SentSpace is a modular framework for streamlined evaluation of text. SentSpacecharacterizes textual input using diverse lexical, syntactic, and semantic features derivedfrom corpora and psycholinguistic experiments. Core sentence features fall into three primaryfeature spaces: 1) Lexical, 2) Contextual, and 3) Embeddings. To aid in the analysis of computed features, SentSpace provides a web interface for interactive visualization and comparison with text from large corpora. The modular design of SentSpace allows researchersto easily integrate their own feature computation into the pipeline while benefiting from acommon framework for evaluation and visualization. In this manuscript we will describe thedesign of SentSpace, its core feature spaces, and demonstrate an example use case by comparing human-written and machine-generated (GPT2-XL) sentences to each other. We findthat while GPT2-XL-generated text appears fluent at the surface level, psycholinguistic normsand measures of syntactic processing reveal key differences between text produced by humansand machines. Thus, SentSpace provides a broad set of cognitively motivated linguisticfeatures for evaluation of text within natural language processing, cognitive science, as wellas the social sciences.

Multilingual language models achieve impressive zero-shot accuracies in many languages in complex tasks such as Natural Language Inference (NLI). Examples in NLI (and equivalent complex tasks) often pertain to various types of sub-tasks, requiring different kinds of reasoning. Certain types of reasoning have proven to be more difficult to learn in a monolingual context, and in the crosslingual context, similar observations may shed light on zero-shot transfer efficiency and few-shot sample selection. Hence, to investigate the effects of types of reasoning on transfer performance, we propose a category-annotated multilingual NLI dataset and discuss the challenges to scale monolingual annotations to multiple languages. We statistically observe interesting effects that the confluence of reasoning types and language similarities have on transfer performance.

Neural models excel at extracting statistical patterns from large amounts of data, but struggle to learn patterns or reason about language from only a few examples. In this paper, we ask: Can we learn explicit rules that generalize well from only a few examples? We explore this question using program synthesis. We develop a synthesis model to learn phonology rules as programs in a domain-specific language. We test the ability of our models to generalize from few training examples using our new dataset of problems from the Linguistics Olympiad, a challenging set of tasks that require strong linguistic reasoning ability. In addition to being highly sample-efficient, our approach generates human-readable programs, and allows control over the generalizability of the learnt programs.

Automatic fact-checking is crucial for recognizing misinformation spreading on the internet. Most existing fact-checkers break down the process into several subtasks, one of which determines candidate evidence sentences that can potentially support or refute the claim to be verified; typically, evidence sentences with gold-standard labels are needed for this. In a more realistic setting, however, such sentence-level annotations are not available. In this paper, we tackle the natural language inference (NLI) subtask—given a document and a (sentence) claim, determine whether the document supports or refutes the claim—only using document-level annotations. Using fine-tuned BERT and multiple instance learning, we achieve 81.9% accuracy, significantly outperforming the existing results on the WikiFactCheck-English dataset.

Automated fact checking is becoming increasingly vital as both truthful and fallacious information accumulate online. Research on fact checking has benefited from large-scale datasets such as FEVER and SNLI. However, such datasets suffer from limited applicability due to the synthetic nature of claims and/or evidence written by annotators that differ from real claims and evidence on the internet. To this end, we present WikiFactCheck-English, a dataset of 124k+ triples consisting of a claim, context and an evidence document extracted from English Wikipedia articles and citations, as well as 34k+ manually written claims that are refuted by the evidence documents. This is the largest fact checking dataset consisting of real claims and evidence to date; it will allow the development of fact checking systems that can better process claims and evidence in the real world. We also show that for the NLI subtask, a logistic regression system trained using existing and novel features achieves peak accuracy of 68%, providing a competitive baseline for future work. Also, a decomposable attention model trained on SNLI significantly underperforms the models trained on this dataset, suggesting that models trained on manually generated data may not be sufficiently generalizable or suitable for fact checking real-world claims.

Pre-trained Transformer-based neural architectures have consistently achieved state-of-the-art performance in the Natural Language Inference (NLI) task. Since NLI examples encompass a variety of linguistic, logical, and reasoning phenomena, it remains unclear as to which specific concepts are learnt by the trained systems and where they can achieve strong generalization. To investigate this question, we propose a taxonomic hierarchy of categories that are relevant for the NLI task. We introduce TaxiNLI, a new dataset, that has 10k examples from the MNLI dataset with these taxonomic labels. Through various experiments on TaxiNLI, we observe that whereas for certain taxonomic categories SOTA neural models have achieved near perfect accuracies—a large jump over the previous models—some categories still remain difficult. Our work adds to the growing body of literature that shows the gaps in the current NLI systems and datasets through a systematic presentation and analysis of reasoning categories.