16,878 Hits in 3.6 sec

Generating Contradictory, Neutral, and Entailing Sentences [article]

Yikang Shen, Shawn Tan, Chin-Wei Huang, Aaron Courville
2018 arXiv   pre-print
In our paper, we propose an approach to generating sentences, conditioned on an input sentence and a logical inference label.  ...  The experiment results show that, given our framework, we have clear ways to improve the quality and diversity of generated sentences.  ...  To do so, we require two simultaneous conditions: (i) that the input sentence entails the output sentence; and (ii) that the output sentence entails the input sentence.  ... 
arXiv:1803.02710v1 fatcat:xbizxy5lznbi3jyddgerakpm54

Unsupervised Natural Language Inference Using PHL Triplet Generation [article]

Neeraj Varshney, Pratyay Banerjee, Tejas Gokhale, Chitta Baral
2022 arXiv   pre-print
As a solution, we propose a procedural data generation approach that leverages a set of sentence transformations to collect PHL (Premise, Hypothesis, Label) triplets for training NLI models, bypassing  ...  We investigate it under three settings: PH, P, and NPH that differ in the extent of unlabeled data available for learning.  ...  A sentence is treated as premise, and multiple hypotheses conditioned on each label (Entailment-E, Contradiction-C, and Neutral-N) are generated using a set of sentence transformations. consuming and can  ... 
arXiv:2110.08438v2 fatcat:foxtf2abdrcfzkhtwnhdlxuzkm

Textual Inference: getting logic from humans

Aikaterini-Lida Kalouli, Livy Real, Valeria de Paiva
2017 International Conference on Computational Semantics  
The system provides conceptual semantics for sentences, so that entailment-contradiction-neutrality relations between sentences can be identified.  ...  We also show that detecting the relationship between some pairs of the SICK corpus requires more than just lexical semantics, which provides us with guidelines and intuitions for our further implementation  ...  Thus, A cannot entail B and should rather be neutral with respect to B as B is neutral with respect to A.  ... 
dblp:conf/iwcs/KalouliRP17 fatcat:lcc7h3f6tzayffqdkvidx7exoi

Weakly Supervised Explainable Phrasal Reasoning with Neural Fuzzy Logic [article]

Zijun Wu, Atharva Naik, Zi Xuan Zhang, Lili Mou
2021 arXiv   pre-print
Natural language inference (NLI) aims to determine the logical relationship between two sentences among the target labels Entailment, Contradiction, and Neutral.  ...  Then, the model predicts the NLI label for the aligned phrases, and induces the sentence label by fuzzy logic formulas.  ...  Rule for Neutral: Two sentences are neutral if there exists (at least) one neutral phrase pair, but there does not exist any contradictory phrase pair.  ... 
arXiv:2109.08927v1 fatcat:eoi6rywopbga5byokgl3lvnfti

MonaLog: a Lightweight System for Natural Language Inference Based on Monotonicity [article]

Hai Hu, Qi Chen, Kyle Richardson, Atreyee Mukherjee, Lawrence S. Moss, Sandra Kuebler
2019 arXiv   pre-print
We show that MonaLog is capable of generating large amounts of high-quality training data for BERT, improving its accuracy on SICK.  ...  quantifiers, lexical items and tokenlevel polarity information.  ...  Contradictory sentences To generate sentences contradictory to the input sentence, we do the following: 1) if the sentence starts with "no (some)", replace the first word with "some (no)". 2) If the object  ... 
arXiv:1910.08772v1 fatcat:kelkmh3ymfbwbbffrnzbn5lmqu

Christ – A Contradiction: A Defense of Contradictory Christology

Jc Beall
2019 The Journal of Analytic Theology  
Christ is human (with everything entailed thereby) and Christ is divine (with everything entailed thereby). Being divine entails (among many other of God's properties) being immutable.  ...  In this paper, I put the simplest explanation on the theological table: namely, Christ appears to be contradictory because Christ is contradictory (i.e., some predicate is both true and false of Christ  ...  I claim that a Contradictory Christology -and negation-inconsistent theology in general -is viable.  ... 
doi:10.12978/jat.2019-7.090202010411 fatcat:vc2ebhqv2favtnges35w3okdei

On Contradictory Christology: Preliminary Remarks, Notation and Terminology

Jc Beall
2019 The Journal of Analytic Theology  
In that paper I advance and defend a contradictory Christology which solves the fundamental 'problem' of Christology by holding that Christ is a contradictory being: it is true that Christ is mutable and  ...  it is false that Christ is mutable; it is true that Christ is immutable and it is false that Christ is immutable; and so on for other fundamental-problem properties.  ...  theology in general).  ... 
doi:10.12978/jat.2019-7.352828273037 fatcat:pd7jed2uxffrxkih3uyse56uzm

Contradiction Detection in Persian Text [article]

Zeinab Rahimi, Mehrnoush ShamsFard
2021 arXiv   pre-print
Detection of semantic contradictory sentences is one of the most challenging and fundamental issues for NLP applications such as recognition of textual entailments.  ...  Extracted rules are tested for different categories of contradictory sentences.  ...  entailment. (2) Set of 250 sentence pairs (130 contradictory and 120 neutral sentence pairs) containing different categories of contradictions that were manually generated for testing the system.  ... 
arXiv:2107.01987v1 fatcat:5qdc56ysejhrhndykph66vnkky

Evaluating Compositionality in Sentence Embeddings [article]

Ishita Dasgupta, Demi Guo, Andreas Stuhlmüller, Samuel J. Gershman, Noah D. Goodman
2018 arXiv   pre-print
We find that the performance of state of the art sentence embeddings (InferSent; Conneau et al., 2017) on our new dataset is poor.  ...  This highlights the importance of structured datasets in better understanding and improving AI systems.  ...  NDG is supported by DARPA agreement number FA8750-14-2-0009, and a Sloan Foundation Research Fellowship. Code and data are available at: ishita-dg/ScrambleTests.  ... 
arXiv:1802.04302v2 fatcat:n3fx2tkg35d3bik2bwdshenxay

Contradiction Detection with Contradiction-Specific Word Embedding

Luyang Li, Bing Qin, Ting Liu
2017 Algorithms  
CWE is learned from a training corpus which is automatically generated from the paraphrase database, and is naturally applied as features to carry out contradiction detection in SemEval 2014 benchmark  ...  Contradiction detection is a task to recognize contradiction relations between a pair of sentences.  ...  Given a pair of sentences, we aim to predict the correct semantic relation between them. The semantic relations include: entailment, contradiction and neutral.  ... 
doi:10.3390/a10020059 fatcat:tqzwdrfid5e7tos23zdjihdq5q

Open IE-Triples Inference - Corpora Development and DNN Architectures

Martin Víta, Petr Škoda
2020 Proceedings of the 2020 Federated Conference on Computer Science and Information Systems  
This task is usually stated as a 3-way classification of sentence pairs with respect to entailment relation (entailment, neutral, contradiction).  ...  In this work, we focus on a derived task of relation inference: we propose a method of transforming a general NLI corpus to an annotated corpus for relation inference that utilizes existing NLI annotations  ...  Accuracy w.r.t. the ENTAILMENT label is 0.9853, NEUTRAL label: 0.9412 and CONTRADICTION label: 0.9772.  ... 
doi:10.15439/2020f78 dblp:conf/fedcsis/Vita020 fatcat:zz6cd3xfazdgnbbl5poaczttka

SemEval-2014 Task 1: Evaluation of Compositional Distributional Semantic Models on Full Sentences through Semantic Relatedness and Textual Entailment

Marco Marelli, Luisa Bentivogli, Marco Baroni, Raffaella Bernardi, Stefano Menini, Roberto Zamparelli
2014 Proceedings of the 8th International Workshop on Semantic Evaluation (SemEval 2014)  
Systems were presented with pairs of sentences and were evaluated on their ability to predict human judgments on (i) semantic relatedness and (ii) entailment.  ...  We received 17 submissions in the relatedness subtask (for a total of 66 runs) and 18 in the entailment subtask (65 runs).  ...  Acknowledgments We thank the creators of the ImageFlickr, MSR-Video, and SemEval-2012 STS data sets for granting us permission to use their data for the task.  ... 
doi:10.3115/v1/s14-2001 dblp:conf/semeval/MarelliBBBMZ14 fatcat:smqlwhf4lffabgvvm4sggtqnx4

Natural Language Inference from Multiple Premises [article]

Alice Lai, Yonatan Bisk, Julia Hockenmaier
2017 arXiv   pre-print
We define a novel textual entailment task that requires inference over multiple premise sentences.  ...  We evaluate several strong neural baselines and analyze how the multiple premise task differs from standard textual entailment.  ...  (DARPA) and the Army Research Office (ARO).  ... 
arXiv:1710.02925v1 fatcat:oizpw2eet5aqxlpanzhnav3hfe

e-SNLI: Natural Language Inference with Natural Language Explanations [article]

Oana-Maria Camburu, Tim Rocktäschel, Thomas Lukasiewicz, Phil Blunsom
2018 arXiv   pre-print
We show how our corpus of explanations, which we call e-SNLI, can be used for various goals, such as obtaining full sentence justifications of a model's decisions, improving universal sentence representations  ...  In this work, we extend the Stanford Natural Language Inference dataset with an additional layer of human-annotated natural language explanations of the entailment relations.  ...  For example, high overlap in words between premise and hypothesis usually predicts entailment, while most contradictory sentence pairs have no or very little overlap of words.  ... 
arXiv:1812.01193v2 fatcat:leshzirbs5d7rjmuhliimonzvq

On Contradictory Christology: A Reply to Uckelman's 'Contradictions, Impossibility, and Triviality'

Jc Beall
2019 The Journal of Analytic Theology  
On my usage of 'contradictory sentences' A sentence ‫ܣ‬ is contradictory just if ‫,ܣ‬ together with (possibly no) other true sentences, entails ‫.ܣ!‬  ...  sentences (discussed in the 'Preliminaries' piece and elsewhere in this symposium). 7 A sentence is explosive with respect to a consequence (entailment) relation if and only if the sentence entails  ... 
doi:10.12978/jat.2019-7.139132132131134141 fatcat:pqbouo6vzfdbdj4mokjqwmbskq
« Previous Showing results 1 — 15 out of 16,878 results