Towards Debiasing Fact Verification Models

Fact verification requires validating a claim in the context of evidence. We show, however, that in the popular FEVER dataset this might not necessarily be the case. Claim-only classifiers perform competitively with top evidence-aware models. In this paper, we investigate the cause of this phenomenon, identifying strong cues for predicting labels solely based on the claim, without considering any evidence. We create an evaluation set that avoids those idiosyncrasies. The performance of FEVER-trained models significantly drops when evaluated on this test set. Therefore, we introduce a regularization method which alleviates the effect of bias in the training data, obtaining improvements on the newly created test set. This work is a step towards a more sound evaluation of reasoning capabilities in fact verification models.

A Structured Distributional Model of Sentence Meaning and Processing

Most compositional distributional semantic models represent sentence meaning with a single vector. In this paper, we propose a Structured Distributional Model (SDM) that combines word embeddings with formal semantics and is based on the assumption that sentences represent events and situations. The semantic representation of a sentence is a formal structure derived from Discourse Representation Theory and containing distributional vectors. This structure is dynamically and incrementally built by integrating knowledge about events and their typical participants, as they are activated by lexical items. Event knowledge is modeled as a graph extracted from parsed corpora and encoding roles and relationships between participants that are represented as distributional vectors. SDM is grounded on extensive psycholinguistic research showing that generalized knowledge about events stored in semantic memory plays a key role in sentence comprehension. We evaluate SDM on two recently introduced compositionality datasets, and our results show that combining a simple compositional model with event knowledge constantly improves performances, even with different types of word embeddings.

Unsupervised Text Style Transfer via Iterative Matching and Translation

Text style transfer seeks to learn how to automatically rewrite sentences from a source domain to the target domain in different styles, while simultaneously preserving their semantic contents. A major challenge in this task stems from the lack of parallel data that connects the source and target styles. Existing approaches try to disentangle content and style, but this is quite difficult and often results in poor content-preservation and grammaticality. In contrast, we propose a novel approach by first constructing a pseudo-parallel resource that aligns a subset of sentences with similar content between source and target corpus. And then a standard sequence-to-sequence model can be applied to learn the style transfer. Subsequently, we iteratively refine the learned style transfer function while improving upon the imperfections in our original alignment. Our method is applied to the tasks of sentiment modification and formality transfer, where it outperforms state-of-the-art systems by a large margin. As an auxiliary contribution, we produced a publicly-available test set with human-generated style transfers for future community use.

GraphIE: A Graph-Based Framework for Information Extraction

Most modern Information Extraction (IE) systems are implemented as sequential taggers and focus on modelling local dependencies. Non-local and non-sequential context is, however, a valuable source of information to improve predictions. In this paper, we introduce GraphIE, a framework that operates over a graph representing both local and non-local dependencies between textual units (i.e. words or sentences). The algorithm propagates information between connected nodes through graph convolutions and exploits the richer representation to improve word level predictions. The framework is evaluated on three different tasks, namely social media, textual and visual information extraction. Results show that GraphIE outperforms a competitive baseline (BiLSTM+CRF) in all tasks by a significant margin.

A Rank-Based Similarity Metric for Word Embeddings

Word Embeddings have recently imposed themselves as a standard for representing word meaning in NLP. Semantic similarity between word pairs has become the most common evaluation benchmark for these representations, with vector cosine being typically used as the only similarity metric. In this paper, we report experiments with a rank-based metric for WE, which performs comparably to vector cosine in similarity estimation and outperforms it in the recently-introduced and challenging task of outlier detection, thus suggesting that rank-based measures can improve clustering quality.

BomJi at SemEval-2018 Task 10: Combining Vector-, Pattern- and Graph-based Information to Identify Discriminative Attributes

This paper describes BomJi, a supervised system for capturing discriminative attributes in word pairs (e.g. yellow as discriminative for banana over watermelon). The system relies on an XGB classifier trained on carefully engineered graph-, pattern- and word embedding based features. It participated in the SemEval- 2018 Task 10 on Capturing Discriminative Attributes, achieving an F1 score of 0:73 and ranking 2nd out of 26 participant systems.

Is Structure Necessary for Modeling Argument Expectations in Distributional Semantics?

Despite the number of NLP studies dedicated to thematic fit estimation, little attention has been paid to the related task of composing and updating verb argument expectations. The few exceptions have mostly modeled this phenomenon with structured distributional models, implicitly assuming a similarly structured representation of events. Recent experimental evidence, however, suggests that human processing system could also exploit an unstructured "bag-of-arguments" type of event representation to predict upcoming input. In this paper, we re-implement a traditional structured model and adapt it to compare the different hypotheses concerning the degree of structure in our event knowledge, evaluating their relative performance in the task of the argument expectations update.

Measuring Thematic Fit with Distributional Feature Overlap

In this paper, we introduce a new distributional method for modeling predicate-argument thematic fit judgments. We use a syntax-based DSM to build a prototypical representation of verb-specific roles: for every verb, we extract the most salient second order contexts for each of its roles (i.e. the most salient dimensions of typical role fillers), and then we compute thematic fit as a weighted overlap between the top features of candidate fillers and role prototypes. Our experiments show that our method consistently outperforms a baseline re-implementing a state-of-the-art system, and achieves better or comparable results to those reported in the literature for the other unsupervised systems. Moreover, it provides an explicit representation of the features characterizing verb-specific semantic roles.

German in Flux: Detecting Metaphoric Change via Word Entropy

This paper explores the information-theoretic measure entropy to detect metaphoric change, transferring ideas from hypernym detection to research on language change. We also build the first diachronic test set for German as a standard for metaphoric change annotation. Our model shows high performance, is unsupervised, language-independent and generalizable to other processes of semantic change.

Hypernyms under Siege: Linguistically-motivated Artillery for Hypernymy Detection

The fundamental role of hypernymy in NLP has motivated the development of many methods for the automatic identification of this relation, most of which rely on word distribution. We investigate an extensive number of such unsupervised measures, using several distributional semantic models that differ by context type and feature weighting. We analyze the performance of the different methods based on their linguistic motivation. Comparison to the state-of-the-art supervised methods shows that while supervised methods generally outperform the unsupervised ones, the former are sensitive to the distribution of training instances, hurting their reliability. Being based on general linguistic hypotheses and independent from training data, unsupervised measures are more robust, and therefore are still useful artillery for hypernymy detection.