Answer verification identifies correct solutions among candidates generated by large language models (LLMs). Current approaches typically train verifier models by labeling solutions as correct or incorrect based solely on whether the final answer matches the gold answer. However, this approach neglects any flawed rationale in the solution yielding the correct answer, undermining the verifier’s ability to distinguish between sound and flawed rationales. We empirically show that in StrategyQA, only 19% of LLM-generated solutions with correct answers have valid rationales, thus leading to an unreliable verifier. Furthermore, we demonstrate that training a verifier on valid rationales significantly improves its ability to distinguish valid and flawed rationale. To make a better verifier without extra human supervision, we introduce REPS (Rationale Enhancement through Pairwise Selection), a method for selecting valid rationales from candidates by iteratively applying pairwise self-evaluation using the same LLM that generates the solutions. Verifiers trained on solutions selected by REPS outperform those trained using conventional training methods on three reasoning benchmarks (ARC-Challenge, DROP, and StrategyQA). Our results suggest that training reliable verifiers requires ensuring the validity of rationales in addition to the correctness of the final answers, which would be critical for models assisting humans in solving complex reasoning tasks.

What kinds of and how much data is necessary for language models to induce grammatical knowledge to judge sentence acceptability? Recent language models still have much room for improvement in their data efficiency compared to humans. This paper investigates whether language models efficiently use indirect data (indirect evidence), from which they infer sentence acceptability. In contrast, humans use indirect evidence efficiently, which is considered one of the inductive biases contributing to efficient language acquisition. To explore this question, we introduce the Wug InDirect Evidence Test (WIDET), a dataset consisting of training instances inserted into the pre-training data and evaluation instances. We inject synthetic instances with newly coined wug words into pretraining data and explore the model’s behavior on evaluation data that assesses grammatical acceptability regarding those words. We prepare the injected instances by varying their levels of indirectness and quantity. Our experiments surprisingly show that language models do not induce grammatical knowledge even after repeated exposure to instances with the same structure but differing only in lexical items from evaluation instances in certain language phenomena. Our findings suggest a potential direction for future research: developing models that use latent indirect evidence to induce grammatical knowledge.

The imitation of the children’s language acquisition process has been explored to make language models (LMs) more efficient.In particular, errors caused by children’s regularization (so-called overregularization, e.g., using wroted for the past tense of write) have been widely studied to reveal the mechanisms of language acquisition. Existing research has analyzed regularization in language acquisition only by modeling word inflection directly, which is unnatural in light of human language acquisition. In this paper, we hypothesize that language models that imitate the errors children make during language acquisition have a learning process more similar to humans. To verify this hypothesis, we analyzed the learning curve and error preferences of verb inflections in small-scale LMs using acceptability judgments. We analyze the differences in results by model architecture, data, and tokenization. Our model shows child-like U-shaped learning curves clearly for certain verbs, but the preferences for types of overgeneralization did not fully match the observations in children.

Psycholinguistic research suggests that humans may build a representation of linguistic input that is ‘good-enough’ for the task at hand. This study examines what architectural features make language models learn human-like good-enough language processing. We focus on the number of layers and self-attention heads in Transformers. We create a good-enough language processing (GELP) evaluation dataset (7,680 examples), which is designed to test the effects of two plausibility types, eight construction types, and three degrees of memory cost on language processing. To annotate GELP, we first conduct a crowdsourcing experiment whose design follows prior psycholinguistic studies. Our model evaluation against the annotated GELP then reveals that the full model as well as models with fewer layers and/or self-attention heads exhibit a good-enough performance. This result suggests that models with shallower depth and fewer heads can learn good-enough language processing.

In this study, we create a CConS (Counter-commonsense Contextual Size comparison) dataset to investigate how physical commonsense affects the contextualized size comparison task; the proposed dataset consists of both contexts that fit physical commonsense and those that do not. This dataset tests the ability of language models to predict the size relationship between objects under various contexts generated from our curated noun list and templates. We measure the ability of several masked language models and encoder-decoder models. The results show that while large language models can use prepositions such as “in” and “into” in the provided context to infer size relationships, they fail to use verbs and thus make incorrect judgments led by their prior physical commonsense.

What makes a presupposition of an utterance —information taken for granted by its speaker— different from other pragmatic inferences such as an entailment is projectivity (e.g., the negative sentence the boy did not stop shedding tears presupposes the boy had shed tears before). The projectivity may vary depending on the combination of presupposition triggers and environments. However, prior natural language understanding studies fail to take it into account as they either use no human baseline or include only negation as an entailment-canceling environment to evaluate models’ performance. The current study attempts to reconcile these issues. We introduce a new dataset, projectivity of presupposition (PROPRES), which includes 12k premise–hypothesis pairs crossing six triggers involving some lexical variety with five environments. Our human evaluation reveals that humans exhibit variable projectivity in some cases. However, the model evaluation shows that the best-performed model, DeBERTa, does not fully capture it. Our findings suggest that probing studies on pragmatic inferences should take extra care of the human judgment variability and the combination of linguistic items.

To explain the predicted answers and evaluate the reasoning abilities of models, several studies have utilized underlying reasoning (UR) tasks in multi-hop question answering (QA) datasets. However, it remains an open question as to how effective UR tasks are for the QA task when training models on both tasks in an end-to-end manner. In this study, we address this question by analyzing the effectiveness of UR tasks (including both sentence-level and entity-level tasks) in three aspects: (1) QA performance, (2) reasoning shortcuts, and (3) robustness. While the previous models have not been explicitly trained on an entity-level reasoning prediction task, we build a multi-task model that performs three tasks together: sentence-level supporting facts prediction, entity-level reasoning prediction, and answer prediction. Experimental results on 2WikiMultiHopQA and HotpotQA-small datasets reveal that (1) UR tasks can improve QA performance. Using four debiased datasets that are newly created, we demonstrate that (2) UR tasks are helpful in preventing reasoning shortcuts in the multi-hop QA task. However, we find that (3) UR tasks do not contribute to improving the robustness of the model on adversarial questions, such as sub-questions and inverted questions. We encourage future studies to investigate the effectiveness of entity-level reasoning in the form of natural language questions (e.g., sub-question forms).

Natural language understanding (NLU) studies often exaggerate or underestimate the capabilities of systems, thereby limiting the reproducibility of their findings. These erroneous evaluations can be attributed to the difficulty of defining and testing NLU adequately. In this position paper, we reconsider this challenge by identifying two types of researcher degrees of freedom. We revisit Turing’s original interpretation of the Turing test and reveal that an effective test of NLU does not provide an operational definition; it merely provides inductive evidence that the test subject understands the language sufficiently well to meet stakeholder objectives. In other words, stakeholders are free to arbitrarily define NLU through their objectives. To use the test results as inductive evidence, stakeholders must carefully assess if the interpretation of test scores is valid or not. However, designing and using NLU tests involve other degrees of freedom, such as specifying target skills and defining evaluation metrics. As a result, achieving consensus among stakeholders becomes difficult. To resolve this issue, we propose a validity argument, which is a framework comprising a series of validation criteria across test components. By demonstrating that current practices in NLU studies can be associated with those criteria and organizing them into a comprehensive checklist, we prove that the validity argument can serve as a coherent guideline for designing credible test sets and facilitating scientific communication.

To precisely evaluate a language model’s capability for logical reading comprehension, we present a dataset for testing the understanding of the rationale behind critical reasoning. For questions taken from an existing multiple-choice logical reading comprehension dataset, we crowdsource rationale texts that explain why we should select or eliminate answer options, resulting in 3,003 multiple-choice subquestions that are associated with 943 main questions. Experiments on our dataset show that recent large language models (e.g., InstructGPT) struggle to answer the subquestions even if they are able to answer the main questions correctly. We find that the models perform particularly poorly in answering subquestions written for the incorrect options of the main questions, implying that the models have a limited capability for explaining why incorrect alternatives should be eliminated. These results suggest that our dataset encourages further investigation into the critical reasoning ability of language models while focusing on the elimination process of relevant alternatives.

Extractive question answering (QA) models tend to exploit spurious correlations to make predictions when a training set has unintended biases. This tendency results in models not being generalizable to examples where the correlations do not hold. Determining the spurious correlations QA models can exploit is crucial in building generalizable QA models in real-world applications; moreover, a method needs to be developed that prevents these models from learning the spurious correlations even when a training set is biased. In this study, we discovered that the relative position of an answer, which is defined as the relative distance from an answer span to the closest question-context overlap word, can be exploited by QA models as superficial cues for making predictions. Specifically, we find that when the relative positions in a training set are biased, the performance on examples with relative positions unseen during training is significantly degraded. To mitigate the performance degradation for unseen relative positions, we propose an ensemble-based debiasing method that does not require prior knowledge about the distribution of relative positions. We demonstrate that the proposed method mitigates the models’ reliance on relative positions using the biased and full SQuAD dataset. We hope that this study can help enhance the generalization ability of QA models in real-world applications.

For a natural language understanding benchmark to be useful in research, it has to consist of examples that are diverse and difficult enough to discriminate among current and near-future state-of-the-art systems. However, we do not yet know how best to select text sources to collect a variety of challenging examples. In this study, we crowdsource multiple-choice reading comprehension questions for passages taken from seven qualitatively distinct sources, analyzing what attributes of passages contribute to the difficulty and question types of the collected examples. To our surprise, we find that passage source, length, and readability measures do not significantly affect question difficulty. Through our manual annotation of seven reasoning types, we observe several trends between passage sources and reasoning types, e.g., logical reasoning is more often required in questions written for technical passages. These results suggest that when creating a new benchmark dataset, selecting a diverse set of passages can help ensure a diverse range of question types, but that passage difficulty need not be a priority.

Image captioning models require the high-level generalization ability to describe the contents of various images in words. Most existing approaches treat the image–caption pairs equally in their training without considering the differences in their learning difficulties. Several image captioning approaches introduce curriculum learning methods that present training data with increasing levels of difficulty. However, their difficulty measurements are either based on domain-specific features or prior model training. In this paper, we propose a simple yet efficient difficulty measurement for image captioning using cross-modal similarity calculated by a pretrained vision–language model. Experiments on the COCO and Flickr30k datasets show that our proposed approach achieves superior performance and competitive convergence speed to baselines without requiring heuristics or incurring additional training costs. Moreover, the higher model performance on difficult examples and unseen data also demonstrates the generalization ability.

Debiasing language models from unwanted behaviors in Natural Language Understanding (NLU) tasks is a topic with rapidly increasing interest in the NLP community. Spurious statistical correlations in the data allow models to perform shortcuts and avoid uncovering more advanced and desirable linguistic features.A multitude of effective debiasing approaches has been proposed, but flexibility remains a major issue. For the most part, models must be retrained to find a new set of weights with debiased behavior.We propose a new debiasing method in which we identify debiased pruning masks that can be applied to a finetuned model. This enables the selective and conditional application of debiasing behaviors.We assume that bias is caused by a certain subset of weights in the network; our method is, in essence, a mask search to identify and remove biased weights.Our masks show equivalent or superior performance to the standard counterparts, while offering important benefits.Pruning masks can be stored with high efficiency in memory, and it becomes possible to switch among several debiasing behaviors (or revert back to the original biased model) at inference time. Finally, it opens the doors to further research on how biases are acquired by studying the generated masks. For example, we observed that the early layers and attention heads were pruned more aggressively, possibly hinting towards the location in which biases may be encoded.

Several multi-hop reading comprehension datasets have been proposed to resolve the issue of reasoning shortcuts by which questions can be answered without performing multi-hop reasoning. However, the ability of multi-hop models to perform step-by-step reasoning when finding an answer to a comparison question remains unclear. It is also unclear how questions about the internal reasoning process are useful for training and evaluating question-answering (QA) systems. To evaluate the model precisely in a hierarchical manner, we first propose a dataset, HieraDate, with three probing tasks in addition to the main question: extraction, reasoning, and robustness. Our dataset is created by enhancing two previous multi-hop datasets, HotpotQA and 2WikiMultiHopQA, focusing on multi-hop questions on date information that involve both comparison and numerical reasoning. We then evaluate the ability of existing models to understand date information. Our experimental results reveal that the multi-hop models do not have the ability to subtract two dates even when they perform well in date comparison and number subtraction tasks. Other results reveal that our probing questions can help to improve the performance of the models (e.g., by +10.3 F1) on the main QA task and our dataset can be used for data augmentation to improve the robustness of the models.

The possible consequences for the same context may vary depending on the situation we refer to. However, current studies in natural language processing do not focus on situated commonsense reasoning under multiple possible scenarios. This study frames this task by asking multiple questions with the same set of possible endings as candidate answers, given a short story text. Our resulting dataset, Possible Stories, consists of more than 4.5K questions over 1.3K story texts in English. We discover that even current strong pretrained language models struggle to answer the questions consistently, highlighting that the highest accuracy in an unsupervised setting (60.2%) is far behind human accuracy (92.5%). Through a comparison with existing datasets, we observe that the questions in our dataset contain minimal annotation artifacts in the answer options. In addition, our dataset includes examples that require counterfactual reasoning, as well as those requiring readers’ reactions and fictional information, suggesting that our dataset can serve as a challenging testbed for future studies on situated commonsense reasoning.

Crowdsourcing is widely used to create data for common natural language understanding tasks. Despite the importance of these datasets for measuring and refining model understanding of language, there has been little focus on the crowdsourcing methods used for collecting the datasets. In this paper, we compare the efficacy of interventions that have been proposed in prior work as ways of improving data quality. We use multiple-choice question answering as a testbed and run a randomized trial by assigning crowdworkers to write questions under one of four different data collection protocols. We find that asking workers to write explanations for their examples is an ineffective stand-alone strategy for boosting NLU example difficulty. However, we find that training crowdworkers, and then using an iterative process of collecting data, sending feedback, and qualifying workers based on expert judgments is an effective means of collecting challenging data. But using crowdsourced, instead of expert judgments, to qualify workers and send feedback does not prove to be effective. We observe that the data from the iterative protocol with expert assessments is more challenging by several measures. Notably, the human–model gap on the unanimous agreement portion of this data is, on average, twice as large as the gap for the baseline protocol data.

Natural Language Inference (NLI) datasets contain examples with highly ambiguous labels. While many research works do not pay much attention to this fact, several recent efforts have been made to acknowledge and embrace the existence of ambiguity, such as UNLI and ChaosNLI. In this paper, we explore the option of training directly on the estimated label distribution of the annotators in the NLI task, using a learning loss based on this ambiguity distribution instead of the gold-labels. We prepare AmbiNLI, a trial dataset obtained from readily available sources, and show it is possible to reduce ChaosNLI divergence scores when finetuning on this data, a promising first step towards learning how to capture linguistic ambiguity. Additionally, we show that training on the same amount of data but targeting the ambiguity distribution instead of gold-labels can result in models that achieve higher performance and learn better representations for downstream tasks.

Question answering (QA) models for reading comprehension have achieved human-level accuracy on in-distribution test sets. However, they have been demonstrated to lack robustness to challenge sets, whose distribution is different from that of training sets. Existing data augmentation methods mitigate this problem by simply augmenting training sets with synthetic examples sampled from the same distribution as the challenge sets. However, these methods assume that the distribution of a challenge set is known a priori, making them less applicable to unseen challenge sets. In this study, we focus on question-answer pair generation (QAG) to mitigate this problem. While most existing QAG methods aim to improve the quality of synthetic examples, we conjecture that diversity-promoting QAG can mitigate the sparsity of training sets and lead to better robustness. We present a variational QAG model that generates multiple diverse QA pairs from a paragraph. Our experiments show that our method can improve the accuracy of 12 challenge sets, as well as the in-distribution accuracy.

Question answering (QA) models for reading comprehension have been demonstrated to exploit unintended dataset biases such as question–context lexical overlap. This hinders QA models from generalizing to under-represented samples such as questions with low lexical overlap. Question generation (QG), a method for augmenting QA datasets, can be a solution for such performance degradation if QG can properly debias QA datasets. However, we discover that recent neural QG models are biased towards generating questions with high lexical overlap, which can amplify the dataset bias. Moreover, our analysis reveals that data augmentation with these QG models frequently impairs the performance on questions with low lexical overlap, while improving that on questions with high lexical overlap. To address this problem, we use a synonym replacement-based approach to augment questions with low lexical overlap. We demonstrate that the proposed data augmentation approach is simple yet effective to mitigate the degradation problem with only 70k synthetic examples.

Machine reading comprehension (MRC) has received considerable attention as a benchmark for natural language understanding. However, the conventional task design of MRC lacks explainability beyond the model interpretation, i.e., reading comprehension by a model cannot be explained in human terms. To this end, this position paper provides a theoretical basis for the design of MRC datasets based on psychology as well as psychometrics, and summarizes it in terms of the prerequisites for benchmarking MRC. We conclude that future datasets should (i) evaluate the capability of the model for constructing a coherent and grounded representation to understand context-dependent situations and (ii) ensure substantive validity by shortcut-proof questions and explanation as a part of the task design.

A multi-hop question answering (QA) dataset aims to test reasoning and inference skills by requiring a model to read multiple paragraphs to answer a given question. However, current datasets do not provide a complete explanation for the reasoning process from the question to the answer. Further, previous studies revealed that many examples in existing multi-hop datasets do not require multi-hop reasoning to answer a question. In this study, we present a new multi-hop QA dataset, called 2WikiMultiHopQA, which uses structured and unstructured data. In our dataset, we introduce the evidence information containing a reasoning path for multi-hop questions. The evidence information has two benefits: (i) providing a comprehensive explanation for predictions and (ii) evaluating the reasoning skills of a model. We carefully design a pipeline and a set of templates when generating a question-answer pair that guarantees the multi-hop steps and the quality of the questions. We also exploit the structured format in Wikidata and use logical rules to create questions that are natural but still require multi-hop reasoning. Through experiments, we demonstrate that our dataset is challenging for multi-hop models and it ensures that multi-hop reasoning is required.

A challenge in creating a dataset for machine reading comprehension (MRC) is to collect questions that require a sophisticated understanding of language to answer beyond using superficial cues. In this work, we investigate what makes questions easier across recent 12 MRC datasets with three question styles (answer extraction, description, and multiple choice). We propose to employ simple heuristics to split each dataset into easy and hard subsets and examine the performance of two baseline models for each of the subsets. We then manually annotate questions sampled from each subset with both validity and requisite reasoning skills to investigate which skills explain the difference between easy and hard questions. From this study, we observed that (i) the baseline performances for the hard subsets remarkably degrade compared to those of entire datasets, (ii) hard questions require knowledge inference and multiple-sentence reasoning in comparison with easy questions, and (iii) multiple-choice questions tend to require a broader range of reasoning skills than answer extraction and description questions. These results suggest that one might overestimate recent advances in MRC.

Knowing the quality of reading comprehension (RC) datasets is important for the development of natural-language understanding systems. In this study, two classes of metrics were adopted for evaluating RC datasets: prerequisite skills and readability. We applied these classes to six existing datasets, including MCTest and SQuAD, and highlighted the characteristics of the datasets according to each metric and the correlation between the two classes. Our dataset analysis suggests that the readability of RC datasets does not directly affect the question difficulty and that it is possible to create an RC dataset that is easy to read but difficult to answer.

This paper proposes a methodology for building a specialized Japanese data set for recognizing temporal relations and discourse relations. In addition to temporal and discourse relations, multi-layered situational relations that distinguish generic and specific states belonging to different layers in a discourse are annotated. Our methodology has been applied to 170 text fragments taken from Wikinews articles in Japanese. The validity of our methodology is evaluated and analyzed in terms of degree of annotator agreement and frequency of errors.