Motivation: Solid-state drives (SSDs) are now ubiquitous in computing systems, yet their internal data layout and operations remain largely invisible and difficult for students to conceptualize. Existing operating systems (OS) teaching materials provide minimal learning support for these abstract processes, which in turn limits student understanding. Method: We designed SSDVis, a web-based interactive SSD visualizer that enables students to simulate file operations and observe SSD internals through coordinated visualizations and step-by-step execution. We evaluated it in a third-year OS course by surveying students (N=154) on their usage and perceptions of SSDVis. Results: 95.4% of survey participants reported using the visualizer extensively, and the overwhelming majority reported improved understanding of SSD data layout, file operations, and garbage collection when using SSDVis. Students valued the interactive features for predicting and verifying behavior, including the step-by-step mode. While usage for understanding wear leveling was lower, this likely reflects the topic's inherent complexity to observe indirectly in elaborate scenarios, rather than usability issues. Implications: Our findings show that a pedagogically designed SSD visualizer has the potential to effectively support the learning of modern OS topics with complex hidden mechanisms.

We develop quantum RNNs with cells based on Parametrised Quantum Circuits (PQCs).PQCs can provide a form of hybrid quantum-classical computation where the input and the output is in the form of classical data.The previous "hidden" state is the quantum state from the previous time-step, and an angle encoding is used to define a (non-linear) mapping from a classical word embedding into the quantum Hilbert space.Measurements of the quantum state provide classical statistics which are used for classification.We report results which are competitive with various RNN baselines on the Rotten Tomatoes dataset, as well as emulator results which demonstrate the feasibility of running such models on quantum hardware.

We apply a generative segmental model of task structure, guided by narration, to action segmentation in video.We focus on unsupervised and weakly-supervised settings where no action labels are known during training.Despite its simplicity, our model performs competitively with previous work on a dataset of naturalistic instructional videos.Our model allows us to vary the sources of supervision used in training, and we find that both task structure and narrative language provide large benefits in segmentation quality.

The task of session search focuses on using interaction data to improve relevance for the user's next query at the session level. In this paper, we formulate session search as a personalization task under the framework of learning to rank. Personalization approaches re-rank results to match a user model. Such user models are usually accumulated over time based on the user's browsing behaviour. We use a pre-computed and transparent set of user models based on concepts from the social science literature. Interaction data are used to map each session to these user models. Novel features are then estimated based on such models as well as sessions' interaction data. Extensive experiments on test collections from the TREC session track show statistically significant improvements over current session search algorithms.

This paper is about learning word representations using grammatical type information. We use the syntactic types of Combinatory Categorial Grammar to develop multilinear representations, i.e. maps with n arguments, for words with different functional types. The multilinear maps of words compose with each other to form sentence representations. We extend the skipgram algorithm from vectors to multi- linear maps to learn these representations and instantiate it on unary and binary maps for transitive verbs. These are evaluated on verb and sentence similarity and disambiguation tasks and a subset of the SICK relatedness dataset. Our model performs better than previous type- driven models and is competitive with state of the art representation learning methods such as BERT and neural sentence encoders.

Recent work has shown how predictive modeling can endow agents with rich knowledge of their surroundings, improving their ability to act in complex environments. We propose question-answering as a general paradigm to decode and understand the representations that such agents develop, applying our method to two recent approaches to predictive modelling - action-conditional CPC (Guo et al., 2018) and SimCore (Gregor et al., 2019). After training agents with these predictive objectives in a visually-rich, 3D environment with an assortment of objects, colors, shapes, and spatial configurations, we probe their internal state representations with a host of synthetic (English) questions, without backpropagating gradients from the question-answering decoder into the agent. The performance of different agents when probed in this way reveals that they learn to encode factual, and seemingly compositional, information about objects, properties and spatial relations from their physical environment. Our approach is intuitive, i.e. humans can easily interpret the responses of the model as opposed to inspecting continuous vectors, and model-agnostic, i.e. applicable to any modeling approach. By revealing the implicit knowledge of objects, quantities, properties and relations acquired by agents as they learn, question-conditional agent probing can stimulate the design and development of stronger predictive learning objectives.

Prior work has shown that, on small amounts of training data, syntactic neural language models learn structurally sensitive generalisations more successfully than sequential language models. However, their computational complexity renders scaling difficult, and it remains an open question whether structural biases are still necessary when sequential models have access to ever larger amounts of training data. To answer this question, we introduce an efficient knowledge distillation (KD) technique that transfers knowledge from a syntactic language model trained on a small corpus to an LSTM language model, hence enabling the LSTM to develop a more structurally sensitive representation of the larger training data it learns from. On targeted syntactic evaluations, we find that, while sequential LSTMs perform much better than previously reported, our proposed technique substantially improves on this baseline, yielding a new state of the art. Our findings and analysis affirm the importance of structural biases, even in models that learn from large amounts of data.

Amandla Mabona, Laura Rimell, Stephen Clark, Andreas Vlachos. 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.

Abstract We present two studies on neural network architectures that learn to represent sentences by composing their words according to automatically induced binary trees, without ever being shown a correct parse tree. We use Tree-Long Short-Term Memories (LSTMs) as our composition function, applied along a tree structure found by a differentiable natural language chart parser. The models simultaneously optimise both the composition function and the parser, thus eliminating the need for externally provided parse trees, which are normally required for Tree-LSTMs. They can therefore be seen as tree-based recurrent neural networks that are unsupervised with respect to the parse trees. Due to being fully differentiable, the models are easily trained with an off-the-shelf gradient descent method and backpropagation. In the first part of this paper, we introduce a model based on the CKY chart parser, and evaluate its downstream performance on a natural language inference task and a reverse dictionary task. Further, we show how its performance can be improved with an attention mechanism which fully exploits the parse chart, by attending over all possible subspans of the sentence. We find that our approach is competitive against similar models of comparable size and outperforms Tree-LSTMs that use trees produced by a parser. Finally, we present an alternative architecture based on a shift-reduce parser. We perform an analysis of the trees induced by both our models, to investigate whether they are consistent with each other and across re-runs, and whether they resemble the trees produced by a standard parser.

Adhiguna Kuncoro, Chris Dyer, John Hale, Dani Yogatama, Stephen Clark, Phil Blunsom. Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2018.

Research in computational semantics is increasingly guided by our understanding of human semantic processing.However, semantic models are typically studied in the context of natural language processing system performance.In this paper, we present a systematic evaluation and comparison of a range of widely-used, stateof-the-art semantic models in their ability to predict patterns of conceptual representation in the human brain.Our results provide new insights both for the design of computational semantic models and for further research in cognitive neuroscience.

Multi-modal semantics, which aims to ground semantic representations in perception, has relied on feature norms or raw image data for perceptual input. In this paper we examine grounding semantic representations in raw auditory data, using standard evaluations for multi-modal semantics. After having shown the quality of such auditorily grounded representations, we show how they can be applied to tasks where auditory perception is relevant, including two unsupervised categorization experiments, and provide further analysis. We find that features transfered from deep neural networks outperform bag of audio words approaches. To our knowledge, this is the first work to construct multi-modal models from a combination of textual information and auditory information extracted from deep neural networks, and the first work to evaluate the performance of tri-modal (textual, visual and auditory) semantic models.

Important advances have recently been made using computational semantic models to decode brain activity patterns associated with concepts; however, this work has almost exclusively focused on concrete nouns. How well these models extend to decoding abstract nouns is largely unknown. We address this question by applying state-of-the-art computational models to decode functional Magnetic Resonance Imaging (fMRI) activity patterns, elicited by participants reading and imagining a diverse set of both concrete and abstract nouns. One of the models we use is linguistic, exploiting the recent word2vec skipgram approach trained on Wikipedia. The second is visually grounded, using deep convolutional neural networks trained on Google Images. Dual coding theory considers concrete concepts to be encoded in the brain both linguistically and visually, and abstract concepts only linguistically. Splitting the fMRI data according to human concreteness ratings, we indeed observe that both models significantly decode the most concrete nouns; however, accuracy is significantly greater using the text-based models for the most abstract nouns. More generally this confirms that current computational models are sufficiently advanced to assist in investigating the representational structure of abstract concepts in the brain.

We address the problem of automatically acquiring knowledge of event sequences from text, with the aim of providing a predictive model for use in narrative generation systems. We present a neural network model that simultaneously learns embeddings for words describing events, a function to compose the embeddings into a representation of the event, and a coherence function to predict the strength of association between two events. We introduce a new development of the narrative cloze evaluation task, better suited to a setting where rich information about events is available. We compare models that learn vector-space representations of the events denoted by verbs in chains centering on a single protagonist. We find that recent work on learning vector-space embeddings to capture word meaning can be effectively applied to this task, including simple incorporation of a verb's arguments in the representation by vector addition. These representations provide a good initialization for learning the richer, compositional model of events with a neural network, vastly outperforming a number of baselines and competitive alternatives.

Multi-modal distributional models learn grounded representations for improved performance in semantics.Deep visual representations, learned using convolutional neural networks, have been shown to achieve particularly high performance.In this study, we systematically compare deep visual representation learning techniques, experimenting with three well-known network architectures.In addition, we explore the various data sources that can be used for retrieving relevant images, showing that images from search engines perform as well as, or better than, those from manually crafted resources such as ImageNet.Furthermore, we explore the optimal number of images and the multi-lingual applicability of multi-modal semantics.We hope that these findings can serve as a guide for future research in the field.

Property norms have the potential to aid a wide range of semantic tasks, provided that they can be obtained for large numbers of concepts. Recent work has focused on text as the main source of information for automatic property extraction. In this paper we examine property norm prediction from visual, rather than textual, data, using cross-modal maps learnt between property norm and visual spaces. We also investigate the importance of having a complete feature norm dataset, for both training and testing. Finally, we evaluate how these datasets and cross-modal maps can be used in an image retrieval task.

We present expected F-measure training for shift-reduce parsing with RNNs, which enables the learning of a global parsing model optimized for sentence-level F1.We apply the model to CCG parsing, where it improves over a strong greedy RNN baseline, by 1.47% F1, yielding state-of-the-art results for shiftreduce CCG parsing.

This article introduces RELPRON, a large data set of subject and object relative clauses, for the evaluation of methods in compositional distributional semantics. RELPRON targets an intermediate level of grammatical complexity between content-word pairs and full sentences. The task involves matching terms, such as “wisdom,” with representative properties, such as “quality that experience teaches.” A unique feature of RELPRON is that it is built from attested properties, but without the need for them to appear in relative clause format in the source corpus. The article also presents some initial experiments on RELPRON, using a variety of composition methods including simple baselines, arithmetic operators on vectors, and finally, more complex methods in which argument-taking words are represented as tensors. The latter methods are based on the Categorial framework, which is described in detail. The results show that vector addition is difficult to beat—in line with the existing literature—but that an implementation of the Categorial framework based on the Practical Lexical Function model is able to match the performance of vector addition. The article finishes with an in-depth analysis of RELPRON, showing how results vary across subject and object relative clauses, across different head nouns, and how the methods perform on the subtasks necessary for capturing relative clause semantics, as well as providing a qualitative analysis highlighting some of the more common errors. Our hope is that the competitive results presented here, in which the best systems are on average ranking one out of every two properties correctly for a given term, will inspire new approaches to the RELPRON ranking task and other tasks based on linguistically interesting constructions.

Within the categorical compositional distributional model of meaning, we provide semantic interpretations for the subject and object roles of the possessive relative pronoun ‘whose’. This is done in terms of Frobenius algebras over compact closed categories. These algebras and their diagrammatic language expose how meanings of words in relative clauses interact with each other. We show how our interpretation is related to Montague-style semantics and provide a truth-theoretic interpretation. We also show how vector spaces provide a concrete interpretation and provide preliminary corpus-based experimental evidence. In a prequel to this article, we used similar methods and dealt with the case of subject and object relative pronouns.

We present a compositional distributional semantic model which is an implementation of the tensor-based framework of Coecke et al. (2011). It is an extended skipgram model (Mikolov et al., 2013) which we apply to adjective-noun combinations, learning nouns as vectors and adjectives as matrices. We also propose a novel measure of adjective similarity, and show that adjective matrix representations lead to improved performance in adjective and adjective-noun similarity tasks, as well as in the detection of semantically anomalous adjective-noun pairs.

Multi-modal semantics has relied on feature norms or raw image data for perceptual input.In this paper we examine grounding semantic representations in raw auditory data, using standard evaluations for multi-modal semantics, including measuring conceptual similarity and relatedness.We also evaluate cross-modal mappings, through a zero-shot learning task mapping between linguistic and auditory modalities.In addition, we evaluate multimodal representations on an unsupervised musical instrument clustering task.To our knowledge, this is the first work to combine linguistic and auditory information into multi-modal representations.

We demonstrate the advantage of specializing semantic word embeddings for either similarity or relatedness.We compare two variants of retrofitting and a joint-learning approach, and find that all three yield specialized semantic spaces that capture human intuitions regarding similarity and relatedness better than unspecialized spaces.We also show that using specialized spaces in NLP tasks and applications leads to clear improvements, for document classification and synonym selection, which rely on either similarity or relatedness but not both.

This paper is concerned with the task of bilingual lexicon induction using imagebased features.By applying features from a convolutional neural network (CNN), we obtain state-of-the-art performance on a standard dataset, obtaining a 79% relative improvement over previous work which uses bags of visual words based on SIFT features.The CNN image-based approach is also compared with state-of-the-art linguistic approaches to bilingual lexicon induction, even outperforming these for one of three language pairs on another standard dataset.Furthermore, we shed new light on the type of visual similarity metric to use for genuine similarity versus relatedness tasks, and experiment with using multiple layers from the same network in an attempt to improve performance.

Word ordering is a fundamental problem in text generation. In this article, we study word ordering using a syntax-based approach and a discriminative model. Two grammar formalisms are considered: Combinatory Categorial Grammar (CCG) and dependency grammar. Given the search for a likely string and syntactic analysis, the search space is massive, making discriminative training challenging. We develop a learning-guided search framework, based on best-first search, and investigate several alternative training algorithms. The framework we present is flexible in that it allows constraints to be imposed on output word orders. To demonstrate this flexibility, a variety of input conditions are considered. First, we investigate a “pure” word-ordering task in which the input is a multi-set of words, and the task is to order them into a grammatical and fluent sentence. This task has been tackled previously, and we report improved performance over existing systems on a standard Wall Street Journal test set. Second, we tackle the same reordering problem, but with a variety of input conditions, from the bare case with no dependencies or POS tags specified, to the extreme case where all POS tags and unordered, unlabeled dependencies are provided as input (and various conditions in between). When applied to the NLG 2011 shared task, our system gives competitive results compared with the best-performing systems, which provide a further demonstration of the practical utility of our system.

This paper presents the first dependency model for a shift-reduce CCG parser. Mod-elling dependencies is desirable for a num-ber of reasons, including handling the “spurious ” ambiguity of CCG; fitting well with the theory of CCG; and optimizing for structures which are evaluated at test time. We develop a novel training tech-nique using a dependency oracle, in which all derivations are hidden. A challenge arises from the fact that the oracle needs to keep track of exponentially many gold-standard derivations, which is solved by integrating a packed parse forest with the beam-search decoder. Standard CCGBank tests show the model achieves up to 1.05 labeled F-score improvements over three existing, competitive CCG parsing models.

Distributional semantic models (DSMs) have been effective at representing semantics at the word level, and research has recently moved on to building distributional representations for larger segments of text.In this paper, we introduce novel ways of applying context selection and normalisation to vary model sparsity and the range of values of the DSM vectors.We show how these methods enhance the quality of the vectors and thus result in improved low dimensional and composed representations.We demonstrate these effects on standard word and phrase datasets, and on a new definition retrieval task and dataset.

Compositional distributional semantics is a subfield of Computational Linguistics which investigates methods for representing the meanings of phrases and sentences.In this paper, we explore implementations of a framework based on Combinatory Categorial Grammar (CCG), in which words with certain grammatical types have meanings represented by multilinear maps (i.e.multi-dimensional arrays, or tensors).An obstacle to full implementation of the framework is the size of these tensors.We examine the performance of lower dimensional approximations of transitive verb tensors on a sentence plausibility/selectional preference task.We find that the matrices perform as well as, and sometimes even better than, full tensors, allowing a reduction in the number of parameters needed to model the framework.

Linguistic steganography is concerned with hiding information in natural language text. One of the major transformations used in linguistic steganography is synonym substitution. However, few existing studies have studied the practical application of this approach. In this article we propose two improvements to the use of synonym substitution for encoding hidden bits of information. First, we use the Google n-gram corpus for checking the applicability of a synonym in context, and we evaluate this method using data from the SemEval lexical substitution task and human annotated data. Second, we address the problem that arises from words with more than one sense, which creates a potential ambiguity in terms of which bits are represented by a particular word. We develop a novel method in which words are the vertices in a graph, synonyms are linked by edges, and the bits assigned to a word are determined by a vertex coding algorithm. This method ensures that each word represents a unique sequence of bits, without cutting out large numbers of synonyms, and thus maintains a reasonable embedding capacity.

Semantic parsing is the task of translating natural language utterances into a machine-interpretable meaning representation. Most approaches to this task have been evaluated on a small number of existing corpora which assume that all utterances must be interpreted according to a database and typically ignore context. In this paper we present a new, publicly available corpus for context-dependent semantic parsing. The MRL used for the annotation was designed to support a portable, interactive tourist information system. We develop a semantic parser for this corpus by adapting the imitation learning algorithm DAgger without requiring alignment information during training. DAgger improves upon independently trained classifiers by 9.0 and 4.8 points in F-score on the development and test sets respectively.

We present a novel unsupervised approach to detecting the compositionality of multi-word expressions.We compute the compositionality of a phrase through substituting the constituent words with their "neighbours" in a semantic vector space and averaging over the distance between the original phrase and the substituted neighbour phrases.Several methods of obtaining neighbours are presented.The results are compared to existing supervised results and achieve state-of-the-art performance on a verb-object dataset of human compositionality ratings.

This paper develops a compositional vector-based semantics of subject and object relative pronouns within a categorical framework. Frobenius algebras are used to formalise the operations required to model the semantics of relative pronouns, including passing information between the relative clause and the modified noun phrase, as well as copying, combining, and discarding parts of the relative clause. We develop two instantiations of the abstract semantics, one based on a truth-theoretic approach and one based on corpus statistics.

We study a range of syntactic processing tasks using a general statistical framework that consists of a global linear model, trained by the generalized perceptron together with a generic beam-search decoder. We apply the framework to word segmentation, joint segmentation and POS-tagging, dependency parsing, and phrase-structure parsing. Both components of the framework are conceptually and computationally very simple. The beam-search decoder only requires the syntactic processing task to be broken into a sequence of decisions, such that, at each stage in the process, the decoder is able to consider the top-n candidates and generate all possibilities for the next stage. Once the decoder has been defined, it is applied to the training data, using trivial updates according to the generalized perceptron to induce a model. This simple framework performs surprisingly well, giving accuracy results competitive with the state-of-the-art on all the tasks we consider. The computational simplicity of the decoder and training algorithm leads to significantly higher test speeds and lower training times than their main alternatives, including log-linear and large-margin training algorithms and dynamic-programming for decoding. Moreover, the framework offers the freedom to define arbitrary features which can make alternative training and decoding algorithms prohibitively slow. We discuss how the general framework is applied to each of the problems studied in this article, making comparisons with alternative learning and decoding algorithms. We also show how the comparability of candidates considered by the beam is an important factor in the performance. We argue that the conceptual and computational simplicity of the framework, together with its language-independent nature, make it a competitive choice for a range of syntactic processing tasks and one that should be considered for comparison by developers of alternative approaches.

This article describes a number of log-linear parsing models for an automatically extracted lexicalized grammar. The models are “full” parsing models in the sense that probabilities are defined for complete parses, rather than for independent events derived by decomposing the parse tree. Discriminative training is used to estimate the models, which requires incorrect parses for each sentence in the training data as well as the correct parse. The lexicalized grammar formalism used is Combinatory Categorial Grammar (CCG), and the grammar is automatically extracted from CCGbank, a CCG version of the Penn Treebank. The combination of discriminative training and an automatically extracted grammar leads to a significant memory requirement (up to 25 GB), which is satisfied using a parallel implementation of the BFGS optimization algorithm running on a Beowulf cluster. Dynamic programming over a packed chart, in combination with the parallel implementation, allows us to solve one of the largest-scale estimation problems in the statistical parsing literature in under three hours. A key component of the parsing system, for both training and testing, is a Maximum Entropy supertagger which assigns CCG lexical categories to words in a sentence. The supertagger makes the discriminative training feasible, and also leads to a highly efficient parser. Surprisingly, given CCG's “spurious ambiguity,” the parsing speeds are significantly higher than those reported for comparable parsers in the literature. We also extend the existing parsing techniques for CCG by developing a new model and efficient parsing algorithm which exploits all derivations, including CCG's nonstandard derivations. This model and parsing algorithm, when combined with normal-form constraints, give state-of-the-art accuracy for the recovery of predicate-argument dependencies from CCGbank. The parser is also evaluated on DepBank and compared against the RASP parser, outperforming RASP overall and on the majority of relation types. The evaluation on DepBank raises a number of issues regarding parser evaluation. This article provides a comprehensive blueprint for building a wide-coverage CCG parser. We demonstrate that both accurate and highly efficient parsing is possible with CCG.

With performance above 97% accuracy for newspaper text, part of speech (POS) tagging might be considered a solved problem. Previous studies have shown that allowing the parser to resolve POS tag ambiguity does not improve performance. However, for grammar formalisms which use more fine-grained grammatical categories, for example TAG and CCG, tagging accuracy is much lower. In fact, for these formalisms, premature ambiguity resolution makes parsing infeasible.We describe a multi-tagging approach which maintains a suitable level of lexical category ambiguity for accurate and efficient CCG parsing. We extend this multi-tagging approach to the POS level to overcome errors introduced by automatically assigned POS tags. Although POS tagging accuracy seems high, maintaining some POS tag ambiguity in the language processing pipeline results in more accurate CCG supertagging.

This paper describes and evaluates log-linear parsing models for Combinatory Categorial Grammar (CCG). A parallel implementation of the L-BFGS optimisation algorithm is described, which runs on a Beowulf cluster allowing the complete Penn Treebank to be used for estimation. We also develop a new efficient parsing algorithm for CCG which maximises expected recall of dependencies. We compare models which use all CCG derivations, including non-standard derivations, with normal-form models. The performances of the two models are comparable and the results are competitive with existing wide-coverage CCG parsers.

This paper describes a wide-coverage statistical parser that uses Combinatory Categorial Grammar (CCG) to derive dependency structures. The parser differs from most existing wide-coverage treebank parsers in capturing the long-range dependencies inherent in constructions such as coordination, extraction, raising and control, as well as the standard local predicate-argument dependencies. A set of dependency structures used for training and testing the parser is obtained from a treebank of CCG normal-form derivations, which have been derived (semi-) automatically from the Penn Treebank. The parser correctly recovers over 80% of labelled dependencies, and around 90% of unlabelled dependencies.

This article concerns the estimation of a particular kind of probability, namely, the probability of a noun sense appearing as a particular argument of a predicate. In order to overcome the accompanying sparse-data problem, the proposal here is to define the probabilities in terms of senses from a semantic hierarchy and exploit the fact that the senses can be grouped into classes consisting of semantically similar senses. There is a particular focus on the problem of how to determine a suitable class for a given sense, or, alternatively, how to determine a suitable level of generalization in the hierarchy. A procedure is developed that uses a chi-square test to determine a suitable level of generalization. In order to test the performance of the estimation method, a pseudo-disambiguation task is used, together with two alternative estimation methods. Each method uses a different generalization procedure; the first alternative uses the minimum description length principle, and the second uses Resnik's measure of selectional preference. In addition, the performance of our method is investigated using both the standard Pearson chi-square statistic and the log-likelihood chi-square statistic.