions: the latent variable model and the label encoding with global information. Although the two approaches compete with one another, we find they are also highly complementary. We propose a combination of the two approaches, and we will show the proposed method outperforms all of the existing methods on abbreviation generation datasets. In order to reduce computational complexity of learning non-local information, we further present an online training method, which can arrive the objective optimum with accelerated training speed. We used a Chinese newswire dataset and a English biomedical dataset for experiments. Experiments revealed that the proposed abbreviation generator with non-local information achieved the best results for both the Chinese and English languages.

MOTIVATION: Event extraction using expressive structured representations has been a significant focus of recent efforts in biomedical information extraction. However, event extraction resources and methods have so far focused almost exclusively on molecular-level entities and processes, limiting their applicability. RESULTS: We extend the event extraction approach to biomedical information extraction to encompass all levels of biological organization from the molecular to the whole organism. We present the ontological foundations, target types and guidelines for entity and event annotation and introduce the new multi-level event extraction (MLEE) corpus, manually annotated using a structured representation for event extraction. We further adapt and evaluate named entity and event extraction methods for the new task, demonstrating that both can be achieved with performance broadly comparable with that for established molecular entity and event extraction tasks. AVAILABILITY: The resources and methods introduced in this study are available from http://nactem.ac.uk/MLEE/. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: The Genia task, when it was introduced in 2009, was the first community-wide effort to address a fine-grained, structural information extraction from biomedical literature. Arranged for the second time as one of the main tasks of BioNLP Shared Task 2011, it aimed to measure the progress of the community since 2009, and to evaluate generalization of the technology to full text papers. The Protein Coreference task was arranged as one of the supporting tasks, motivated from one of the lessons of the 2009 task that the abundance of coreference structures in natural language text hinders further improvement with the Genia task. RESULTS: The Genia task received final submissions from 15 teams. The results show that the community has made a significant progress, marking 74% of the best F-score in extracting bio-molecular events of simple structure, e.g., gene expressions, and 45% ~ 48% in extracting those of complex structure, e.g., regulations. The Protein Coreference task received 6 final submissions. The results show that the coreference resolution performance in biomedical domain is lagging behind that in newswire domain, cf. 50% vs. 66% in MUC score. Particularly, in terms of protein coreference resolution the best system achieved 34% in F-score. CONCLUSIONS: Detailed analysis performed on the results improves our insight into the problem and suggests the directions for further improvements.

BACKGROUND: Current research has shown that major difficulties in event extraction for the biomedical domain are traceable to coreference. Therefore, coreference resolution is believed to be useful for improving event extraction. To address coreference resolution in molecular biology literature, the Protein Coreference (COREF) task was arranged in the BioNLP Shared Task (BioNLP-ST, hereafter) 2011, as a supporting task. However, the shared task results indicated that transferring coreference resolution methods developed for other domains to the biological domain was not a straight-forward task, due to the domain differences in the coreference phenomena. RESULTS: We analyzed the contribution of domain-specific information, including the information that indicates the protein type, in a rule-based protein coreference resolution system. In particular, the domain-specific information is encoded into semantic classification modules for which the output is used in different components of the coreference resolution. We compared our system with the top four systems in the BioNLP-ST 2011; surprisingly, we found that the minimal configuration had outperformed the best system in the BioNLP-ST 2011. Analysis of the experimental results revealed that semantic classification, using protein information, has contributed to an increase in performance by 2.3% on the test data, and 4.0% on the development data, in F-score. CONCLUSIONS: The use of domain-specific information in semantic classification is important for effective coreference resolution. Since it is difficult to transfer domain-specific information across different domains, we need to continue seek for methods to utilize such information in coreference resolution.

We present the preparation, resources, results and analysis of three tasks of the BioNLP Shared Task 2011: the main tasks on Infectious Diseases (ID) and Epigenetics and Post-translational Modifications (EPI), and the supporting task on Entity Relations (REL). The two main tasks represent extensions of the event extraction model introduced in the BioNLP Shared Task 2009 (ST'09) to two new areas of biomedical scientific literature, each motivated by the needs of specific biocuration tasks. The ID task concerns the molecular mechanisms of infection, virulence and resistance, focusing in particular on the functions of a class of signaling systems that are ubiquitous in bacteria. The EPI task is dedicated to the extraction of statements regarding chemical modifications of DNA and proteins, with particular emphasis on changes relating to the epigenetic control of gene expression. By contrast to these two application-oriented main tasks, the REL task seeks to support extraction in general by separating challenges relating to part-of relations into a subproblem that can be addressed by independent systems. Seven groups participated in each of the two main tasks and four groups in the supporting task. The participating systems indicated advances in the capability of event extraction methods and demonstrated generalization in many aspects: from abstracts to full texts, from previously considered subdomains to new ones, and from the ST'09 extraction targets to other entities and events. The highest performance achieved in the supporting task REL, 58% F-score, is broadly comparable with levels reported for other relation extraction tasks. For the ID task, the highest-performing system achieved 56% F-score, comparable to the state-of-the-art performance at the established ST'09 task. In the EPI task, the best result was 53% F-score for the full set of extraction targets and 69% F-score for a reduced set of core extraction targets, approaching a level of performance sufficient for user-facing applications. In this study, we extend on previously reported results and perform further analyses of the outputs of the participating systems. We place specific emphasis on aspects of system performance relating to real-world applicability, considering alternate evaluation metrics and performing additional manual analysis of system outputs. We further demonstrate that the strengths of extraction systems can be combined to improve on the performance achieved by any system in isolation. The manually annotated corpora, supporting resources, and evaluation tools for all tasks are available from http://www.bionlp-st.org and the tasks continue as open challenges for all interested parties.

OBJECTIVE: A system that translates narrative text in the medical domain into structured representation is in great demand. The system performs three sub-tasks: concept extraction, assertion classification, and relation identification. DESIGN: The overall system consists of five steps: (1) pre-processing sentences, (2) marking noun phrases (NPs) and adjective phrases (APs), (3) extracting concepts that use a dosage-unit dictionary to dynamically switch two models based on Conditional Random Fields (CRF), (4) classifying assertions based on voting of five classifiers, and (5) identifying relations using normalized sentences with a set of effective discriminating features. MEASUREMENTS: Macro-averaged and micro-averaged precision, recall and F-measure were used to evaluate results. RESULTS: The performance is competitive with the state-of-the-art systems with micro-averaged F-measure of 0.8489 for concept extraction, 0.9392 for assertion classification and 0.7326 for relation identification. CONCLUSIONS: The system exploits an array of common features and achieves state-of-the-art performance. Prudent feature engineering sets the foundation of our systems. In concept extraction, we demonstrated that switching models, one of which is especially designed for telegraphic sentences, improved extraction of the treatment concept significantly. In assertion classification, a set of features derived from a rule-based classifier were proven to be effective for the classes such as conditional and possible. These classes would suffer from data scarcity in conventional machine-learning methods. In relation identification, we use two-staged architecture, the second of which applies pairwise classifiers to possible candidate classes. This architecture significantly improves performance.

OBJECTIVE: To create a highly accurate coreference system in discharge summaries for the 2011 i2b2 challenge. The coreference categories include Person, Problem, Treatment, and Test. DESIGN: An integrated coreference resolution system was developed by exploiting Person attributes, contextual semantic clues, and world knowledge. It includes three subsystems: Person coreference system based on three Person attributes, Problem/Treatment/Test system based on numerous contextual semantic extractors and world knowledge, and Pronoun system based on a multi-class support vector machine classifier. The three Person attributes are patient, relative and hospital personnel. Contextual semantic extractors include anatomy, position, medication, indicator, temporal, spatial, section, modifier, equipment, operation, and assertion. The world knowledge is extracted from external resources such as Wikipedia. MEASUREMENTS: Micro-averaged precision, recall and F-measure in MUC, BCubed and CEAF were used to evaluate results. RESULTS: The system achieved an overall micro-averaged precision, recall and F-measure of 0.906, 0.925, and 0.915, respectively, on test data (from four hospitals) released by the challenge organizers. It achieved a precision, recall and F-measure of 0.905, 0.920 and 0.913, respectively, on test data without Pittsburgh data. We ranked the first out of 20 competing teams. Among the four sub-tasks on Person, Problem, Treatment, and Test, the highest F-measure was seen for Person coreference. CONCLUSIONS: This system achieved encouraging results. The Person system can determine whether personal pronouns and proper names are coreferent or not. The Problem/Treatment/Test system benefits from both world knowledge in evaluating the similarity of two mentions and contextual semantic extractors in identifying semantic clues. The Pronoun system can automatically detect whether a Pronoun mention is coreferent to that of the other four types. This study demonstrates that it is feasible to accomplish the coreference task in discharge summaries.

OBJECTIVE: To develop a system to extract follow-up information from radiology reports. The method may be used as a component in a system which automatically generates follow-up information in a timely fashion. METHODS: A novel method of combining an LSP (labeled sequential pattern) classifier with a CRF (conditional random field) recognizer was devised. The LSP classifier filters out irrelevant sentences, while the CRF recognizer extracts follow-up and time phrases from candidate sentences presented by the LSP classifier. MEASUREMENTS: The standard performance metrics of precision (P), recall (R), and F measure (F) in the exact and inexact matching settings were used for evaluation. RESULTS: Four experiments conducted using 20,000 radiology reports showed that the CRF recognizer achieved high performance without time-consuming feature engineering and that the LSP classifier further improved the performance of the CRF recognizer. The performance of the current system is P=0.90, R=0.86, F=0.88 in the exact matching setting and P=0.98, R=0.93, F=0.95 in the inexact matching setting. CONCLUSION: The experiments demonstrate that the system performs far better than a baseline rule-based system and is worth considering for deployment trials in an alert generation system. The LSP classifier successfully compensated for the inherent weakness of CRF, that is, its inability to use global information.

Abstract As described in this paper, we specifically examine the structural learning problem of a supertagging task. Supertagging is a task to assign the most probable lexical entry to each word in a sentence. A supertagger is extremely important for a lexicalized grammar parser because an accurate supertagger can greatly reduce lexical ambiguity in downstream parser. Supertagging is more challenging than conventional sequence labeling tasks (e.g., part-of-speech tagging). First, the supertags are numerous. Supertags are the lexical entries defined in a lexicalized grammar, which consists of rich syntactic/semantic information. Second, the inter-supertag relation is more complex. A proper supertag assignment is expected to be compatible with other supertag assignments in a sentence to construct a parse tree. Commonly used adjacent label features (e.g., first-order edge feature) in a sequence labeling model are too rough for the supertagging task. Long-range information is extremely important for the supertagging task. Two approaches to consider long-range information in a supertagger's training stage are proposed. Specifically, we propose a dependency-informed supertagger to use word-to-word dependency derived from a dependency parser and generate long-range features as soft constraints in the training. In the forest-guided supertagger, we constrain the classifier to learn in a grammar-satisfying space and use a CFG filter to impose grammar constraints for the update of model parameters. The experiments show that the proposed structure-guided supertaggers perform significantly better than the baseline supertaggers. Based on the improved supertaggers, theF-score of the final parser is also improved. Using the forest-guided supertagger in a shift-reduce HPSG parser, we achieved a competitive parsing performance of 89.31%F-score with higher parsing speed than that of a state-of-the-art HPSG parser.

This paper is about supervised and semi-supervised dimensionality reduction (DR) by generating spectral embeddings from multi-output data based on the pairwise proximity information. Two flexible and generic frameworks are proposed to achieve supervised DR (SDR) for multilabel classification. One is able to extend any existing single-label SDR to multilabel via sample duplication, referred to as MESD. The other is a multilabel design framework that tackles the SDR problem by computing weight (proximity) matrices based on simultaneous feature and label information, referred to as MOPE, as a generalization of many current techniques. A diverse set of different schemes for label-based proximity calculation, as well as a mechanism for combining label-based and feature-based weight information by considering information importance and prioritization, are proposed for MOPE. Additionally, we summarize many current spectral methods for unsupervised DR (UDR), single/multilabel SDR, and semi-supervised DR (SSDR) and express them under a common template representation as a general guide to researchers in the field. We also propose a general framework for achieving SSDR by combining existing SDR and UDR models, and also a procedure of reducing the computational cost via learning with a target set of relation features. The effectiveness of our proposed methodologies is demonstrated with experiments with document collections for multilabel text categorization from the natural language processing domain.

Bilingual dictionaries can be automatically extended by new translations using comparable corpora. The general idea is based on the assumption that similar words have similar contexts across languages. However, previous studies have mainly focused on Indo-European languages, or use only a bag-of-words model to describe the context. Furthermore, we argue that it is helpful to extract only the statistically significant context, instead of using all context. The present approach addresses these issues in the following manner. First, based on the context of a word with an unknown translation (query word), we extract salient pivot words. Pivot words are words for which a translation is already available in a bilingual dictionary. For the extraction of salient pivot words, we use a Bayesian estimation of the point-wise mutual information to measure statistical significance. In the second step, we match these pivot words across languages to identify translation candidates for the query word. We therefore calculate a similarity score between the query word and a translation candidate using the probability that the same pivots will be extracted for both the query word and the translation candidate. The proposed method uses several context positions, namely, a bag-of-words of one sentence, and the successors, predecessors, and siblings with respect to the dependency parse tree of the sentence. In order to make these context positions comparable across Japanese and English, which are unrelated languages, we use several heuristics to adjust the dependency trees appropriately. We demonstrate that the proposed method significantly increases the accuracy of word translations, as compared to previous methods.

UNLABELLED: Often, the most informative genes have to be selected from different gene sets and several computer gene ranking algorithms have been developed to cope with the problem. To help researchers decide which algorithm to use, we developed the analysis of gene ranking algorithms (AGRA) system that offers a novel technique for comparing ranked lists of genes. The most important feature of AGRA is that no previous knowledge of gene ranking algorithms is needed for their comparison. Using the text mining system finding-associated concepts with text analysis. AGRA defines what we call biomedical concept space (BCS) for each gene list and offers a comparison of the gene lists in six different BCS categories. The uploaded gene lists can be compared using two different methods. In the first method, the overlap between each pair of two gene lists of BCSs is calculated. The second method offers a text field where a specific biomedical concept can be entered. AGRA searches for this concept in each gene lists' BCS, highlights the rank of the concept and offers a visual representation of concepts ranked above and below it. AVAILABILITY AND IMPLEMENTATION: Available at http://agra.fzv.uni-mb.si/, implemented in Java and running on the Glassfish server. CONTACT: [email protected].

MOTIVATION: Discovering useful associations between biomedical concepts has been one of the main goals in biomedical text-mining, and understanding their biomedical contexts is crucial in the discovery process. Hence, we need a text-mining system that helps users explore various types of (possibly hidden) associations in an easy and comprehensible manner. RESULTS: This article describes FACTA+, a real-time text-mining system for finding and visualizing indirect associations between biomedical concepts from MEDLINE abstracts. The system can be used as a text search engine like PubMed with additional features to help users discover and visualize indirect associations between important biomedical concepts such as genes, diseases and chemical compounds. FACTA+ inherits all functionality from its predecessor, FACTA, and extends it by incorporating three new features: (i) detecting biomolecular events in text using a machine learning model, (ii) discovering hidden associations using co-occurrence statistics between concepts, and (iii) visualizing associations to improve the interpretability of the output. To the best of our knowledge, FACTA+ is the first real-time web application that offers the functionality of finding concepts involving biomolecular events and visualizing indirect associations of concepts with both their categories and importance. AVAILABILITY: FACTA+ is available as a web application at http://refine1-nactem.mc.man.ac.uk/facta/, and its visualizer is available at http://refine1-nactem.mc.man.ac.uk/facta-visualizer/. CONTACT: [email protected].

MOTIVATION: Understanding key biological processes (bioprocesses) and their relationships with constituent biological entities and pharmaceutical agents is crucial for drug design and discovery. One way to harvest such information is searching the literature. However, bioprocesses are difficult to capture because they may occur in text in a variety of textual expressions. Moreover, a bioprocess is often composed of a series of bioevents, where a bioevent denotes changes to one or a group of cells involved in the bioprocess. Such bioevents are often used to refer to bioprocesses in text, which current techniques, relying solely on specialized lexicons, struggle to find. RESULTS: This article presents a range of methods for finding bioprocess terms and events. To facilitate the study, we built a gold standard corpus in which terms and events related to angiogenesis, a key biological process of the growth of new blood vessels, were annotated. Statistics of the annotated corpus revealed that over 36% of the text expressions that referred to angiogenesis appeared as events. The proposed methods respectively employed domain-specific vocabularies, a manually annotated corpus and unstructured domain-specific documents. Evaluation results showed that, while a supervised machine-learning model yielded the best precision, recall and F1 scores, the other methods achieved reasonable performance and less cost to develop. AVAILABILITY: The angiogenesis vocabularies, gold standard corpus, annotation guidelines and software described in this article are available at http://text0.mib.man.ac.uk/~mbassxw2/angiogenesis/ CONTACT: [email protected].

BACKGROUND: Bio-molecular event extraction from literature is recognized as an important task of bio text mining and, as such, many relevant systems have been developed and made available during the last decade. While such systems provide useful services individually, there is a need for a meta-service to enable comparison and ensemble of such services, offering optimal solutions for various purposes. RESULTS: We have integrated nine event extraction systems in the U-Compare framework, making them intercompatible and interoperable with other U-Compare components. The U-Compare event meta-service provides various meta-level features for comparison and ensemble of multiple event extraction systems. Experimental results show that the performance improvements achieved by the ensemble are significant. CONCLUSIONS: While individual event extraction systems themselves provide useful features for bio text mining, the U-Compare meta-service is expected to improve the accessibility to the individual systems, and to enable meta-level uses over multiple event extraction systems such as comparison and ensemble.

This paper presents the preparation, results and analysis of the BioNLP’09 shared task on event extraction, organized to address the automatic extraction of fine-grained information from the scientific literature on molecular biology. A representation of this information was defined taking into account both the biological and computational requirements of the task, and corpus resources manually annotated by domain experts provided to task participants. To create a basis for further progress, emphasis was placed on providing fine-grained evaluation that isolates different subtasks and allows the analysis of different aspects of the results through various evaluation criteria. In introducing this new task to the community, we made an effort to reduce the cost of participation by making common natural language processing tools, data, and evaluation methods easily accessible. The task received community-wide participation, establishing the state-of-the-art performance at fine-grained event extraction as well as allowing the identification of remaining challenges and suggesting directions for future improvements. All the resources and results of the shared task are publicly available and an online evaluation on blind test data accessible at http://www-tsujii.is.s.u-tokyo.ac.jp/GENIA/SharedTask/.

This article presents a novel approach to event extraction from biological text using Markov Logic. It can be described by three design decisions: (1) instead of building a pipeline using local classifiers, we design and learn a joint probabilistic model over events in a sentence; (2) instead of developing specific inference and learning algorithms for our joint model, we apply Markov Logic, a general purpose Statistical Relation Learning language, for this task; (3) we represent events as relations over the token indices of a sentence, as opposed to structures that relate event entities to gene or protein mentions. In this article, we extend our original work by providing an error analysis for binding events. Moreover, we investigate the impact of different loss functions to precision, recall and F-measure. Finally, we show how to extract events of different types that share the same event clue. This extension allowed us to improve our performance our performance even further, leading to the third best scores for task 1 (in close range to the second place) and the best results for task 2 with a 14% point margin.

This paper describes the rational behind designing workflow systems based on the Unix make by showing a number of idioms useful for workflows comprising many tasks. It also demonstrates a specific design and implementation of such a workflow system called GXP make. GXP make supports all the features of GNU make and extends its platforms from single node systems to clusters, clouds, supercomputers, and distributed systems. Interestingly, it is achieved by a very small code base that does not modify GNU make implementation at all. While being not ideal for performance, it achieved a useful performance and scalability of dispatching one million tasks in approximately 16,000 seconds (60 tasks per second, including dependence analysis) on an 8 core Intel Nehalem node. For real applications, recognition and classification of protein-protein interactions from biomedical texts on a supercomputer with more than 8,000 cores are described.

MOTIVATION: There has recently been a notable shift in biomedical information extraction (IE) from relation models toward the more expressive event model, facilitated by the maturation of basic tools for biomedical text analysis and the availability of manually annotated resources. The event model allows detailed representation of complex natural language statements and can support a number of advanced text mining applications ranging from semantic search to pathway extraction. A recent collaborative evaluation demonstrated the potential of event extraction systems, yet there have so far been no studies of the generalization ability of the systems nor the feasibility of large-scale extraction. RESULTS: This study considers event-based IE at PubMed scale. We introduce a system combining publicly available, state-of-the-art methods for domain parsing, named entity recognition and event extraction, and test the system on a representative 1% sample of all PubMed citations. We present the first evaluation of the generalization performance of event extraction systems to this scale and show that despite its computational complexity, event extraction from the entire PubMed is feasible. We further illustrate the value of the extraction approach through a number of analyses of the extracted information. AVAILABILITY: The event detection system and extracted data are open source licensed and available at http://bionlp.utu.fi/.

UNLABELLED: Text mining from the biomedical literature is of increasing importance, yet it is not easy for the bioinformatics community to create and run text mining workflows due to the lack of accessibility and interoperability of the text mining resources. The U-Compare system provides a wide range of bio text mining resources in a highly interoperable workflow environment where workflows can very easily be created, executed, evaluated and visualized without coding. We have linked U-Compare to Taverna, a generic workflow system, to expose text mining functionality to the bioinformatics community. AVAILABILITY: http://u-compare.org/taverna.html, http://u-compare.org.

MOTIVATION: Metabolic and signaling pathways are an increasingly important part of organizing knowledge in systems biology. They serve to integrate collective interpretations of facts scattered throughout literature. Biologists construct a pathway by reading a large number of articles and interpreting them as a consistent network, but most of the models constructed currently lack direct links to those articles. Biologists who want to check the original articles have to spend substantial amounts of time to collect relevant articles and identify the sections relevant to the pathway. Furthermore, with the scientific literature expanding by several thousand papers per week, keeping a model relevant requires a continuous curation effort. In this article, we present a system designed to integrate a pathway visualizer, text mining systems and annotation tools into a seamless environment. This will enable biologists to freely move between parts of a pathway and relevant sections of articles, as well as identify relevant papers from large text bases. The system, PathText, is developed by Systems Biology Institute, Okinawa Institute of Science and Technology, National Centre for Text Mining (University of Manchester) and the University of Tokyo, and is being used by groups of biologists from these locations.

MOTIVATION: The ultimate goal of abbreviation management is to disambiguate every occurrence of an abbreviation into its expanded form (concept or sense). To collect expanded forms for abbreviations, previous studies have recognized abbreviations and their expanded forms in parenthetical expressions of bio-medical texts. However, expanded forms extracted by abbreviation recognition are mixtures of concepts/senses and their term variations. Consequently, a list of expanded forms should be structured into a sense inventory, which provides possible concepts or senses for abbreviation disambiguation. RESULTS: A sense inventory is a key to robust management of abbreviations. Therefore, we present a supervised approach for clustering expanded forms. The experimental result reports 0.915 F1 score in clustering expanded forms. We then investigate the possibility of conflicts of protein and gene names with abbreviations. Finally, an experiment of abbreviation disambiguation on the sense inventory yielded 0.984 accuracy and 0.986 F1 score using the dataset obtained from MEDLINE abstracts. AVAILABILITY: The sense inventory and disambiguator of abbreviations are accessible at http://www.nactem.ac.uk/software/acromine/ and http://www.nactem.ac.uk/software/acromine_disambiguation/.

MOTIVATION: Text mining technologies have been shown to reduce the laborious work involved in organizing the vast amount of information hidden in the literature. One challenge in text mining is linking ambiguous word forms to unambiguous biological concepts. This article reports on a comprehensive study on resolving the ambiguity in mentions of biomedical named entities with respect to model organisms and presents an array of approaches, with focus on methods utilizing natural language parsers. RESULTS: We build a corpus for organism disambiguation where every occurrence of protein/gene entity is manually tagged with a species ID, and evaluate a number of methods on it. Promising results are obtained by training a machine learning model on syntactic parse trees, which is then used to decide whether an entity belongs to the model organism denoted by a neighbouring species-indicating word (e.g. yeast). The parser-based approaches are also compared with a supervised classification method and results indicate that the former are a more favorable choice when domain portability is of concern. The best overall performance is obtained by combining the strengths of syntactic features and supervised classification. AVAILABILITY: The corpus and demo are available at http://www.nactem.ac.uk/deca_details/start.cgi, and the software is freely available as U-Compare components (Kano et al., 2009): NaCTeM Species Word Detector and NaCTeM Species Disambiguator. U-Compare is available at http://-compare.org/

In the recent decades, high-throughput screening methods were established, bringing forth major breakthroughs in the fields of molecular biology and biomedicine. Since researchers in these fields need to interpret an enormous quantity of data and the publication rates of scientific articles are exploding, demands on text mining technology are growing with each passing year.

Currently, relation extraction (RE) and event extraction (EE) are the two main streams of biological information extraction. In 2009, the majority of these RE and EE research efforts were centered around the BioCreative II.5 Protein-Protein Interaction (PPI) challenge and the "BioNLP event extraction shared task." Although these challenges took somewhat different approaches, they share the same ultimate goal of extracting bio-knowledge from the literature. This paper compares the two challenge task definitions, and presents a unified system that was successfully applied in both these and several other PPI extraction task settings. The AkaneRE system has three parts: A core engine for RE, a pool of modules for specific solutions, and a configuration language to adapt the system to different tasks. The core engine is based on machine learning, using either Support Vector Machines or Statistical Classifiers and features extracted from given training data. The specific modules solve tasks like sentence boundary detection, tokenization, stemming, part-of-speech tagging, parsing, named entity recognition, generation of potential relations, generation of machine learning features for each relation, and finally, assignment of confidence scores and ranking of candidate relations. With these components, the AkaneRE system produces state-of-the-art results, and the system is freely available for academic purposes at http://www-tsujii.is.s.u-tokyo.ac.jp/satre/akane/.

We propose a method for classifying opinions which captures the role of linguistic modalities in the sentence. We use features than simple bag-of-words or opinion-holding predicates. The method is based on a machine learning and utilizes opinion-holding predicates and linguistic modalities as features. Two different detectors help to classify the opinions: the opinion-holding predicate detector and the modality detector. An opinion in the target is first parsed into a dependency structure, and then the opinion-holding predicates and modalities stick onto the leaf nodes of the dependency tree. The whole tree is regarded as input features of the opinion, and it becomes the input of tree kernel support vector machines. We have applied method to opinions in Japanese about television programs, and have confirmed the effectiveness of the method against conventional bag-of-words features, or against simple opinion-holding predicates features

This paper presents a novel document classification method using all substrings as features. Although tokenized words are not enough for determining a class of a document, learning by using all substrings has a prohibitive computational cost because the number of all candidate substrings can be very large. We show that the idea of equivalent classes of substrings can help determine all effective substrings exhaustively in linear time. Moreover, by applying L1 regularization to our model, we obtain a compact result, which makes an inference extremely efficient in time and space, and robust even if we use substrings of all lengths. In experiments, we show that our method can extract effective substrings efficiently, and achieved more accurate results and the its inference was faster than the results using previous methods.

SUMMARY: Due to the increasing number of text mining resources (tools and corpora) available to biologists, interoperability issues between these resources are becoming significant obstacles to using them effectively. UIMA, the Unstructured Information Management Architecture, is an open framework designed to aid in the construction of more interoperable tools. U-Compare is built on top of the UIMA framework, and provides both a concrete framework for out-of-the-box text mining and a sophisticated evaluation platform allowing users to run specific tools on any target text, generating both detailed statistics and instance-based visualizations of outputs. U-Compare is a joint project, providing the world's largest, and still growing, collection of UIMA-compatible resources. These resources, originally developed by different groups for a variety of domains, include many famous tools and corpora. U-Compare can be launched straight from the web, without needing to be manually installed. All U-Compare components are provided ready-to-use and can be combined easily via a drag-and-drop interface without any programming. External UIMA components can also simply be mixed with U-Compare components, without distinguishing between locally and remotely deployed resources. AVAILABILITY: http://u-compare.org/

MOTIVATION: While text mining technologies for biomedical research have gained popularity as a way to take advantage of the explosive growth of information in text form in biomedical papers, selecting appropriate natural language processing (NLP) tools is still difficult for researchers who are not familiar with recent advances in NLP. This article provides a comparative evaluation of several state-of-the-art natural language parsers, focusing on the task of extracting protein-protein interaction (PPI) from biomedical papers. We measure how each parser, and its output representation, contributes to accuracy improvement when the parser is used as a component in a PPI system. RESULTS: All the parsers attained improvements in accuracy of PPI extraction. The levels of accuracy obtained with these different parsers vary slightly, while differences in parsing speed are larger. The best accuracy in this work was obtained when we combined Miyao and Tsujii's Enju parser and Charniak and Johnson's reranking parser, and the accuracy is better than the state-of-the-art results on the same data. AVAILABILITY: The PPI extraction system used in this work (AkanePPI) is available online at http://www-tsujii.is.s.u-tokyo.ac.jp/downloads/downloads.cgi. The evaluated parsers are also available online from each developer's site.

BACKGROUND: The number of corpora, collections of structured texts, has been increasing, as a result of the growing interest in the application of natural language processing methods to biological texts. Many named entity recognition (NER) systems have been developed based on these corpora. However, in the biomedical community, there is yet no general consensus regarding named entity annotation; thus, the resources are largely incompatible, and it is difficult to compare the performance of systems developed on resources that were divergently annotated. On the other hand, from a practical application perspective, it is desirable to utilize as many existing annotated resources as possible, because annotation is costly. Thus, it becomes a task of interest to integrate the heterogeneous annotations in these resources. RESULTS: We explore the potential sources of incompatibility among gene and protein annotations that were made for three common corpora: GENIA, GENETAG and AIMed. To show the inconsistency in the corpora annotations, we first tackle the incompatibility problem caused by corpus integration, and we quantitatively measure the effect of this incompatibility on protein mention recognition. We find that the F-score performance declines tremendously when training with integrated data, instead of training with pure data; in some cases, the performance drops nearly 12%. This degradation may be caused by the newly added heterogeneous annotations, and cannot be fixed without an understanding of the heterogeneities that exist among the corpora. Motivated by the result of this preliminary experiment, we further qualitatively analyze a number of possible sources for these differences, and investigate the factors that would explain the inconsistencies, by performing a series of well-designed experiments. Our analyses indicate that incompatibilities in the gene/protein annotations exist mainly in the following four areas: the boundary annotation conventions, the scope of the entities of interest, the distribution of annotated entities, and the ratio of overlap between annotated entities. We further suggest that almost all of the incompatibilities can be prevented by properly considering the four aspects aforementioned. CONCLUSION: Our analysis covers the key similarities and dissimilarities that exist among the diverse gene/protein corpora. This paper serves to improve our understanding of the differences in the three studied corpora, which can then lead to a better understanding of the performance of protein recognizers that are based on the corpora.

The bilingual lexicon is an expensive but critical resource for multilingual applications in natural language processing. This article proposes an integrated framework for building a bilingual lexicon between the Chinese and Japanese languages. Since the language pair Chinese-Japanese does not include English, which is a central language of the world, few large-scale bilingual resources between Chinese and Japanese have been constructed. One solution to alleviate this problem is to build a Chinese-Japanese bilingual lexicon through English as the pivot language. In addition to the pivotal approach, we can make use of the characteristics of Chinese and Japanese languages that use Han characters. We incorporate a translation model obtained from a small Chinese-Japanese lexicon and use the similarity of the hanzi and kanji characters by using the log-linear model. Our experimental results show that the use of the pivotal approach can improve the translation performance over the translation model built from a small Chinese-Japanese lexicon. The results also demonstrate that the similarity between the hanzi and kanji characters provides a positive effect for translating technical terms.

This paper presents a framework for searching text regions with specifying annotated information in tag-annotated text by using Region Algebra. We extend the efficient algorithm for region algebra to handle both nested and crossed regions and introduce variables for attribute values to treat tag-annotations in which attributes indicate another tag regions. Our framework have been implemented in a text search engine for MEDLINE, which is a large textbase of abstracts in medical science. Experiments in tag-annotated MEDLINE abstracts demonstrate the effectiveness of specifying annotations and the efficiency of our framework.

Kleio is an advanced information retrieval (IR) system developed at the UK National Centre for Text Mining (NaCTeM)1. The system offers textual and metadata searches across MEDLINE and provides enhanced searching functionality by leveraging terminology management technologies.

UNLABELLED: FACTA is a text search engine for MEDLINE abstracts, which is designed particularly to help users browse biomedical concepts (e.g. genes/proteins, diseases, enzymes and chemical compounds) appearing in the documents retrieved by the query. The concepts are presented to the user in a tabular format and ranked based on the co-occurrence statistics. Unlike existing systems that provide similar functionality, FACTA pre-indexes not only the words but also the concepts mentioned in the documents, which enables the user to issue a flexible query (e.g. free keywords or Boolean combinations of keywords/concepts) and receive the results immediately even when the number of the documents that match the query is very large. The user can also view snippets from MEDLINE to get textual evidence of associations between the query terms and the concepts. The concept IDs and their names/synonyms for building the indexes were collected from several biomedical databases and thesauri, such as UniProt, BioThesaurus, UMLS, KEGG and DrugBank. AVAILABILITY: The system is available at http://www.nactem.ac.uk/software/facta/

A recent posting to the BioNLP mailing list notes that the past few months of 2008 have seen the appearance of over fifty papers on biomedical natural language processing/text mining (BioNLP). This number (which included medical, as well as genomic work) represents about as many papers on genomic language processing as existed in all of PubMed at the end of 2003 [1] – just five years ago, and the current supplement in BMC Bioinformatics presents another ten! These papers have in common the fact that they are follow-on work to papers originally published in the proceedings of the BioNLP 2008 workshop at the annual meeting of the Association for Computational Linguistics (ACL). All have gone through a separate rigorous review process and represent an advance beyond the work originally presented at the workshop. Like the annual BioNLP workshop itself, they represent a wide cross-section of the type of work that goes on in BioNLP today.

BACKGROUND: Advanced Text Mining (TM) such as semantic enrichment of papers, event or relation extraction, and intelligent Question Answering have increasingly attracted attention in the bio-medical domain. For such attempts to succeed, text annotation from the biological point of view is indispensable. However, due to the complexity of the task, semantic annotation has never been tried on a large scale, apart from relatively simple term annotation. RESULTS: We have completed a new type of semantic annotation, event annotation, which is an addition to the existing annotations in the GENIA corpus. The corpus has already been annotated with POS (Parts of Speech), syntactic trees, terms, etc. The new annotation was made on half of the GENIA corpus, consisting of 1,000 Medline abstracts. It contains 9,372 sentences in which 36,114 events are identified. The major challenges during event annotation were (1) to design a scheme of annotation which meets specific requirements of text annotation, (2) to achieve biology-oriented annotation which reflect biologists' interpretation of text, and (3) to ensure the homogeneity of annotation quality across annotators. To meet these challenges, we introduced new concepts such as Single-facet Annotation and Semantic Typing, which have collectively contributed to successful completion of a large scale annotation. CONCLUSION: The resulting event-annotated corpus is the largest and one of the best in quality among similar annotation efforts. We expect it to become a valuable resource for NLP (Natural Language Processing)-based TM in the bio-medical domain.

BACKGROUND: Associating literature with pathways poses new challenges to the Text Mining (TM) community. There are three main challenges to this task: (1) the identification of the mapping position of a specific entity or reaction in a given pathway, (2) the recognition of the causal relationships among multiple reactions, and (3) the formulation and implementation of required inferences based on biological domain knowledge. RESULTS: To address these challenges, we constructed new resources to link the text with a model pathway; they are: the GENIA pathway corpus with event annotation and NF-kB pathway. Through their detailed analysis, we address the untapped resource, 'bio-inference,' as well as the differences between text and pathway representation. Here, we show the precise comparisons of their representations and the nine classes of 'bio-inference' schemes observed in the pathway corpus. CONCLUSIONS: We believe that the creation of such rich resources and their detailed analysis is the significant first step for accelerating the research of the automatic construction of pathway from text.

BACKGROUND: Previous studies of named entity recognition have shown that a reasonable level of recognition accuracy can be achieved by using machine learning models such as conditional random fields or support vector machines. However, the lack of training data (i.e. annotated corpora) makes it difficult for machine learning-based named entity recognizers to be used in building practical information extraction systems. RESULTS: This paper presents an active learning-like framework for reducing the human effort required to create named entity annotations in a corpus. In this framework, the annotation work is performed as an iterative and interactive process between the human annotator and a probabilistic named entity tagger. Unlike active learning, our framework aims to annotate all occurrences of the target named entities in the given corpus, so that the resulting annotations are free from the sampling bias which is inevitable in active learning approaches. CONCLUSION: We evaluate our framework by simulating the annotation process using two named entity corpora and show that our approach can reduce the number of sentences which need to be examined by the human annotator. The cost reduction achieved by the framework could be drastic when the target named entities are sparse.

Probabilistic modeling of lexicalized grammars is difficult because these grammars exploit complicated data structures, such as typed feature structures. This prevents us from applying common methods of probabilistic modeling in which a complete structure is divided into sub-structures under the assumption of statistical independence among sub-structures. For example, part-of-speech tagging of a sentence is decomposed into tagging of each word, and CFG parsing is split into applications of CFG rules. These methods have relied on the structure of the target problem, namely lattices or trees, and cannot be applied to graph structures including typed feature structures. This article proposes the feature forest model as a solution to the problem of probabilistic modeling of complex data structures including typed feature structures. The feature forest model provides a method for probabilistic modeling without the independence assumption when probabilistic events are represented with feature forests. Feature forests are generic data structures that represent ambiguous trees in a packed forest structure. Feature forest models are maximum entropy models defined over feature forests. A dynamic programming algorithm is proposed for maximum entropy estimation without unpacking feature forests. Thus probabilistic modeling of any data structures is possible when they are represented by feature forests. This article also describes methods for representing HPSG syntactic structures and predicate-argument structures with feature forests. Hence, we describe a complete strategy for developing probabilistic models for HPSG parsing. The effectiveness of the proposed methods is empirically evaluated through parsing experiments on the Penn Treebank, and the promise of applicability to parsing of real-world sentences is discussed.

There are two streams of research in Computational Linguistics and Natural language Processing, the empiricist and rationalist traditions. Theories and computational techniques in these two streams have been developed separately and different in nature. Although the two traditions have been considered irreconcilable and have often been antagonistic toward each other, I have contention with this assertion, and thus claim that these two research streams in linguistics, despite or due to their differences, can be complementary to each other and should be combined into a unified methodology. I will demonstrate in my talk that there have been interesting developments in this direction of integration, and would like to discuss some of the recent results with their implications on engineering application.

MOTIVATION: One of the bottlenecks of biomedical data integration is variation of terms. Exact string matching often fails to associate a name with its biological concept, i.e. ID or accession number in the database, due to seemingly small differences of names. Soft string matching potentially enables us to find the relevant ID by considering the similarity between the names. However, the accuracy of soft matching highly depends on the similarity measure employed. RESULTS: We used logistic regression for learning a string similarity measure from a dictionary. Experiments using several large-scale gene/protein name dictionaries showed that the logistic regression-based similarity measure outperforms existing similarity measures in dictionary look-up tasks. AVAILABILITY: A dictionary look-up system using the similarity measures described in this article is available at http://text0.mib.man.ac.uk/software/mldic/.

This paper introduces a novel framework for the accurate retrieval of relational concepts from huge texts. Prior to retrieval, all sentences are annotated with predicate argument structures and ontological identifiers by applying a deep parser and a term recognizer. During the run time, user requests are converted into queries of region algebra on these annotations. Structural matching with pre-computed semantic annotations establishes the accurate and efficient retrieval of relational concepts. This framework was applied to a text retrieval system for MEDLINE. Experiments on the retrieval of biomedical correlations revealed that the cost is sufficiently small for real-time applications and that the retrieval precision is significantly improved.

We present a practical HPSG parser for English, an intelligent search engine to retrieve MEDLINE abstracts that represent biomedical events and an efficient MED-LINE search tool helping users to find information about biomedical entities such as genes, proteins, and the interactions between them.

This paper presents techniques to apply semi-CRFs to Named Entity Recognition tasks with a tractable computational cost. Our framework can handle an NER task that has long named entities and many labels which increase the computational cost. To reduce the computational cost, we propose two techniques: the first is the use of feature forests, which enables us to pack feature-equivalent states, and the second is the introduction of a filtering process which significantly reduces the number of candidate states. This framework allows us to use a rich set of features extracted from the chunk-based representation that can capture informative characteristics of entities. We also introduce a simple trick to transfer information about distant entities by embedding label information into non-entity labels. Experimental results show that our model achieves an F-score of 71.48% on the JNLPBA 2004 shared task without using any external resources or post-processing techniques.

BACKGROUND: Automatic recognition of relations between a specific disease term and its relevant genes or protein terms is an important practice of bioinformatics. Considering the utility of the results of this approach, we identified prostate cancer and gene terms with the ID tags of public biomedical databases. Moreover, considering that genetics experts will use our results, we classified them based on six topics that can be used to analyze the type of prostate cancers, genes, and their relations. METHODS: We developed a maximum entropy-based named entity recognizer and a relation recognizer and applied them to a corpus-based approach. We collected prostate cancer-related abstracts from MEDLINE, and constructed an annotated corpus of gene and prostate cancer relations based on six topics by biologists. We used it to train the maximum entropy-based named entity recognizer and relation recognizer. RESULTS: Topic-classified relation recognition achieved 92.1% precision for the relation (an increase of 11.0% from that obtained in a baseline experiment). For all topics, the precision was between 67.6 and 88.1%. CONCLUSION: A series of experimental results revealed two important findings: a carefully designed relation recognition system using named entity recognition can improve the performance of relation recognition, and topic-classified relation recognition can be effectively addressed through a corpus-based approach using manual annotation and machine learning techniques.

This paper defines a generative probabilistic model of parse trees, which we call PCFG-LA. This model is an extension of PCFG in which non-terminal symbols are augmented with latent variables. Fine-grained CFG rules are automatically induced from a parsed corpus by training a PCFG-LA model using an EM-algorithm. Because exact parsing with a PCFG-LA is NP-hard, several approximations are described and empirically compared. In experiments using the Penn WSJ corpus, our automatically trained model gave a performance of 86.6% (F1, sentences ≤ 40 words), which is comparable to that of an unlexicalized PCFG parser created using extensive manual feature selection.

This paper reports the development of log-linear models for the disambiguation in wide-coverage HPSG parsing. The estimation of log-linear models requires high computational cost, especially with wide-coverage grammars. Using techniques to reduce the estimation cost, we trained the models using 20 sections of Penn Tree-bank. A series of experiments empirically evaluated the estimation techniques, and also examined the performance of the disambiguation models on the parsing of real-world sentences.

Motivation: The sheer volume of textually described biomedical knowledge exerts the need for natural language processing (NLP) applications in order to allow flexible and efficient access to relevant information. Specialized semantic networks (such as biomedical ontologies, terminologies or semantic lexicons) can significantly enhance these applications by supplying the necessary terminological information in a machine-readable form. With the explosive growth of bio-literature, new terms (representing newly identified concepts or variations of the existing terms) may not be explicitly described within the network and hence cannot be fully exploited by NLP applications. Linguistic and statistical clues can be used to extract many new terms from free text. The extracted terms still need to be correctly positioned relative to other terms in the network. Classification as a means of semantic typing represents the first step in updating a semantic network with new terms. Results: The MaSTerClass system implements the case-based reasoning methodology for the classification of biomedical terms. Availability: MaSTerClass is available at http://www.cbr-masterclass.org. It is distributed under an open source licence for educational and research purposes. The software requires Java, JWDSP, Ant, MySQL and X-hive to be installed and licences obtained separately where needed. Contact: [email protected] Supplementary information: Available at http://www.cbr-masterclass.org

This paper presents a method to develop a class of variable memory Markov models that have higher memory capacity than traditional (uniform memory) Markov models. The structure of the variable memory models is induced from a manually annotated corpus through a decision tree learning algorithm. A series of comparative experiments show the resulting models outperform uniform memory Markov models in a part-of-speech tagging task.

This paper presents an algorithm to generate possible variants for biomedical terms. The algorithm gives each variant its generation probability representing its plausibility, which is potentially useful for query and dictionary expansions. The probabilistic rules for generating variants are automatically learned from raw texts using an existing abbreviation extraction technique. Our method, therefore, requires no linguistic knowledge or labor-intensive natural language resource. We conducted an experiment using 83,142 MEDLINE abstracts for rule induction and 18,930 abstracts for testing. The results indicate that our method will significantly increase the number of retrieved documents for long biomedical terms.

Abstract We review recent results in literature data mining for biology and discuss the need and the steps for a challenge evaluation for this field. Literature data mining has progressed from simple recognition of terms to extraction of interaction relationships from complex sentences, and has broadened from recognition of protein interactions to a range of problems such as improving homology search, identifying cellular location, and so on. To encourage participation and accelerate progress in this expanding field, we propose creating challenge evaluations, and we describe two specific applications in this context. Contact: [email protected]@[email protected]@[email protected] * To whom correspondence should be addressed.

No abstract available.

In this paper we present part-of-speech taggers based on hidden Markov models, which adopt a less strict Markov assumption to consider rich contexts. In models whose parameters are very specific like lexicalized ones, sparse-data problem is very serious and also conditional probabilities tend to be estimated unreliably. To overcome data-sparseness, a simplified version of the well-known back-off smoothing method is used. To mitigate unreliable estimation problem, our models assume joint independence instead of conditional independence because joint probabilities have the same degree of estimation reliability. In experiments for the Brown corpus, models with rich contexts achieve relatively high accuracy and some models assuming joint independence show better results than the corresponding HMMs.

&% ' )( * '+ -, .! !/ 10 2 3 4 4 5 6 74 8 ) 7 9 !: ; 74 < 64 )0 >= ?4 4 ) ;@ -% 9 ; 7 A% 9 B C % ED F HG -'+ -, 7 / ! ; 74 -F H " 7 JI K ML C / N " 8 A% 9 -% ED O QP R !S 9T VU XW MW ZY [ '[ ]\ _9ba Mc R NT ed gf h d 5i kj lc m nd od p 7q Vr \ _s ;t *u q vt w _x Xy B&x zw _s *u 3\ t e{ |w E^} V{ q & q 9{ | q s * ;{ |t B w x A w E v Ew w E9 Ew 3 B _ 9 5 9& ' _w B ' w E v Ew E B E E E 9 E\ r \ _^ q nq t * * ;{ |{ o 7{ V V t ew E v _w 9 ¡\ _ r ¢ P 9[ ET ¤£ ¥h !P R !S §¦ d g¨ p 7q nr \ s ;t eu q vt ©w _x Xª 5w Eu r &t q s ©} n{ q 9

We propose indices to measure the difficulty of the named entity (NE) task by looking at test corpora, based on expressions inside and outside the NEs. These indices are intended to estimate the difficulty of each task without actually using an NE system and to be unbiased towards a specific system. The values of the indices are compared with the systems' performance in Japanese documents. We also discuss the difference between NE classes with the indices and show useful clues which will make it easier to recognize NEs.

We have witnessed significant progress in NLP applications such as information extraction (IE), summarization, machine translation, cross-lingual information retrieval (CLIR), etc. The progress will be accelerated by advances in speech technology, which not only enables us to interact with systems via speech but also to store and retrieve texts input via speech.

The panel intends to pick up some of the issues discussed in the Summit and discuss them further in the final session from broader perspectives. Since the Summit has not even started yet, I will just enumerate in this paper a list of possible perspectives on MT that I hope are relevant to our discussion.

Experience-Guided Machine Translation (EGMT) seeks to represent the translators’ knowledge of translation as experiences and translates by analogy. The transfer in EGMT finds the experiences most similar to a new text and its parts, segments it into units of translation and translates them by analogy to the experiences and then assembles them into a whole. A research prototype of analogical transfer from Chinese to English is built to prove the viability of the approach in the exploration of new architecture of machine translation. The paper discusses how the experiences are represented and selected with respect to a new text. It describes how units of translation are defined, partial translation is derived and composed into a whole.

This paper outlines the linguistic semantic commitments underlying an application which automatically constructs depictions of verbal spatial descriptions.Our approach draws on the ideational view of linguistic semantics developed by Ronald Langacker in his theory of Cognitive Grammar, and the conceptual representation of physical objects from the two-level semantics of Bierwisch and Lang.In particular the dimensions of the process of conventwnal imagery are used as a metric for the design of our own conceptual representation.

This paper describes the outline of our Japanese to English machine translation system, which is supported by the Agency of Science and Technology of the Japanese Government. Many new methodologies are introduced to obtain high-quality translation results. The analysis is based on case grammar, which is suitable for a word-order-free language such as Japanese. The dictionary is rich enough to handle many specific expressions. It contains not only case frame information, but also semantic information, idiomatic expressions, and many others. In the transfer phase, the system applies many structural transformations, so that the structural difference of the same contents in Japanese and English can be relieved. In the generation phase, many structural transformations are again applied so that the ellipsis problems can be avoided, and that better stylistic expressions can be obtained. The system is running mainly for the abstracts of scientific and technical papers. The evaluation method of the translated results is also discussed, with many example translations.

Due %0 the advance of electronic instrumentation, auwmsted ship increases in number.)

A powerful grammar writing system has been developed. This grammar writing system is called GRADE (GRAmmar DEscriber). GRADE allows a grammar writer to write grammars including analysis, transfer, and generation using the same expression. GRADE has powerful grammar writing facility. GRADE allows a grammar writer to control the process of a machine translation. GRADE also has a function to use grammatical rules written in a word dictionary. GRADE has been used for more than a year as the software of the machine translation project from Japanese into English, which is supported by the Japanese Government and called Mu-project.

Analysis grammar of Japanese in the Mu-project is presented. It is emphasized that rules expressing constraints on single linguistic structures and rules for selecting the most preferable readings are completely different in nature, and that rules for selecting preferale readings should be utilized in analysis grammars of practical MT systems. It is also claimed that procedural control is essential in integrating such rules into a unified grammar. Some sample rules are given to make the points of discussion clear and concrete.