Although psycholinguists and psychologists have long studied the tendency of linguistic strings to evoke mental images in hearers or readers, most computational studies have applied this concept of imageability only to isolated words. Using recent developments in text-to-image generation models, such as DALLE mini, we propose computational methods that use generated images to measure the imageability of both single English words and connected text. We sample text prompts for image generation from three corpora: human-generated image captions, news article sentences, and poem lines. We subject these prompts to different deformances to examine the model’s ability to detect changes in imageability caused by compositional change. We find high correlation between the proposed computational measures of imageability and human judgments of individual words. We also find the proposed measures more consistently respond to changes in compositionality than baseline approaches. We discuss possible effects of model training and implications for the study of compositionality in text-to-image models.

We explore the task of quotability identification, in which, given a document, we aim to identify which of its passages are the most quotable, i.e. the most likely to be directly quoted by later derived documents. We approach quotability identification as a passage ranking problem and evaluate how well both feature-based and BERT-based (Devlin et al., 2019) models rank the passages in a given document by their predicted quotability. We explore this problem through evaluations on five datasets that span multiple languages (English, Latin) and genres of literature (e.g. poetry, plays, novels) and whose corresponding derived documents are of multiple types (news, journal articles). Our experiments confirm the relatively strong performance of BERT-based models on this task, with the best model, a RoBERTA sequential sentence tagger, achieving an average rho of 0.35 and NDCG@1, 5, 50 of 0.26, 0.31 and 0.40, respectively, across all five datasets.

Growing concern with online misinformation has encouraged NLP research on fact verification. Since writers often base their assertions on structured data, we focus here on verifying textual statements given evidence in tables. Starting from the Table Parsing (TAPAS) model developed for question answering (Herzig et al., 2020), we find that modeling table structure improves a language model pre-trained on unstructured text. Pre-training language models on English Wikipedia table data further improves performance. Pre-training on a question answering task with column-level cell rank information achieves the best performance. With improved pre-training and cell embeddings, this approach outperforms the state-of-the-art Numerically-aware Graph Neural Network table fact verification model (GNN-TabFact), increasing statement classification accuracy from 72.2% to 73.9% even without modeling numerical information. Incorporating numerical information with cell rankings and pre-training on a question-answering task increases accuracy to 76%. We further analyze accuracy on statements implicating single rows or multiple rows and columns of tables, on different numerical reasoning subtasks, and on generalizing to detecting errors in statements derived from the ToTTo table-to-text generation dataset.

Writers often repurpose material from existing texts when composing new documents. Because most documents have more than one source, we cannot trace these connections using only models of document-level similarity. Instead, this paper considers methods for local text reuse detection (LTRD), detecting localized regions of lexically or semantically similar text embedded in otherwise unrelated material. In extensive experiments, we study the relative performance of four classes of neural and bag-of-words models on three LTRD tasks – detecting plagiarism, modeling journalists’ use of press releases, and identifying scientists’ citation of earlier papers. We conduct evaluations on three existing datasets and a new, publicly-available citation localization dataset. Our findings shed light on a number of previously-unexplored questions in the study of LTRD, including the importance of incorporating document-level context for predictions, the applicability of of-the-shelf neural models pretrained on “general” semantic textual similarity tasks such as paraphrase detection, and the trade-offs between more efficient bag-of-words and feature-based neural models and slower pairwise neural models.

We present our work on automatically detecting isnads, the chains of authorities for a re-port that serve as citations in hadith and other classical Arabic texts. We experiment with both sequence labeling methods for identifying isnads in a single pass and a hybrid “retrieve-and-tag” approach, in which a retrieval model first identifies portions of the text that are likely to contain start points for isnads, then a sequence labeling model identifies the exact starting locations within these much smaller retrieved text chunks. We find that the usefulness of full-document sequence to sequence models is limited due to memory limitations and the ineffectiveness of such models at modeling very long documents. We conclude by sketching future improvements on the tagging task and more in-depth analysis of the people and relationships involved in the social network that influenced the evolution of the written tradition over time.

Code-switching has long interested linguists, with computational work in particular focusing on speech and social media data (Sitaram et al., 2019). This paper contrasts these informal instances of code-switching to its appearance in more formal registers, by examining the mixture of languages in the Deutsches Textarchiv (DTA), a corpus of 1406 primarily German books from the 17th to 19th centuries. We automatically annotate and manually inspect spans of six embedded languages (Latin, French, English, Italian, Spanish, and Greek) in the corpus. We quantitatively analyze the differences between code-switching patterns in these books and those in more typically studied speech and social media corpora. Furthermore, we address the practical task of predicting code-switching from features of the matrix language alone in the DTA corpus. Such classifiers can help reduce errors when optical character recognition or speech transcription is applied to a large corpus with rare embedded languages.

We describe and evaluate the Finite-State Arabic Morphologizer (FSAM) – a concatenative (prefix-stem-suffix) and templatic (root- pattern) morphologizer that generates and analyzes undiacritized Modern Standard Arabic (MSA) words, and diacritizes them. Our bidirectional unified-architecture finite state machine (FSM) is based on morphotactic MSA grammatical rules. The FSM models the root-pattern structure related to semantics and syntax, making it readily scalable unlike stem-tabulations in prevailing systems. We evaluate the coverage and accuracy of our model, with coverage being percentage of words in Tashkeela (a large corpus) that can be analyzed. Accuracy is computed against a gold standard, comprising words and properties, created from the intersection of UD PADT treebank and Tashkeela. Coverage of analysis (extraction of root and properties from word) is 82%. Accuracy results are: root computed from a word (92%), word generation from a root (100%), non-root properties of a word (97%), and diacritization (84%). FSAM’s non-root results match or surpass MADAMIRA’s, and root result comparisons are not made because of the concatenative nature of publicly available morphologizers.

Language models have broad adoption in predictive typing tasks. When the typing history contains numerous errors, as in open-vocabulary predictive typing with brain-computer interface (BCI) systems, we observe significant performance degradation in both n-gram and recurrent neural network language models trained on clean text. In evaluations of ranking character predictions, training recurrent LMs on noisy text makes them much more robust to noisy histories, even when the error model is misspecified. We also propose an effective strategy for combining evidence from multiple ambiguous histories of BCI electroencephalogram measurements.

Brain-computer interfaces and other augmentative and alternative communication devices introduce language-modeing challenges distinct from other character-entry methods. In particular, the acquired signal of the EEG (electroencephalogram) signal is noisier, which, in turn, makes the user intent harder to decipher. In order to adapt to this condition, we propose to maintain ambiguous history for every time step, and to employ, apart from the character language model, word information to produce a more robust prediction system. We present preliminary results that compare this proposed Online-Context Language Model (OCLM) to current algorithms that are used in this type of setting. Evaluation on both perplexity and predictive accuracy demonstrates promising results when dealing with ambiguous histories in order to provide to the front end a distribution of the next character the user might type.

We propose a novel approach to OCR post-correction that exploits repeated texts in large corpora both as a source of noisy target outputs for unsupervised training and as a source of evidence when decoding. A sequence-to-sequence model with attention is applied for single-input correction, and a new decoder with multi-input attention averaging is developed to search for consensus among multiple sequences. We design two ways of training the correction model without human annotation, either training to match noisily observed textual variants or bootstrapping from a uniform error model. On two corpora of historical newspapers and books, we show that these unsupervised techniques cut the character and word error rates nearly in half on single inputs and, with the addition of multi-input decoding, can rival supervised methods.