ACL 2023 Papers

Argument maps structure discourse into nodes in a tree with each node being an argument that supports or opposes its parent argument. This format is more comprehensible and less redundant compared to an unstructured one. Exploring those maps and maintaining their structure by placing new arguments under suitable parents is more challenging for users with huge maps that are typical in online discussions. To support those users, we introduce the task of node placement: suggesting candidate nodes as parents for a new contribution. We establish an upper-bound of human performance, and conduct experiments with models of various sizes and training strategies. We experiment with a selection of maps from Kialo, drawn from a heterogeneous set of domains. Based on an annotation study, we highlight the ambiguity of the task that makes it challenging for both humans and models. We examine the unidirectional relation between tree nodes and show that encoding a node into different embeddings for each of the parent and child cases improves performance. We further show the few-shot effectiveness of our approach.

The accurate prediction of lexical relations between words is a challenging task in Natural Language Processing (NLP). The most recent advances in this direction come with the use of pre-trained language models (PTLMs). A PTLM typically needs “well-formed” verbalized text to interact with it, either to fine-tune it or to exploit it. However, there are indications that commonly used PTLMs already encode enough linguistic knowledge to allow the use of minimal (or none) textual context for some linguistically motivated tasks, thus notably reducing human effort, the need for data pre-processing, and favoring techniques that are language neutral since do not rely on syntactic structures. In this work, we explore this idea for the tasks of lexical relation classification (LRC) and graded Lexical Entailment (LE). After fine-tuning PTLMs for LRC with different verbalizations, our evaluation results show that very simple prompts are competitive for LRC and significantly outperform graded LE SoTA. In order to gain a better insight into this phenomenon, we perform a number of quantitative statistical analyses on the results, as well as a qualitative visual exploration based on embedding projections.

Through the use of first name substitution experiments, prior research has demonstrated the tendency of social commonsense reasoning models to systematically exhibit social biases along the dimensions of race, ethnicity, and gender (An et al., 2023). Demographic attributes of first names, however, are strongly correlated with corpus frequency and tokenization length, which may influence model behavior independent of or in addition to demographic factors. In this paper, we conduct a new series of first name substitution experiments that measures the influence of these factors while controlling for the others. We find that demographic attributes of a name (race, ethnicity, and gender) and name tokenization length are both factors that systematically affect the behavior of social commonsense reasoning models.

Metaphors are highly creative constructs of human language that grow old and eventually die. Popular datasets used for metaphor processing tasks were constructed from dated source texts. In this paper, we propose NewsMet, a large high-quality contemporary dataset of news headlines hand-annotated with metaphorical verbs. The dataset comprises headlines from various sources including political, satirical, reliable and fake. Our dataset serves the purpose of evaluation for the tasks of metaphor interpretation and generation. The experiments reveal several insights and limitations of using LLMs to automate metaphor processing tasks as frequently seen in the recent literature. The dataset is publicly available for research purposes https://github.com/AxleBlaze3/NewsMet_Metaphor_Dataset.

Hot news is one of the most popular topics in daily conversations. However, news grounded conversation has long been stymied by the lack of well-designed task definition and scarce data. In this paper, we propose a novel task, Proactive News Grounded Conversation, in which a dialogue system can proactively lead the conversation based on some key topics of the news. In addition, both information-seeking and chit-chat scenarios are included realistically, where the user may ask a series of questions about the news details or express their opinions and be eager to chat. To further develop this novel task, we collect a human-to-human Chinese dialogue dataset NewsDialogues, which includes 1K conversations with a total of 14.6K utterances and detailed annotations for target topics and knowledge spans. Furthermore, we propose a method named Predict-Generate-Rank, consisting of a generator for grounded knowledge prediction and response generation, and a ranker for the ranking of multiple responses to alleviate the exposure bias. We conduct comprehensive experiments to demonstrate the effectiveness of the proposed method and further present several key findings and challenges to prompt future research.

We present a state-of-the-art neural approach to the unsupervised reconstruction of ancient word forms. Previous work in this domain used expectation-maximization to predict simple phonological changes between ancient word forms and their cognates in modern languages. We extend this work with neural models that can capture more complicated phonological and morphological changes. At the same time, we preserve the inductive biases from classical methods by building monotonic alignment constraints into the model and deliberately underfitting during the maximization step. We evaluate our performance on the task of reconstructing Latin from a dataset of cognates across five Romance languages, achieving a notable reduction in edit distance from the target word forms compared to previous methods.

Neural topic models have been widely used to extract common topics across documents. Recently, contrastive learning has been applied to variational autoencoder-based neural topic models, achieving promising results. However, due to the limitation of the unidirectional structure of the variational autoencoder, the encoder is enhanced with the contrastive loss instead of the decoder, leading to a gap between model training and evaluation. To address the limitation, we propose a novel neural topic modeling framework based on cycle adversarial training and contrastive learning to apply contrastive learning on the generator directly. Specifically, a self-supervised contrastive loss is proposed to make the generator capture similar topic information, which leads to better topic-word distributions. Meanwhile, a discriminative contrastive loss is proposed to cooperate with the self-supervised contrastive loss to balance the generation and discrimination. Moreover, based on the reconstruction ability of the cycle generative adversarial network, a novel data augmentation strategy is designed and applied to the topic distribution directly. Experiments have been conducted on four benchmark datasets and results show that the proposed approach outperforms competitive baselines.

We propose a semi-supervised text classifier based on self-training using one positive and one negative property of neural networks. One of the weaknesses of self-training is the semantic drift problem, where noisy pseudo-labels accumulate over iterations and consequently the error rate soars. In order to tackle this challenge, we reshape the role of pseudo-labels and create a hierarchical order of information. In addition, a crucial step in self-training is to use the classifier confidence prediction to select the best candidate pseudo-labels. This step cannot be efficiently done by neural networks, because it is known that their output is poorly calibrated. To overcome this challenge, we propose a hybrid metric to replace the plain confidence measurement. Our metric takes into account the prediction uncertainty via a subsampling technique. We evaluate our model in a set of five standard benchmarks, and show that it significantly outperforms a set of ten diverse baseline models. Furthermore, we show that the improvement achieved by our model is additive to language model pretraining, which is a widely used technique for using unlabeled documents.

We tackle the problem of neural machine translation of mathematical formulae between ambiguous presentation languages and unambiguous content languages. Compared to neural machine translation on natural language, mathematical formulae have a much smaller vocabulary and much longer sequences of symbols, while their translation requires extreme precision to satisfy mathematical information needs. In this work, we perform the tasks of translating from LaTeX to Mathematica as well as from LaTeX to semantic LaTeX. While recurrent, recursive, and transformer networks struggle with preserving all contained information, we find that convolutional sequence-to-sequence networks achieve 95.1% and 90.7% exact matches, respectively.

State-of-the-art techniques common to low resource Machine Translation (MT) are applied to improve MT of spoken language text to Sign Language (SL) glosses. In our experiments, we improve the performance of the transformer-based models via (1) data augmentation, (2) semi-supervised Neural Machine Translation (NMT), (3) transfer learning and (4) multilingual NMT. The proposed methods are implemented progressively on two German SL corpora containing gloss annotations. Multilingual NMT combined with data augmentation appear to be the most successful setting, yielding statistically significant improvements as measured by three automatic metrics (up to over 6 points BLEU), and confirmed via human evaluation. Our best setting outperforms all previous work that report on the same test-set and is also confirmed on a corpus of the American Sign Language (ASL).

Parameter-efficient tuning (PET) methods fit pre-trained language models (PLMs) to downstream tasks by either computing a small compressed update for a subset of model parameters, or appending and fine-tuning a small number of new model parameters to the pre-trained network. Hand-designed PET architectures from the literature perform well in practice, but have the potential to be improved via automated neural architecture search (NAS). We propose an efficient NAS method for learning PET architectures via structured and unstructured pruning. We present experiments on GLUE demonstrating the effectiveness of our algorithm and discuss how PET architectural design choices affect performance in practice.

Word-level saliency explanations (“heat maps over words”) are often used to communicate feature-attribution in text-based models. Recent studies found that superficial factors such as word length can distort human interpretation of the communicated saliency scores. We conduct a user study to investigate how the marking of a word’s *neighboring words* affect the explainee’s perception of the word’s importance in the context of a saliency explanation. We find that neighboring words have significant effects on the word’s importance rating. Concretely, we identify that the influence changes based on neighboring direction (left vs. right) and a-priori linguistic and computational measures of phrases and collocations (vs. unrelated neighboring words).Our results question whether text-based saliency explanations should be continued to be communicated at word level, and inform future research on alternative saliency explanation methods.

Negation in natural language may affect many NLP applications, e.g., information extraction and sentiment analysis. The key sub-task of negation detection is negation scope resolution which aims to extract the portion of a sentence that is being negated by a negation cue (e.g., keyword “not” and never”) in the sentence. Due to the long spans, existing methods tend to make wrong predictions around the scope boundaries. In this paper, we propose a simple yet effective model named R-BSL which engages the Boundary Shift Loss to refine the predicted boundary. On multiple benchmark datasets, we show that the extremely simple R-BSL achieves best results.

Computational notebooks, such as Jupyter notebooks, are interactive computing environments that are ubiquitous among data scientists to perform data wrangling and analytic tasks. To measure the performance of AI pair programmers that automatically synthesize programs for those tasks given natural language (NL) intents from users, we build ARCADE, a benchmark of 1078 code generation problems using the pandas data analysis framework in data science notebooks. ARCADE features multiple rounds of NL-to-code problems from the same notebook. It requires a model to understand rich multi-modal contexts, such as existing notebook cells and their execution states as well as previous turns of interaction. To establish a strong baseline on this challenging task, we develop PaChiNCo, a 62B code language model (LM) for Python computational notebooks, which significantly outperforms public code LMs. Finally, we explore few-shot prompting strategies to elicit better code with step-by-step decomposition and NL explanation, showing the potential to improve the diversity and explainability of model predictions. Arcade is publicly available at https://github.com/google-research/arcade-nl2code/.

Reasoning has been a central topic in artificial intelligence from the beginning. The recent progress made on distributed representation and neural networks continues to improve the state-of-the-art performance of natural language inference. However, it remains an open question whether the models perform real reasoning to reach their conclusions or rely on spurious correlations. Adversarial attacks have proven to be an important tool to help evaluate the Achilles’ heel of the victim models. In this study, we explore the fundamental problem of developing attack models based on logic formalism. We propose NatLogAttack to perform systematic attacks centring around natural logic, a classical logic formalism that is traceable back to Aristotle’s syllogism and has been closely developed for natural language inference. The proposed framework renders both label-preserving and label-flipping attacks. We show that compared to the existing attack models, NatLogAttack generates better adversarial examples with fewer visits to the victim models. The victim models are found to be more vulnerable under the label-flipping setting. NatLogAttack provides a tool to probe the existing and future NLI models’ capacity from a key viewpoint and we hope more logic-based attacks will be further explored for understanding the desired property of reasoning.

Despite growing interest in applications based on natural customer support conversations,there exist remarkably few publicly available datasets that reflect the expected characteristics of conversations in these settings. Existing task-oriented dialogue datasets, which were collected to benchmark dialogue systems mainly in written human-to-bot settings, are not representative of real customer support conversations and do not provide realistic benchmarks for systems that are applied to natural data. To address this gap, we introduce NatCS, a multi-domain collection of spoken customer service conversations. We describe our process for collecting synthetic conversations between customers and agents based on natural language phenomena observed in real conversations. Compared to previous dialogue datasets, the conversations collected with our approach are more representative of real human-to-human conversations along multiple metrics. Finally, we demonstrate potential uses of NatCS, including dialogue act classification and intent induction from conversations as potential applications, showing that dialogue act annotations in NatCS provide more effective training data for modeling real conversations compared to existing synthetic written datasets. We publicly release NatCS to facilitate research in natural dialog systems

Large-scale language model pretraining is a very successful form of self-supervised learning in natural language processing, but it is increasingly expensive to perform as the models and pretraining corpora have become larger over time. We propose NarrowBERT, a modified transformer encoder that increases the throughput for masked language model pretraining by more than 2x. NarrowBERT sparsifies the transformer model such that the self-attention queries and feedforward layers only operate on the masked tokens of each sentence during pretraining, rather than all of the tokens as with the usual transformer encoder. We also show that NarrowBERT increases the throughput at inference time by as much as 3.5x with minimal (or no) performance degradation on sentence encoding tasks like MNLI. Finally, we examine the performance of NarrowBERT on the IMDB and Amazon reviews classification and CoNLL NER tasks and show that it is also comparable to standard BERT performance.

Pretrained language models have demonstrated extraordinary capabilities in language generation. However, real-world tasks often require controlling the distribution of generated text in order to mitigate bias, promote fairness, and achieve personalization. Existing techniques for controlling the distribution of generated text only work with quantified distributions, which require pre-defined categories, proportions of the distribution, or an existing corpus following the desired distributions. However, many important distributions, such as personal preferences, are unquantified. In this work, we tackle the problem of generating text following arbitrary distributions (quantified and unquantified) by proposing NANO, a few-shot human-in-the-loop training algorithm that continuously learns from human feedback. NANO achieves state-of-the-art results on single topic/attribute as well as quantified distribution control compared to previous works. We also show that NANO is able to learn unquantified distributions, achieves personalization, and captures differences between different individuals’ personal preferences with high sample efficiency.

We present, Naamapadam, the largest publicly available Named Entity Recognition (NER) dataset for the 11 major Indian languages from two language families. The dataset contains more than 400k sentences annotated with a total of at least 100k entities from three standard entity categories (Person, Location, and, Organization) for 9 out of the 11 languages. The training dataset has been automatically created from the Samanantar parallel corpus by projecting automatically tagged entities from an English sentence to the corresponding Indian language translation. We also create manually annotated testsets for 9 languages. We demonstrate the utility of the obtained dataset on the Naamapadam-test dataset. We also release IndicNER, a multilingual IndicBERT model fine-tuned on Naamapadam training set. IndicNER achieves an F1 score of more than 80 for 7 out of 9 test languages. The dataset and models are available under open-source licences at https://ai4bharat.iitm.ac.in/naamapadam.