SwePub - sökning: WFRF:(Bernardy Jean Philippe 1...

Numrering	Referens	Omslagsbild	Hitta
1.	Berman, Alexander, 1979, et al. (författare) Explaining predictions with enthymematic counterfactuals 2022 Ingår i: Proceedings of the 1st Workshop on Bias, Ethical AI, Explainability and the role of Logic and Logic Programming, BEWARE-22. - : CEUR Workshop Proceedings. - 1613-0073. Konferensbidrag (refereegranskat)abstract When people are subject to high-stakes decisions informed by computer models, they have a legitimate interest in understanding the basis for the model’s judgements and whether actions can be taken to turn a dispreferred decision into a preferred one. For example, if an application for a loan is denied by the model, the applicant has an interest in understanding the conditions that would yield an approval. In this paper, we argue that these kinds of counterfactual (or contrastive) explanations rest on domain-specific and commonsensical principles that can be negotiated, and sketch a method for incorporating such principles in an explanatory dialogue system using enthymematic reasoning.
2.	Bernardy, Jean-Philippe, 1978, et al. (författare) A comparison of C++ concepts and Haskell type classes 2008 Ingår i: ACM SIGPLAN Workshop on Generic Programming 2008. Victoria, Canada, 20th September 2008. - 9781605580609 ; , s. 37-48 Konferensbidrag (refereegranskat)abstract Earlier studies have introduced a list of high-level evaluation criteria to assess how well a language supports generic programming. Since each language that meets all criteria is considered generic, those criteria are not fine-grained enough to differentiate between two languages for generic programming.We refine these criteria into a taxonomy that capturesdifferences between type classes in Haskell and concepts inC++, and discuss which differences are incidental and whichones are due to other language features. The taxonomy allows for an improved understanding of language support for generic programming, and the comparison is useful for the ongoing discussions among language designers and users of both languages.
3.	Bernardy, Jean-Philippe, 1978, et al. (författare) Generic programming with C++ concepts and Haskell type classes—a comparison 2010 Ingår i: Journal of Functional Programming. - 1469-7653 .- 0956-7968. ; 20:3-4, s. 271-302 Tidskriftsartikel (refereegranskat)abstract Earlier studies have introduced a list of high-level evaluation criteria to assess how well a language supports generic programming. Languages that meet all criteria include Haskell, because of its type classes, and C++ with the concept feature. We refine these criteria into a taxonomy that captures commonalities and differences between type classes in Haskell and concepts in C++, and discuss which differences are incidental and which ones are due to other language features. The taxonomy allows for an improved understanding of language support for generic programming, and the comparison is useful for the ongoing discussions among language designers and users of both languages.
4.	Bernardy, Jean-Philippe, 1978, et al. (författare) Rational Speech Act models are utterance-independent updates of world priors 2022 Ingår i: Proceedings of SemDial 2022 (DubDial): The 26th Workshop on the Semantics and Pragmatics of Dialogue, August, 22-24, 2022, Dublin / Eleni Gregoromichelaki, Julian Hough, John D. Kelleher (Editors). - 2308-2275. Konferensbidrag (refereegranskat)abstract A popular framework for modelling pragmatic effects is the “rational speech act” (RSA) model introduced by Frank and Goodman (2012). The idea behind RSA is that, to interpret an utter- ance, a rational (pragmatic) listener reasons about a speaker who chooses their utterance by reasoning about the listener, using a literal se- mantic model. In the present work, we take the RSA model at face value, but we reformulate it in information-theoretic terms. We find that the pragmatic listener model can be reconceived as an update of the prior over worlds that can be provided independently of the speaker’s actual utterance. This update consists in a preference for world states which are the most specific to a given utterance in the set of possible ut- terances given by the pragmatic context. Our reformulation allows us to deduce general prop- erties of pragmatic reasoning problems. As an example, we show that RSA does not predict certain quantity implicatures in the presence of bell-curve priors.
5.	Maraev, Vladislav, 1986, et al. (författare) Non-humorous use of laughter in spoken dialogue systems 2021 Ingår i: Proceedings of the 7th Linguistic and Cognitive Approaches To Dialog Agents Workshop, LaCATODA 2021, Montreal, Canada, August 21, 2021 (online) / Rafal Rzepka, Jordi Vallverdú, Andre Wlodarczyk, Michal Ptaszynski and Pawel Dybala (Eds.). - : CEUR-WS. - 1613-0073. Konferensbidrag (refereegranskat)abstract In this paper we argue that laughter, an ambiguous yet ubiquitous signal in everyday interactions, can act as an important feature for task-oriented dialogue systems. We show which components of a dialogue system should be affected and modified, and more specifically how particular types of laughter can be accounted for in a dialogue manager as instances of short answers, feedbacks and vocalisations accompanying them.
6.	Maraev, Vladislav, 1986, et al. (författare) Predicting laughter relevance spaces in dialogue 2021 Ingår i: Marchi E., Siniscalchi S.M., Cumani S., Salerno V.M., Li H. (eds) Increasing Naturalness and Flexibility in Spoken Dialogue Interaction. Lecture Notes in Electrical Engineering, vol 714.. - Singapore : Springer. - 1876-1100. - 9789811593222 ; , s. 41-51 Bokkapitel (refereegranskat)
7.	Maraev, Vladislav, 1986, et al. (författare) Why Should I Turn Left? : Towards Active Explainability for Spoken Dialogue Systems. 2021 Ingår i: Proceedings of the Reasoning and Interaction Conference (ReInAct 2021). - : Association for Computational Linguistics. Konferensbidrag (refereegranskat)abstract In this paper we argue that to make dialogue systems able to actively explain their decisions they can make use of enthymematic reasoning. We motivate why this is an appropriate strategy and integrate it within our own proof-theoretic dialogue manager framework based on linear logic. In particular, this enables a dialogue system to provide reasonable answers to why-questions that query information previously given by the system.
8.	Abel, Andreas, 1974, et al. (författare) A Unified View of Modalities in Type Systems 2020 Ingår i: Proceedings of the ACM on Programming Languages. - : Association for Computing Machinery (ACM). - 2475-1421. ; 4:ICFP Tidskriftsartikel (refereegranskat)abstract We propose to unify the treatment of a broad range of modalities in typed lambda calculi. We do so by defining a generic structure of modalities, and show that this structure arises naturally from the structure of intuitionistic logic, and as such finds instances in a wide range of type systems previously described in literature. Despite this generality, this structure has a rich metatheory, which we expose.
9.	Adouane, Wafia, 1985, et al. (författare) A Comparison of Character Neural Language Model and Bootstrapping for Language Identification in Multilingual Noisy Texts 2018 Ingår i: Proceedings of the Second Workshop on Subword and Character Level Models in NLP (SCLeM), June 6, 2018 New Orleans, Louisiana. - New Orleans, Louisiana USA. - 9781948087186 Konferensbidrag (refereegranskat)abstract This paper seeks to examine the effect of including background knowledge in the form of character pre-trained neural language model (LM), and data bootstrapping to overcome the problem of unbalanced limited resources. As a test, we explore the task of language identification in mixed-language short non-edited texts with an under-resourced language, namely the case of Algerian Arabic for which both labelled and unlabelled data are limited. We compare the performance of two traditional machine learning methods and a deep neural networks (DNNs) model. The results show that overall DNNs perform better on labelled data for the majority categories and struggle with the minority ones. While the effect of the untokenised and unlabelled data encoded as LM differs for each category, bootstrapping, however, improves the performance of all systems and all categories. These methods are language independent and could be generalised to other under-resourced languages for which a small labelled data and a larger unlabelled data are available.
10.	Adouane, Wafia, 1985, et al. (författare) Identifying Sentiments in Algerian Code-switched User-generated Comments 2020 Ingår i: Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), Marseille, 11–16 May 2020. - Paris : The European Language Resources Association. - 9791095546344 Konferensbidrag (refereegranskat)abstract We present in this paper our work on Algerian language, an under-resourced North African colloquial Arabic variety, for which we built a comparably large corpus of more than 36,000 code-switched user-generated comments annotated for sentiments. We opted for this data domain because Algerian is a colloquial language with no existing freely available corpora. Moreover, we compiled sentiment lexicons of positive and negative unigrams and bigrams reflecting the code-switches present in the language. We compare the performance of four models on the task of identifying sentiments, and the results indicate that a CNN model trained end-to-end fits better our unedited code-switched and unbalanced data across the predefined sentiment classes. Additionally, injecting the lexicons as background knowledge to the model boosts its performance on the minority class with a gain of 10.54 points on the F-score. The results of our experiments can be used as a baseline for future research for Algerian sentiment analysis.
11.	Adouane, Wafia, 1985, et al. (författare) Improving Neural Network Performance by Injecting Background Knowledge: Detecting Code-switching and Borrowing in Algerian texts 2018 Ingår i: Proceedings of the Third Workshop on Computational Approaches to Linguistic Code-switching, Melbourne, Australia, July 19, 2018. - Melbourne, Australia : Association for Computational Linguistics. - 9781948087452 Konferensbidrag (refereegranskat)abstract We explore the effect of injecting back- ground knowledge to different deep neural network (DNN) configurations in order to mitigate the problem of the scarcity of annotated data when applying these models on datasets of low-resourced languages. The background knowledge is encoded in the form of lexicons and pre-trained sub-word embeddings. The DNN models are evaluated on the task of detecting code-switching and borrowing points in non-standardised user-generated Algerian texts. Overall results show that DNNs benefit from adding background knowledge. However, the gain varies between models and categories. The proposed DNN architectures are generic and could be applied to other low-resourced languages.
12.	Adouane, Wafia, 1985, et al. (författare) Neural Models for Detecting Binary Semantic Textual Similarity for Algerian and MSA 2019 Ingår i: Proceedings of the Fourth Arabic Natural Language Processing Workshop, WANLP 2019, Jul 28-Aug 2, Florence, Italy. pp. 78-87. - Florence, Italy : Association for Computational Linguistics. - 9781950737321 Konferensbidrag (refereegranskat)abstract We explore the extent to which neural networks can learn to identify semantically equivalent sentences from a small variable dataset using an end-to-end training. We collect a new noisy non-standardised user-generated Algerian (ALG) dataset and also translate it to Modern Standard Arabic (MSA) which serves as its regularised counterpart. We compare the performance of various models on both datasets and report the best performing configurations. The results show that relatively simple models composed of 2 LSTM layers outperform by far other more sophisticated attention-based architectures, for both ALG and MSA datasets.
13.	Adouane, Wafia, 1985, et al. (författare) Normalising Non-standardised Orthography in Algerian Code-switched User-generated Data 2019 Ingår i: The 5th Workshop on Noisy User-generated Text (W-NUT), November 4, 2019, Hong Kong / Wei Xu, Alan Ritter, Tim Baldwin, Afshin Rahimi (Editors). - Stroudsburg, PA : Association for Computational Linguistics. - 9781950737840 Konferensbidrag (refereegranskat)abstract We work with Algerian, an under-resourced non-standardised Arabic variety, for which we compile a new parallel corpus consist- ing of user-generated textual data matched with normalised and corrected human annota- tions following data-driven and our linguisti- cally motivated standard. We use an end-to- end deep neural model designed to deal with context-dependent spelling correction and nor- malisation. Results indicate that a model with two CNN sub-network encoders and an LSTM decoder performs the best, and that word context matters. Additionally, pre- processing data token-by-token with an edit- distance based aligner significantly improves the performance. We get promising results for the spelling correction and normalisation, as a pre-processing step for downstream tasks, on detecting binary Semantic Textual Similarity.
14.	Adouane, Wafia, 1985, et al. (författare) When is Multi-task Learning Beneficial for Low-Resource Noisy Code-switched User-generated Algerian Texts? 2020 Ingår i: Fourth Workshop on Computational Approaches to Linguistic Code-Switching. Language Resources and Evaluation Conference (LREC 2020), Marseille, 11–16 May 202. - Paris : European Language Resources Association. - 9791095546665 Konferensbidrag (refereegranskat)abstract We investigate when is it beneficial to simultaneously learn representations for several tasks, in low-resource settings. For this, we work with noisy user-generated texts in Algerian, a low-resource non-standardised Arabic variety. That is, to mitigate the problem of the data scarcity, we experiment with jointly learning progressively 4 tasks, namely code-switch detection, named entity recognition, spell normalisation and correction, and identifying users’ sentiments. The selection of these tasks is motivated by the lack of labelled data for automatic morpho-syntactic or semantic sequence-tagging tasks for Algerian, in contrast to the case of much multi-task learning for NLP. Our empirical results show that multi-task learning is beneficial for some tasks in particular settings, and that the effect of each task on another, the order of the tasks, and the size of the training data of the task with more data do matter. Moreover, the data augmentation that we performed with no external resources has been shown to be beneficial for certain tasks.
15.	Algebraic Structures in Natural Language 2022 Samlingsverk (redaktörskap) (övrigt vetenskapligt/konstnärligt)abstract The rapid progress of deep learning in artificial intelligence, particularly in natural language processing, has raised serious questions about the role of classical symbolic algebraic systems in the representation and acquisition of linguistic knowledge. This volume brings together chapters by leading researchers in computational linguistics, cognitive science, and natural language processing on this set of issues. It offers a variety of views, and it presents leading edge work on the main topic.
16.	Algehed, Maximilian, 1995, et al. (författare) Dynamic IFC Theorems for Free! 2021 Ingår i: 2021 IEEE 34TH COMPUTER SECURITY FOUNDATIONS SYMPOSIUM (CSF 2021). - : IEEE. - 2374-8303 .- 1940-1434. - 9781728176079 ; , s. 65-78 Konferensbidrag (refereegranskat)abstract We show that noninterference and transparency, the key soundness theorems for dynamic IFC libraries, can be obtained "for free", as direct consequences of the more general parametricity theorem of type abstraction. This allows us to give very short soundness proofs for dynamic IFC libraries such as faceted values and LIO. Our proofs stay short even when fully mechanized for Agda implementations of the libraries in terms of type abstraction.
17.	Algehed, Maximilian, 1995, et al. (författare) Simple Noninterference from Parametricity 2019 Ingår i: Proceedings of the ACM on Programming Languages. - : Association for Computing Machinery (ACM). - 2475-1421. ; 3 Tidskriftsartikel (refereegranskat)abstract In this paper we revisit the connection between parametricity and noninterference. Our primary contribution is a proof of noninterference for a polyvariant variation of the Dependency Core Calculus of in the Calculus of Constructions. The proof is modular: it leverages parametricity for the Calculus of Constructions and the encoding of data abstraction using existential types. This perspective gives rise to simple and understandable proofs of noninterference from parametricity. All our contributions have been mechanised in the Agda proof assistant.
18.	Amanaki, Erini, et al. (författare) Fine-grained Entailment: Resources for Greek NLI and Precise Entailment 2022 Ingår i: Proceedings of the Workshop on Dataset Creation for Lower-Resourced Languages within the 13th Language Resources and Evaluation Conference. - Marseille, France : European Language Resources Association (ELRA). - 9782493814067 Konferensbidrag (refereegranskat)abstract In this paper, we present a number of fine-grained resources for Natural Language Inference (NLI). In particular, we present a number of resources and validation methods for Greek NLI and a resource for precise NLI. First, we extend the Greek version of the FraCaS test suite to include examples where the inference is directly linked to the syntactic/morphological properties of Greek. The new resource contains an additional 428 examples, making it in total a dataset of 774 examples. Expert annotators have been used in order to create the additional resource, while extensive validation of the original Greek version of the FraCaS by non-expert and expert subjects is performed. Next, we continue the work initiated by (CITATION), according to which a subset of the RTE problems have been labeled for missing hypotheses and we present a dataset an order of magnitude larger, annotating the whole SuperGlUE/RTE dataset with missing hypotheses. Lastly, we provide a de-dropped version of the Greek XNLI dataset, where the pronouns that are missing due to the pro-drop nature of the language are inserted. We then run some models to see the effect of that insertion and report the results.
19.	Bernardy, Jean-Philippe, 1978, et al. (författare) A Compositional Bayesian Semantics for Natural Language 2018 Ingår i: Proceedings of the First International Workshop on Language Cognition and Computational Models, COLING 2018, August 20, 2018 Santa Fe, New Mexico, USA. - : COLING. - 1525-2477. - 9781948087575 Konferensbidrag (refereegranskat)abstract We propose a compositional Bayesian semantics that interprets declarative sentences in a natural language by assigning them probability conditions. These are conditional probabilities that estimate the likelihood that a competent speaker would endorse an assertion, given certain hypotheses. Our semantics is implemented in a functional programming language. It estimates the marginal probability of a sentence through Markov Chain Monte Carlo (MCMC) sampling of objects in vector space models satisfying specified hypotheses. We apply our semantics to examples with several predicates and generalised quantifiers, including higher-order quantifiers. It captures the vagueness of predication (both gradable and non-gradable), without positing a precise boundary for classifier application. We present a basic account of semantic learning based on our semantic system. We compare our proposal to other current theories of probabilistic semantics, and we show that it offers several important advantages over these accounts.
20.	Bernardy, Jean-Philippe, 1978, et al. (författare) A Computational Interpretation of Parametricity 2012 Ingår i: IEEE Symposium on Logic in Computer Science. 27th Annual ACM/IEEE Symposium on Logic in Computer Science (LICS), Dubrovnik, Croatia, June 25-28, 2012. - 1043-6871. - 9780769547695 ; , s. 135-144 Konferensbidrag (refereegranskat)abstract Reynolds' abstraction theorem has recently been extended to lambda-calculi with dependent types. In this paper, we show how this theorem can be internalized. More precisely, we describe an extension of Pure Type Systems with a special parametricity rule (with computational content), and prove fundamental properties such as Church-Rosser's and strong normalization. All instances of the abstraction theorem can be both expressed and proved in the calculus itself. Moreover, one can apply parametricity to the parametricity rule: parametricity is itself parametric.
21.	Bernardy, Jean-Philippe, 1978, et al. (författare) A Computational Treatment of Anaphora and its Algorithmic Implementation 2021 Ingår i: Journal of Logic, Language and Information. - : Springer Science and Business Media LLC. - 0925-8531 .- 1572-9583. ; 30, s. 1-29 Tidskriftsartikel (refereegranskat)abstract In this paper, we propose a framework capable of dealing with anaphora and ellipsis which is both general and algorithmic. This generality is ensured by the compination of two general ideas. First, we use a dynamic semantics which reperent effects using a monad structure. Second we treat scopes flexibly, extending them as needed. We additionally implement this framework as an algorithm which translates abstract syntax to logical formulas. We argue that this framework can provide a unified account of a large number of anaphoric phenomena. Specifically, we show its effectiveness in dealing with pronominal and VP-anaphora, strict and lazy pronouns, lazy identity, bound variable anaphora, e-type pronouns, and cataphora. This means that in particular we can handle complex cases like Bach-Peters sentences, which require an account dealing simultaneously with several phenomena. We use Haskell as a meta-language to present the theory, which also consitutes an implementation of all the phenomena discussed in the paper. To demonstrate coverage, we propose a test suite that can be used to evaluate computational approaches to anaphora.
22.	Bernardy, Jean-Philippe, 1978, et al. (författare) A Logic with Measurable Spaces for Natural Language Semantics 2019 Ingår i: TbiLLC 2019: Thirteenth International Tbilisi Symposium on Language, Logic and Computation,16-20 September 2019.. Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract We present a Logic with Measurable Spaces (LMS) and argue that it is suitable to represent the semantics of many natural language phenomena. LMS draws inspiration from several sources. It is decidable (like description logics). It features Sigma spaces (like Martin-Löf type-theory). It internalises the notion of the cardinality (in fact, here, measures) of spaces and ratios thereof, allowing to capture the notion of event probability. In addition, LMS is arguably a concise system. Thanks to all these qualities, we hope that LMS can play a role in the foundations of natural language semantics.
23.	Bernardy, Jean-Philippe, 1978, et al. (författare) A logic with measurable spaces for natural language semantics 2020 Ingår i: Applied Mathematics, Informatics And Mechanics. - 1512-0074. ; 2020, s. 31-44 Tidskriftsartikel (övrigt vetenskapligt/konstnärligt)abstract We present a Logic with Measurable Spaces (LMS) and argue that it is suitable to represent the semantics of a number of natural lan- guage phenomena. LMS draws inspiration from several sources. It is decidable (like descriptive logics). It features Sigma spaces (like Martin-Lf type-theory). It internalises the notion of the cardinality (in fact, here, measures) of spaces and ratios thereof, allow- ing to capture the notion of event probability. In addition to being a powerful system, it is also concise and has a precise semantics in terms of integrals. Thanks to all these qualities, we hope that LMS can play a role in the foundations of natural language semantics.
24.	Bernardy, Jean-Philippe, 1978, et al. (författare) A Neural Model for Compositional Word Embeddings and Sentence Processing 2022 Ingår i: Proceedings of the Workshop on Cognitive Modeling and Computational Linguistics, May 26, 2022, Dublin, Ireland. - Stroudsburg, PA : Association of Computational Linguistics. - 9781955917292 Konferensbidrag (refereegranskat)abstract We propose a new neural model for word em- beddings, which uses Unitary Matrices as the primary device for encoding lexical information. It uses simple matrix multiplication to de- rive matrices for large units, yielding a sentence processing model that is strictly compositional, does not lose information over time steps, and is transparent, in the sense that word embed- dings can be analysed regardless of context. This model does not employ activation functions, and so the network is fully accessible to analysis by the methods of linear algebra at each point in its operation on an input sequence. We test it in two NLP agreement tasks and obtain rule like perfect accuracy, with greater stability than current state-of-the-art systems. Our proposed model goes some way towards offering a class of computationally powerful deep learning systems that can be fully understood and compared to human cognitive processes for natural language learning and representation.
25.	Bernardy, Jean-Philippe, 1978, et al. (författare) A Presheaf Model of Parametric Type Theory 2015 Ingår i: Electronical Notes in Theoretical Computer Science. - : Elsevier BV. - 1571-0661. ; 319, s. 67-82 Tidskriftsartikel (refereegranskat)abstract We extend Martin-Löf's Logical Framework with special constructions and typing rules providing internalized parametricity. Compared to previous similar proposals, this version comes with a denotational semantics which is a refinement of the standard presheaf semantics of dependent type theory. Further, this presheaf semantics is a refinement of the one used to interpret nominal sets with restrictions. The present calculus is a candidate for the core of a proof assistant with internalized parametricity.
26.	Bernardy, Jean-Philippe, 1978 (författare) A pretty but not greedy printer (functional pearl) 2017 Ingår i: Proceedings of the ACM on Programming Languages. - : Association for Computing Machinery (ACM). - 2475-1421. ; 1:ICFP, s. 1-21 Tidskriftsartikel (refereegranskat)abstract This paper proposes a new specification of pretty printing which is stronger than the state of the art: we require the output to be the shortest possible, and we also offer the ability to align sub-documents at will. We argue that our specification precludes a greedy implementation. Yet, we provide an implementation which behaves linearly in the size of the output. The derivation of the implementation demonstrates functional programming methodology.
27.	Bernardy, Jean-Philippe, 1978 (författare) A Theory of Parametric Polymorphism and an Application 2011 Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract This thesis revisits the well-known notion of parametric polymorphismin the light of modern developments in type-theory. Additionally,applications of parametric polymorphism are also presented.The first part of the thesis presents a theoretical investigation ofthe semantics of parametric polymorphism of and within type-theorieswith dependent types. It is shown how the meaning of polymorphic,possibly dependent, types can be reflected within type-theory itself,via a simple syntactic transformation. This self-referential property opens the door to internalise the transformation in type-theory, and westudy one possible way to do so. We also examine how the translationrelates to various specific features of type-theory, such as proofirrelevance and realizability. The second part is concerned an application of parametric polymorphismrelevant to software engineers. We present a schema to reducepolymorphic properties to equivalent monomorphic properties, for thepurpose of testing. Our proof uses parametricity and properties ofinitial algebras.
28.	Bernardy, Jean-Philippe, 1978, et al. (författare) A Wide-Coverage Symbolic Natural Language Inference System 2019 Ingår i: Proceedings of the 22nd Nordic Conference on Computational Linguistics, 30 September – 2 October, 2019, Turku, Finland / Mareike Hartmann, Barbara Plank (Editors). - University of Linköping : Linköping University Electronic Press. - 1650-3686 .- 1650-3740. - 9789179299958 Konferensbidrag (refereegranskat)
29.	Bernardy, Jean-Philippe, 1978, et al. (författare) An abstraction theorem for pure type systems 2010 Tidskriftsartikel (övrigt vetenskapligt/konstnärligt)
30.	Bernardy, Jean-Philippe, 1978, et al. (författare) Applied Temporal Analysis: A Complete Run of the FraCaS Test Suite 2021 Ingår i: Proceedings of the 14th International Conference on Computational Semantics, June 17 - 18, 2021, Groningen, The Netherlands (Online). - Stroudsburg, PA : The Association for Computational Linguistics. - 9781954085190 Konferensbidrag (refereegranskat)abstract In this paper, we propose an implementation of temporal semantics that translates syntax trees to logical formulas, suitable for consumption by the Coq proof assistant. The analysis supports a wide range of phenomena including: temporal references, temporal adverbs, aspectual classes and progressives. The new semantics are built on top of a previous system handling all sections of the FraCaS test suite except the temporal reference section, and we obtain an accuracy of 81 percent overall and 73 percent for the problems explicitly marked as related to temporal reference. To the best of our knowledge, this is the best performance of a logical system on the whole of the FraCaS.
31.	Bernardy, Jean-Philippe, 1978, et al. (författare) Assessing the Unitary RNN as an End-to-End Compositional Model of Syntax 2022 Ingår i: EPTCS 366. Proceedings End-to-End Compositional Models of Vector-Based Semantics, NUI Galway, 15-16 August, 2022, edited by: Michael Moortgat and Gijs Wijnholds. - Waterloo, Australia : Open Publishing Association. - 2075-2180. Konferensbidrag (refereegranskat)abstract We show that both an LSTM and a unitary-evolution recurrent neural network (URN) can achieve encouraging accuracy on two types of syntactic patterns: context-free long distance agreement, and mildly context-sensitive cross serial dependencies. This work extends recent experiments on deeply nested context-free long distance dependencies, with similar results. URNs differ from LSTMs in that they avoid non-linear activation functions, and they apply matrix multiplication to word embeddings encoded as unitary matrices. This permits them to retain all information in the processing of an input string over arbitrary distances. It also causes them to satisfy strict compositionality. URNs constitute a significant advance in the search for explainable models in deep learning applied to NLP.
32.	Bernardy, Jean-Philippe, 1978, et al. (författare) Bayesian Inference Semantics: A Modelling System and A Test Suite 2019 Ingår i: Proceedings of the Eighth Joint Conference on Lexical and Computational Semantics (*SEM), 6-7 June 2019, Minneapolis, Minnesota, USA / Rada Mihalcea, Ekaterina Shutova, Lun-Wei Ku, Kilian Evang, Soujanya Poria (Editors). - Stroudsburg, PA : Association for Computational Linguistics. - 9781948087933 Konferensbidrag (refereegranskat)abstract We present BIS, a Bayesian Inference Semantics, for probabilistic reasoning in natural language. The current system is based on the framework of Bernardy et al. (2018), but departs from it in important respects. BIS makes use of Bayesian learning for inferring a hypothesis from premises. This involves estimating the probability of the hypothesis, given the data supplied by the premises of an argument. It uses a syntactic parser to generate typed syntactic structures that serve as input to a model generation system. Sentences are interpreted compositionally to probabilistic programs, and the corresponding truth values are estimated using sampling methods. BIS successfully deals with various probabilistic semantic phenomena, including frequency adverbs, generalised quantifiers, generics, and vague predicates. It performs well on a number of interesting probabilistic reasoning tasks. It also sustains most classically valid inferences (instantiation, de Morgan’s laws, etc.). To test BIS we have built an experimental test suite with examples of a range of probabilistic and classical inference patterns.
33.	Bernardy, Jean-Philippe, 1978, et al. (författare) Bayesian Inference Semantics for Natural Language 2022 Ingår i: Probabilistic Approaches to Linguistic Theory / edited by Jean-Philippe Bernardy, Rasmus Blanck, Stergios Chatzikyriakidis, Shalom Lappin, Aleksandre Maskharashvili.. - Stanford : CSLI Publications. - 9781684000791 ; , s. 161-228 Bokkapitel (refereegranskat)abstract We present a Bayesian Inference Semantics for natural language, which computes the probability conditions of sentences compositionally, through semantic functions matching with the types of their syntactic constituents. This system captures the vagueness of classifier terms and scalar modifiers. It also offers a straightforward treatment of the sorites paradox. Our framework expresses probabilistic inferences, which rely on lexically encoded priors, and it captures the effect of informational update on these inferences, through Bayesian modelling. The central device with which we represent probabilistic interpretation is the assignment of measurable spaces to objects and properties. We estimate the probability of a predication by measuring the density of relevant objects in the space of the property that the predicate denotes. We explore two alternative models for the priors. The first one is based on Gaussian distributions, but it exhibits computational intractability with some cases of Monte Carlo sampling. The second is based on uniform densities, and in a number of important instances, it allows us to avoid Monte Carlo sampling. We construct a test suite to illustrate the range of syntactic and semantic constructions, and the inference types, that our system covers.
34.	Bernardy, Jean-Philippe, 1978 (författare) Can Recurrent Neural Networks Learn Nested Recursion? 2018 Ingår i: Linguistic Issues in Language Technology. - 1945-3604. ; 16:1, s. 1-20 Tidskriftsartikel (refereegranskat)abstract Context-free grammars (CFG) were one of the first formal tools used to model natural languages, and they remain relevant today as the basis of several frameworks. A key ingredient of CFG is the presence of nested recursion. In this paper, we investigate experimentally the capability of several recurrent neural networks (RNNs) to learn nested recursion. More precisely, we measure an upper bound of their capability to do so, by simplifying the task to learning a generalized Dyck language, namely one composed of matching parentheses of various kinds. To do so, we present the RNNs with a set of random strings having a given maximum nesting depth and test its ability to predict the kind of closing parenthesis when facing deeper nested strings. We report mixed results: when generalizing to deeper nesting levels, the accuracy of standard RNNs is significantly higher than random, but still far from perfect. Additionally, we propose some non-standard stack-based models which can approach perfect accuracy, at the cost of robustness.
35.	Bernardy, Jean-Philippe, 1978, et al. (författare) Can the Transformer Learn Nested Recursion with Symbol Masking? 2021 Ingår i: Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, August 1 - 6, 2021, Online / Chengqing Zong, Fei Xia, Wenjie Li, Roberto Navigli (Editors). - Stroudsburg, PA : The Association for Computational Linguistics. - 9781954085541 Konferensbidrag (refereegranskat)abstract We investigate if, given a simple symbol masking strategy, self-attention models are capable of learning nested structures and generalise over their depth. We do so in the simplest setting possible, namely languages consisting of nested parentheses of several kinds. We use encoder-only models, which we train to predict randomly masked symbols, in a BERT-like fashion. We find that the accuracy is well above random baseline, with accuracy consistently above 50% both when increasing nesting depth and distances between training and testing. However, we find that the predictions made correspond to a simple parenthesis counting strategy, rather than a push-down automaton. This suggests that self-attention models are not suitable for tasks which require generalisation to more complex instances of recursive structures than those found in the training set.
36.	Bernardy, Jean-Philippe, 1978, et al. (författare) Certified Context-Free Parsing: A formalisation of Valiant's Algorithm in Agda 2016 Ingår i: Logical Methods in Computer Science. - 1860-5974. ; 12:2, s. 28- Tidskriftsartikel (refereegranskat)abstract Valiant (1975) has developed an algorithm for recognition of context free languages. As of today, it remains the algorithm with the best asymptotic complexity for this purpose. In this paper, we present an algebraic specification, implementation, and proof of correctness of a generalisation of Valiant’s algorithm. The generalisation can be used for recognition, parsing or generic calculation of the transitive closure of upper triangular matrices. The proof is certified by the Agda proof assistant. The certification is representative of state-of-the-art methods for specification and proofs in proof assistants based on type-theory. As such, this paper can be read as a tutorial for the Agda system.
37.	Bernardy, Jean-Philippe, 1978, et al. (författare) Compiling Linear Logic using Continuations 2014 Tidskriftsartikel (övrigt vetenskapligt/konstnärligt)abstract As System F is the logical foundation of functional programming, it haslong been expected that Classical Linear Logic (CLL) is the logicalfoundation of concurrent programming. In particular, thanks to aninvolutive negation, its language of propositions correspond to protocols.This means that CLL provides the principles to integrate concurrency intofunctional programming languages.Aiming towards this integration, we translate the concurrency features ofCLL into continuations, essentially via a negation-model of CLL into SystemF. Practically, the translation can be used to embed CLL programs intofunctional languages such as Haskell. We explain the properties of theinterface between the two languages. In particular, using an example, weillustrate how the par (⅋) operator can solve practical functionalprogramming problems.
38.	Bernardy, Jean-Philippe, 1978, et al. (författare) Efficient divide-and-conquer parsing of practical context-free languages 2013 Ingår i: Proceedings of the ACM SIGPLAN International Conference on Functional Programming, ICFP. - New York, NY, USA : ACM. - 9781450323260 ; , s. 111-122 Konferensbidrag (refereegranskat)abstract We present a divide-and-conquer algorithm for parsing context-free languages efficiently. Our algorithm is an instance of Valiant's (1975), who reduced the problem of parsing to matrix multiplications. We show that, while the conquer step of Valiant's is O(n3) in the worst case, it improves to O(log3 n), under certain conditions satisfied by many useful inputs. These conditions occur for example in program texts written by humans. The improvement happens because the multiplications involve an overwhelming majority of empty matrices. This result is relevant to modern computing: divide-and-conquer algorithms can be parallelized relatively easily.
39.	Bernardy, Jean-Philippe, 1978, et al. (författare) Efficient Divide-and-Conquer Parsing of Practical Context-Free Languages 2013 Ingår i: SIGPLAN Notices (ACM Special Interest Group on Programming Languages). - : Association for Computing Machinery (ACM). - 0730-8566 .- 0362-1340 .- 1558-1160. ; 48:9, s. 111-122 Tidskriftsartikel (refereegranskat)abstract We present a divide-and-conquer algorithm for parsing context-free languages efficiently. Our algorithm is an instance of Valiant's (1975), who reduced the problem of parsing to matrix multiplications. We show that, while the conquer step of Valiant's is O(n(3)) in the worst case, it improves to O(log (3) n), under certain conditions satisfied by many useful inputs. These conditions occur for example in program texts written by humans. The improvement happens because the multiplications involve an overwhelming majority of empty matrices. This result is relevant to modern computing: divide-and-conquer algorithms can be parallelized relatively easily.
40.	Bernardy, Jean-Philippe, 1978, et al. (författare) Efficient Parallel and Incremental Parsing 2015 Ingår i: Journal of Functional Programming. - 1469-7653 .- 0956-7968. Tidskriftsartikel (refereegranskat)abstract We present a divide-and-conquer algorithm for parsing context-free languages efficiently. Our algorithm is an instance of Valiant's (1975), who reduced the problem of parsing to matrix multiplications. We show that, while the conquer step of Valiant's is O(n^3), it improves to O(\log^2 n) under certain conditions satisfied by many useful inputs that occur in practice, and if one uses a sparse representation of matrices. The improvement happens because the multiplications involve an overwhelming majority of empty matrices. This result is relevant to modern computing: divide-and-conquer algorithms with a polylogarithmic conquer step can be parallelised relatively easily.
41.	Bernardy, Jean-Philippe, 1978, et al. (författare) Efficient parallel and incremental parsing of practical context-free languages 2015 Ingår i: Journal of Functional Programming. - : Cambridge University Press (CUP). - 1469-7653 .- 0956-7968. ; 25, s. Article Number: UNSP e10- Tidskriftsartikel (refereegranskat)abstract We present a divide-and-conquer algorithm for parsing context-free languages efficiently. Our algorithm is an instance of Valiant's (1975; General context-free recognition in less than cubic time. J. Comput. Syst. Sci. 10(2), 308-314), who reduced the problem of parsing to matrix multiplications. We show that, while the conquer step of Valiant's is O(n(3)), it improves to O(log(2) n) under certain conditions satisfied by many useful inputs that occur in practice, and if one uses a sparse representation of matrices. The improvement happens because the multiplications involve an overwhelming majority of empty matrices. This result is relevant to modern computing: divide-and-conquer algorithms with a polylogarithmic conquer step can be parallelized relatively easily.
42.	Bernardy, Jean-Philippe, 1978, et al. (författare) Evaluating Linear Functions to Symmetric Monoidal Categories 2021 Ingår i: Haskell ’21. Proceedings of the 14th ACM SIGPLAN International Haskell Symposium on Haskell, August 26–27, 2021, Virtual, Republic of Korea / Edited by : Jurriaan Hage. - New York : ACM. - 9781450386159 Konferensbidrag (refereegranskat)
43.	Bernardy, Jean-Philippe, 1978, et al. (författare) Improving the precision of natural textual entailment problem datasets 2020 Ingår i: Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), Marseille, 11–16 May 2020. - Paris : ELRA – European Language Resources Association. - 9791095546344 Konferensbidrag (refereegranskat)abstract In this paper, we propose a method to modify natural textual entailment problem datasets so that they better reflect a more precise notion of entailment. We apply this method to a subset of the Recognizing Textual Entailment datasets. We thus obtain a new corpus of entailment problems, which has the following three characteristics: 1. it is precise (does not leave out implicit hypotheses) 2. it is based on ``real-world'' texts (i.e. most of the premises were written for purposes other than testing textual entailment). 3. its size is 150. % Broadly, the method that we employ is to make any missing hypotheses explicit using a crowd of experts. We discuss the relevance of our method in improving existing NLI datasets to be more fit for precise reasoning and we argue that this corpus can be the basis a first step towards wide-coverage testing of precise natural-language inference systems.
44.	Bernardy, Jean-Philippe, 1978 (författare) Lazy Functional Incremental Parsing 2009 Ingår i: ACM SIGPLAN Haskell Symposium 2009, Edinburgh, Scotland Thursday, 3 September. - 9781605585086 ; , s. 49-60 Konferensbidrag (refereegranskat)abstract Structured documents are commonly edited using a free-form editor. Even though every string is an acceptable input, it makes sense to maintain a structured representation of the edited document. The structured representation has a number of uses: structural navigation (and optional structural editing), structure highlighting, etc. The construction of the structure must be done incrementally to be efficient: the time to process an edit operation should be proportional to the size of the change, and (ideally) independent of the total size of the document.We show that combining lazy evaluation and caching of intermediate (partial) results enables incremental parsing. We build a complete incremental parsing library forinteractive systems with support for error-correction.
45.	Bernardy, Jean-Philippe, 1978, et al. (författare) Learning Syntactic Agreement with Deep Neural Networks 2017 Ingår i: Israel Seminar on Computational Linguistics, September 25, 2017. Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract We consider the extent to which different deep neural network (DNN) con- figurations can learn syntactic relations, by taking up Linzen et al.’s (2016) work on subject-verb agreement with LSTM RNNs. We test their methods on a much larger corpus than they used (a ∼24 million example part of the WaCky corpus, instead of their ∼1.35 million example corpus, both drawn from Wikipedia). We experiment with several different DNN architectures (LSTM RNNs, GRUs, and CNNs), and alternative parameter settings for these systems (vocabulary size, training to test ratio, number of layers, mem- ory size, drop out rate, and lexical embedding dimension size). We also try out our own unsupervised DNN language model. Our results are broadly compat- ible with those that Linzen et al. report. However, we discovered some inter- esting, and in some cases, surprising features of DNNs and language models in their performance of the agreement learning task. In particular, we found that DNNs require large vocabularies to form substantive lexical embeddings in order to learn structural patterns. This finding has significant consequences for our understanding of the way in which DNNs represent syntactic information.
46.	Bernardy, Jean-Philippe, 1978, et al. (författare) Linear Haskell: practical linearity in a higher-order polymorphic language 2018 Ingår i: Proceedings of the ACM on Programming Languages. - : Association for Computing Machinery (ACM). - 2475-1421. ; 2:POPL Tidskriftsartikel (refereegranskat)abstract Linear type systems have a long and storied history, but not a clear path forward to integrate with existing languages such as OCaml or Haskell. In this paper, we study a linear type system designed with two crucial properties in mind: backwards-compatibility and code reuse across linear and non-linear users of a library. Only then can the benefits of linear types permeate conventional functional programming. Rather than bifurcate types into linear and non-linear counterparts, we instead attach linearity to function arrows. Linear functions can receive inputs from linearly-bound values, but can also operate over unrestricted, regular values. To demonstrate the efficacy of our linear type system — both how easy it can be integrated in an existing language implementation and how streamlined it makes it to write programs with linear types — we implemented our type system in ghc, the leading Haskell compiler, and demonstrate two kinds of applications of linear types: mutable data with pure interfaces; and enforcing protocols in I/O-performing functions.
47.	Bernardy, Jean-Philippe, 1978 (författare) Linear Haskell: Today and Tomorrow 2021 Ingår i: Haskell Symposium, 26-27 August, 2021.. Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract Several years after the design of Linear Haskell was proposed, GHC 9 is the first compiler which features linear types. As such Haskell programmers can begin to use linear types today. In the talk I will first recall the main ideas principles behind Linear Haskell. Then I will demonstrate some of what one can do using Linear Haskell as implemented in GHC. Finally I will outline future application of linear types in Haskell and the corresponding research directions.
48.	Bernardy, Jean-Philippe, 1978 (författare) Modalities in Type-Systems 2021 Ingår i: Workshop on Computing Semantics with Types, Frames and Related Structures, July 26-27, 2021, Held virtually. Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract In this talk, I will present a unified framework for extending type systems with modalities. This framework supports several well-known and lesser known applications of modalities. Logical applications include alethic and doxastic logics. Functional programming applications include quantitative and secretive types. Additionally, modalities support tracking (co-)effects in programming languages, which have applications in dynamic semantics.
49.	Bernardy, Jean-Philippe, 1978, et al. (författare) Modelling prosodic structure using Artificial Neural Networks 2017 Ingår i: ExLing 2017. Proceedings of 8 th Tutorial and Research Workshop on Experimental Linguistics, 19-22 June 2017, Heraklion, Crete, Greece / edited by Antonis Botinis. - Athens : University of Athens. - 2529-1092. - 9789604661626 Konferensbidrag (refereegranskat)abstract The ability to accurately perceive whether a speaker is asking a question or is making a statement is crucial for any successful interaction. However, learning and classifying tonal patterns has been a challenging task for automatic speech recognition and for models of tonal representation, as tonal contours are characterized by significant variation. This paper provides a classification model of Cypriot Greek questions and statements. We evaluate two state-of-the-art network architectures: a Long Short-Term Memory (LSTM) network and a convolutional network (ConvNet). The ConvNet outperforms the LSTM in the classification task and exhibited an excellent performance with 95% classification accuracy.
50.	Bernardy, Jean-Philippe, 1978, et al. (författare) Names For Free - Polymorphic Views of Names and Binders 2013 Ingår i: SIGPLAN Notices (ACM Special Interest Group on Programming Languages). - New York, NY, USA : ACM. - 0730-8566. - 9781450323833 ; 48:12, s. 13-24 Konferensbidrag (refereegranskat)abstract We propose a novel technique to represent names and binders inHaskell. The dynamic (run-time) representation is based on de Bruijnindices, but it features an interface to write and manipulatevariables conviently, using Haskell-level lambdas and variables. Thekey idea is to use rich types: a subterm with an additional free variableis viewed either as $\forall v. v \rightarrow \mathsf{Term} (a + v)$or $\exists v. v \times \mathsf{Term} (a + v)$ depending on weather itis constructed or analysed. We demonstrate on a number of exampleshow this approach approach permits to express terms construction and manipulation in a natural way, while retaining the good properties ofrepresentations based on de Bruijn indices.

Skapa referenser, mejla, bekava och länka

Länka till träfflistan

Träfflista för sökning "WFRF:(Bernardy Jean Philippe 1978) "

Avgränsa träffmängd

År