An official website of the United States government
Textual analogies that make comparisons between two concepts are often used for explaining complex ideas, creative writing, and scientific discovery. In this paper, we propose and study a new task, called Analogy Detection and Extraction (AnaDE), which includes three synergistic sub-tasks: 1) detecting documents containing analogies, 2) extracting text segments that make up the analogy, and 3) identifying the (source and target) concepts being compared. To facilitate the study of this new task, we create a benchmark dataset by scraping Metamia.com and investigate the performances of state-of-the-art models on all sub-tasks to establish the first-generation benchmark results for this new task. We find that the Longformer model achieves the best performance on all the three sub-tasks demonstrating its effectiveness for handling long texts. Moreover, smaller models fine-tuned on our dataset perform better than non-finetuned ChatGPT, suggesting high task difficulty. Overall, the models achieve a high performance on documents detection suggesting that it could be used to develop applications like analogy search engines. Further, there is a large room for improvement on the segment and concept extraction tasks.
more »
« less
Validity Assessment of Legal Will Statements as Natural Language Inference
This work introduces a natural language inference (NLI) dataset that focuses on the validity of statements in legal wills. This dataset is unique because: (a) each entailment decision requires three inputs: the statement from the will, the law, and the conditions that hold at the time of the testator’s death; and (b) the included texts are longer than the ones in current NLI datasets. We trained eight neural NLI models in this dataset. All the models achieve more than 80% macro F1 and accuracy, which indicates that neural approaches can handle this task reasonably well. However, group accuracy, a stricter evaluation measure that is calculated with a group of positive and negative examples generated from the same statement as a unit, is in mid 80s at best, which suggests that the models’ understanding of the task remains superficial. Further ablative analyses and explanation experiments indicate that all three text segments are used for prediction, but some decisions rely on semantically irrelevant tokens. This indicates that overfitting on these longer texts likely happens, and that additional research is required for this task to be solved.
more »
« less
- Award ID(s):
- 2217215
- PAR ID:
- 10488663
- Publisher / Repository:
- Association for Computational Linguistics
- Date Published:
- Journal Name:
- Findings of the Association for Computational Linguistics: EMNLP 2022
- Page Range / eLocation ID:
- 6047 to 6056
- Format(s):
- Medium: X
- Location:
- Abu Dhabi, United Arab Emirates
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
null (Ed.)Natural language inference (NLI) is the task of detecting the existence of entailment or contradiction in a given sentence pair. Although NLI techniques could help numerous information retrieval tasks, most solutions for NLI are neural approaches whose lack of interpretability prohibits both straightforward integration and diagnosis for further improvement. We target the task of generating token-level explanations for NLI from a neural model. Many existing approaches for token-level explanation are either computationally costly or require additional annotations for training. In this article, we first introduce a novel method for training an explanation generator that does not require additional human labels. Instead, the explanation generator is trained with the objective of predicting how the model’s classification output will change when parts of the inputs are modified. Second, we propose to build an explanation generator in a multi-task learning setting along with the original NLI task so the explanation generator can utilize the model’s internal behavior. The experiment results suggest that the proposed explanation generator outperforms numerous strong baselines. In addition, our method does not require excessive additional computation at prediction time, which renders it an order of magnitude faster than the best-performing baseline.more » « less
-
null (Ed.)Natural language inference (NLI) is an increasingly important task for natural language understanding, which requires one to infer whether a sentence entails another. However, the ability of NLI models to make pragmatic inferences remains understudied. We create an IMPlicature and PRESupposition diagnostic dataset (IMPPRES), consisting of 32K semi-automatically generated sentence pairs illustrating well-studied pragmatic inference types. We use IMPPRES to evaluate whether BERT, InferSent, and BOW NLI models trained on MultiNLI (Williams et al., 2018) learn to make pragmatic inferences. Although MultiNLI appears to contain very few pairs illustrating these inference types, we find that BERT learns to draw pragmatic inferences. It reliably treats scalar implicatures triggered by “some” as entailments. For some presupposition triggers like “only”, BERT reliably recognizes the presupposition as an entailment, even when the trigger is embedded under an entailment canceling operator like negation. BOW and InferSent show weaker evidence of pragmatic reasoning. We conclude that NLI training encourages models to learn some, but not all, pragmatic inferences.more » « less
-
Discovering the latent topics within texts has been a fundamental task for many applica- tions. However, conventional topic models suffer different problems in different settings. The Latent Dirichlet Allocation (LDA) may not work well for short texts due to the data sparsity (i.e., the sparse word co-occurrence patterns in short documents). The Biterm Topic Model (BTM) learns topics by mod- eling the word-pairs named biterms in the whole corpus. This assumption is very strong when documents are long with rich topic in- formation and do not exhibit the transitivity of biterms. In this paper, we propose a novel way called GraphBTM to represent biterms as graphs and design Graph Convolutional Net- works (GCNs) with residual connections to extract transitive features from biterms. To overcome the data sparsity of LDA and the strong assumption of BTM, we sample a fixed number of documents to form a mini-corpus as a training instance. We also propose a dataset called All News extracted from (Thompson, 2017), in which documents are much longer than 20 Newsgroups. We present an amortized variational inference method for GraphBTM. Our method generates more coherent topics compared with previous approaches. Exper- iments show that the sampling strategy im- proves performance by a large margin.more » « less
-
Discovering the latent topics within texts has been a fundamental task for many applica- tions. However, conventional topic models suffer different problems in different settings. The Latent Dirichlet Allocation (LDA) may not work well for short texts due to the data sparsity (i.e., the sparse word co-occurrence patterns in short documents). The Biterm Topic Model (BTM) learns topics by mod- eling the word-pairs named biterms in the whole corpus. This assumption is very strong when documents are long with rich topic in- formation and do not exhibit the transitivity of biterms. In this paper, we propose a novel way called GraphBTM to represent biterms as graphs and design Graph Convolutional Net- works (GCNs) with residual connections to extract transitive features from biterms. To overcome the data sparsity of LDA and the strong assumption of BTM, we sample a fixed number of documents to form a mini-corpus as a training instance. We also propose a dataset called All News extracted from (Thompson, 2017), in which documents are much longer than 20 Newsgroups. We present an amortized variational inference method for GraphBTM. Our method generates more coherent topics compared with previous approaches. Exper- iments show that the sampling strategy im- proves performance by a large margin.more » « less
