Technology plays a very important role in education and Intelligent Computer-Assisted Language Learning (iCALL) has emerged as a complementary or even alternative method to the conventional language teaching practices. The automatic generation (and correction) of language exercises based on real texts extracted from corpora constitutes a non-trivial challenge to iCALL tutorial systems, and may involve the use of sophisticated Natural Language Processing tools and large-scale linguistic resources. This paper presents the main issues related to the automatic generation of exercises on the Passive transformation, a commonly occurring type of exercises in language textbooks, but also a very complex topic of Portuguese grammar. The paper describes the methods used to produce a large batch of passive-active sentence pairs, where the active sentence was automatically generated from naturally occurring passive sentences, taken from a large-sized, publicly available, corpus. Sentence pairs are ranked by difficulty level. A sample of randomly selected sentence pairs (40 from difficult level, 100 from medium, and 100 from easy level) was manually evaluated by an expert. Results are presented and error analysis is performed. The sentence pairs can be used as prime and correct answer for iCALL systems.

Sharing data in the form of text is important for a wide range of activities but it also raises a concern about privacy when sharing data that could be sensitive. Automated text anonymization is a solution for removing all the sensitive information from documents. However, this is a challenging task due to the unstructured form of textual data and the ambiguity of natural language. In this work, we present our implementation of an automated anonymization system, built in a modular structure, for documents written in Portuguese. Four different methods of anonymization are evaluated and compared. Two methods replace the sensitive information by artificial labels: suppression and tagging. The other two methods replace the information by textual expressions: random substitution and generalization. Evaluation showed that the use of the tagging and the generalization methods facilitates the reading of an anonymized text while preventing some semantic drifts caused by the remotion of the original information.

This paper focuses on the tasks of recovering capitalization and punctuation marks from texts without that information, such as spoken transcripts, produced by automatic speech recognition systems. These two practical rich transcription tasks were performed using the same discriminative approach, based on maximum entropy, suitable for on-the-fly usage. Reported experiments were conducted both over Portuguese and English broadcast news data. Both force aligned and automatic transcripts were used, allowing to measure the impact of the speech recognition errors. Capitalized words and named entities are intrinsically related, and are influenced by time variation effects. For that reason, the so-called language dynamics have been addressed for the capitalization task. Language adaptation results indicate, for both languages, that the capitalization performance is affected by the temporal distance between the training and testing data. In what regards the punctuation task, this paper covers the three most frequent punctuation marks: full stop, comma, and question marks. Different methods were explored for improving the baseline results for full stop and comma. The first uses punctuation information extracted from large written corpora. The second applies different levels of linguistic structure, including lexical, prosodic, and speaker related features. The comma detection improved significantly in the first method, thus indicating that it depends more on lexical features. The second method provided even better results, for both languages and both punctuation marks, best results being achieved mainly for full stop. As for question marks, there is a small gain, but differences are not very significant, due to the relatively small number of question marks in the corpora.

The goal of this work is the automatic selection of materials for a listening comprehension game. We would like to select automatically transcribed sentences from recent broadcast news corpora, in order to gather material for the games with little human effort. The recognized words are used as the ground solution of the exercises, thus sentences with misrecognitions need to be filtered out. Our experiments confirmed the feasibility of the filter chain that automatically selects sentences, although harder confidence thresholds may be needed. Together with the correct words, wrong candidates, namely distractors, are also needed to build the exercises. Two techniques of distractor generation are presented, either based on the confusion networks produced by the recognizer, or on phonetic distances. The experiments confirmed the complementarity of both approaches. Index Terms: CALL, Listening Comprehension, European Portuguese, ASR, distractors

This paper describes our recent work on extending the punctuation module of automatic subtitles for Portuguese Broadcast News. The main improvement was achieved by the use of prosodic information. This enabled the extension of the previous module which covered only full stops and commas, to cover question marks as well. The approach uses lexical, acoustic and prosodic information. Our results show that the latter is relevant for all types of punctuation. An analysis of the results also shows what type of interrogative is better dealt with by our method, taking into account the specificities of Portuguese. This may lead to different results for different types of corpora, depending on the types of interrogatives that are more frequent.

This paper describes the integration of multimedia documents in the Portuguese version of REAP, a tutoring system for vocabulary learning. The documents result from the pipeline processing of Broadcast News videos that automatically segments the audio files, transcribes them, adds punctuation and capitalization, and breaks them into stories classified by topics. The integration of these materials in REAP was done in a way that tries to decrease the impact of potential errors of the automatic chain in the learning process.

This paper describes and evaluates a language independent approach for automatically enriching the speech recognition output with punctuation marks and capitalization information. The two tasks are treated as two classification problems, using a maximum entropy modeling approach, which achieves results within state-of-the-art. The language independence of the approach is attested with experiments conducted on Portuguese, Spanish and English broadcast news corpora. This paper provides the first comparative study between the three languages, concerning these tasks.

This paper investigates the impact of language dynamics on the capitalization of transcriptions of broadcast news. Most of the capitalization information is provided by a large newspaper corpus. Three different speech corpora subsets, from different time periods, are used for evaluation, assessing the importance of available training data in nearby time periods. Results are provided both for manual and automatic transcriptions, showing also the impact of the recognition errors in the capitalization task. Our approach is based on maximum entropy models, uses unlimited vocabulary, and is suitable for language adaptation. The language model for a given language period is produced by retraining a previous language model with data from that time period. The language model produced with this approach can be sorted and then pruned, in order to reduce computational resources, without much impact in the final results.

The application of speech recognition to live subtitling of Broadcast News has motivated the adaptation of the lexical and language models of the recognizer on a daily basis with text material retrieved from online newspapers. This paper studies the impact of this adaptation on two of the blocks following the speech recognition module: capitalization and topic indexation. We describe and evaluate different adaptation approaches that try to explore the language dynamics.

This paper presents a system that helps you cook a recipe through a spoken dialogue tutoring session. We report our experience while creating the first version of a tutoring dialogue system that helps the user cook a selected dish. Having a working framework to support us with the creation of the cooking assistant, the main challenge we faced was the change of paradigm: instead of the system being driven by the user, the user is instructed by the system. The result is a system capable of dictating generic contents to the user. On top of it, the system can be used in several domains where the goal is not the replacement of the user but providing some assistance while (s)he performs some procedural task.

This paper shows results of recovering punctuation over speech transcriptions for a Portuguese broadcast news corpus. The approach is based on maximum entropy models and uses word, part-of-speech, time and speaker information. The contribution of each type of feature is analyzed individually. Separate results for each focus condition are given, making it possible to analyze the differences of performance between planned and spontaneous speech. Index Terms: rich transcription, punctuation recovery, sentence boundary detection, maximum entropy.

In the near future people will be surrounded by intelligent devices embedded in everyday objects where the knowledge and understanding of device attributes and capabilities will be a key enabler. This paper describes the current state of our research in designing distributed knowledge based devices as a solution to adapt spoken dialogue systems within ambient intelligence. In this context a spoken dialogue system is a computational entity that allows universal access to ambient intelligence for anyone, anywhere, at anytime to use any device through any media. Our aim is to build knowledge-based devices to enable dynamic adaptation of the components integrated in the dialogue system architecture. An example focusing household appliances is depicted. 1.

In this paper we describe a multi-purpose Spoken Dialogue System platform associated with two distinct applications as an home intelligent environment and remote access to information databases. These applications differ substantially on contents and possible uses but gives us the chance to develop a platform where we were able to represent diverse services to be accessible by a spoken interface. The implemented voice input/output possibilities and the service independence level opens a wide range of possibilities for the development of new applications using the current components of our Spoken Dialogue System.

We present an evolution of SNePS, the SnePSR (from SNePS with resources) knowledge representation/reasoner system. SnePSR is an intelligent resource-bounded reasoner that allows several resource spending strategies. Since no a priori commitments are made about the way resources are spent, the process of consuming resources can be used to model non-omniscient, non-exhaustive reasoners. SnePSR combines the introduction of resources with the capability to produce conditional answers, which explicitly reveals the impediments that are responsible for the absence of a definite answer. SnePSR avoids two problems that most of the programs, trying to behave intelligently, suffer from: (1) never take into account the fact that reasoning resources are limited; (2) remain silent whenever a definite answer cannot be produced. After briefly presenting how the characteristics that distinguish SNePSR have been incorporated into SNePS we present some case studies of interactions with SNePSR demonstrating some of the system's features.