Human language is known to exhibit a nested, hierarchical structure, allowing us to form complex sentences out of smaller pieces. However, many state-of-the-art neural networks models such as Transformers have no explicit hierarchical structure in their architecture—that is, they don’t have an inductive bias toward hierarchical structure. Additionally, Transformers are known to perform poorly on compositional generalization tasks which require such structures. In this paper, we introduce Treeformer, a general-purpose encoder module inspired by the CKY algorithm which learns a composition operator and pooling function to construct hierarchical encodings for phrases and sentences. Our extensive experiments demonstrate the benefits of incorporating hierarchical structure into the Transformer and show significant improvements in compositional generalization as well as in downstream tasks such as machine translation, abstractive summarization, and various natural language understanding tasks.

Figurative language is commonplace in natural language, and while making communication memorable and creative, can be difficult to understand. In this work, we investigate the robustness of Question Answering (QA) models on figurative text. Yes/no questions, in particular, are a useful probe of figurative language understanding capabilities of large language models. We propose FigurativeQA, a set of 1000 yes/no questions with figurative and non-figurative contexts, extracted from the domains of restaurant and product reviews. We show that state-of-the-art BERT-based QA models exhibit an average performance drop of up to 15% points when answering questions from figurative contexts, as compared to non-figurative ones. While models like GPT-3 and ChatGPT are better at handling figurative texts, we show that further performance gains can be achieved by automatically simplifying the figurative contexts into their non-figurative (literal) counterparts. We find that the best overall model is ChatGPT with chain-of-thought prompting to generate non-figurative contexts. Our work provides a promising direction for building more robust QA models with figurative language understanding capabilities.

There has been significant interest in zero and few-shot learning for dialogue state tracking (DST) due to the high cost of collecting and annotating task-oriented dialogues. Recent work has demonstrated that in-context learning requires very little data and zero parameter updates, and even outperforms trained methods in the few-shot setting. We propose RefPyDST, which advances the state of the art with three advancements to in-context learning for DST.First, we formulate DST as a Python programming task, explicitly modeling language coreference as variable reference in Python. Second, since in-context learning depends highly on the context examples, we propose a method to retrieve a diverse set of relevant examples to improve performance. Finally, we introduce a novel re-weighting method during decoding that takes into account probabilities of competing surface forms, and produces a more accurate dialogue state prediction. We evaluate our approach using MultiWOZ and achieve state-of-the-art multi-domain joint-goal accuracy in zero and few-shot settings.

Code-switching, or switching between languages, occurs for many reasons and has important linguistic, sociological, and cultural implications. Multilingual speakers code-switch for a variety of communicative functions, such as expressing emotions, borrowing terms, making jokes, introducing a new topic, etc. The function of code-switching may be quite useful for the analysis of linguists, cognitive scientists, speech therapists, and others, but is not readily apparent. To remedy this situation, we annotate and release a new dataset of functions of code-switching in Spanish-English. We build the first system (to our knowledge) to automatically identify a wide range of functions for which speakers code-switch in everyday speech, achieving an accuracy of 75% across all functions.

This tutorial reviews the design of common meaning representations, SoTA models for predicting meaning representations, and the applications of meaning representations in a wide range of downstream NLP tasks and real-world applications. Reporting by a diverse team of NLP researchers from academia and industry with extensive experience in designing, building and using meaning representations, our tutorial has three components: (1) an introduction to common meaning representations, including basic concepts and design challenges; (2) a review of SoTA methods on building models for meaning representations; and (3) an overview of applications of meaning representations in downstream NLP tasks and real-world applications. We will also present qualitative comparisons of common meaning representations and a quantitative study on how their differences impact model performance. Finally, we will share best practices in choosing the right meaning representation for downstream tasks.

Figurative language is widespread in human language (Lakoff and Johnson, 2008) posing potential challenges in NLP applications. In this paper, we investigate the effect of figurative language on the task of question answering (QA). We construct FigQA, a test set of 400 yes-no questions with figurative and non-figurative contexts, extracted from product reviews and restaurant reviews. We demonstrate that a state-of-the-art RoBERTa QA model has considerably lower performance in question answering when the contexts are figurative rather than literal, indicating a gap in current models. We propose a general method for improving the performance of QA models by converting the figurative contexts into non-figurative by prompting GPT-3, and demonstrate its effectiveness. Our results indicate a need for building QA models infused with figurative language understanding capabilities.

Previous studies have shown that the Abstract Meaning Representation (AMR) can improve Neural Machine Translation (NMT). However, there has been little work investigating incorporating AMR graphs into Transformer models. In this work, we propose a novel encoder-decoder architecture which augments the Transformer model with a Heterogeneous Graph Transformer (Yao et al., 2020) which encodes source sentence AMR graphs. Experimental results demonstrate the proposed model outperforms the Transformer model and previous non-Transformer based models on two different language pairs in both the high resource setting and low resource setting. Our source code, training corpus and released models are available at https://github.com/jlab-nlp/amr-nmt.

Despite extensive research on parsing of English sentences into Abstract Meaning Representation (AMR) graphs, which are compared to gold graphs via the Smatch metric, full-document parsing into a unified graph representation lacks well-defined representation and evaluation. Taking advantage of a super-sentential level of coreference annotation from previous work, we introduce a simple algorithm for deriving a unified graph representation, avoiding the pitfalls of information loss from over-merging and lack of coherence from under merging. Next, we describe improvements to the Smatch metric to make it tractable for comparing document-level graphs and use it to re-evaluate the best published document-level AMR parser. We also present a pipeline approach combining the top-performing AMR parser and coreference resolution systems, providing a strong baseline for future research.

Leveraging additional unlabeled data to boost model performance is common practice in machine learning and natural language processing. For generation tasks, if there is overlap between the additional data and the target text evaluation data, then training on the additional data is training on answers of the test set. This leads to overly-inflated scores with the additional data compared to real-world testing scenarios and problems when comparing models. We study the AMR dataset and Gigaword, which is popularly used for improving AMR-to-text generators, and find significant overlap between Gigaword and a subset of the AMR dataset. We propose methods for excluding parts of Gigaword to remove this overlap, and show that our approach leads to a more realistic evaluation of the task of AMR-to-text generation. Going forward, we give simple best-practice recommendations for leveraging additional data in AMR-to-text generation.

Materials science literature contains millions of materials synthesis procedures described in unstructured natural language text. Large-scale analysis of these synthesis procedures would facilitate deeper scientific understanding of materials synthesis and enable automated synthesis planning. Such analysis requires extracting structured representations of synthesis procedures from the raw text as a first step. To facilitate the training and evaluation of synthesis extraction models, we introduce a dataset of 230 synthesis procedures annotated by domain experts with labeled graphs that express the semantics of the synthesis sentences. The nodes in this graph are synthesis operations and their typed arguments, and labeled edges specify relations between the nodes. We describe this new resource in detail and highlight some specific challenges to annotating scientific text with shallow semantic structure. We make the corpus available to the community to promote further research and development of scientific information extraction systems.