Why Do You Feel This Way? Summarizing Triggers of Emotions in Social Media Posts

Crises such as the COVID-19 pandemic continuously threaten our world and emotionally affect billions of people worldwide in distinct ways. Understanding the triggers leading to people's emotions is of crucial importance. Social media posts can be a good source of such analysis, yet these texts tend to be charged with multiple emotions, with triggers scattering across multiple sentences. This paper takes a novel angle, namely, emotion detection and trigger summarization, aiming to both detect perceived emotions in text, and summarize events and their appraisals that trigger each emotion. To support this goal, we introduce CovidET (Emotions and their Triggers during Covid-19), a dataset of ~1,900 English Reddit posts related to COVID-19, which contains manual annotations of perceived emotions and abstractive summaries of their triggers described in the post. We develop strong baselines to jointly detect emotions and summarize emotion triggers. Our analyses show that CovidET presents new challenges in emotion-specific summarization, as well as multi-emotion detection in long social media posts.

A Data Cartography based MixUp for Pre-trained Language Models

MixUp is a data augmentation strategy where additional samples are generated during training by combining random pairs of training samples and their labels. However, selecting random pairs is not potentially an optimal choice. In this work, we propose TDMixUp, a novel MixUp strategy that leverages Training Dynamics and allows more informative samples to be combined for generating new data samples. Our proposed TDMixUp first measures confidence, variability, (Swayamdipta et al., 2020), and Area Under the Margin (AUM) (Pleiss et al., 2020) to identify the characteristics of training samples (e.g., as easy-to-learn or ambiguous samples), and then interpolates these characterized samples. We empirically validate that our method not only achieves competitive performance using a smaller subset of the training data compared with strong baselines, but also yields lower expected calibration error on the pre-trained language model, BERT, on both in-domain and out-of-domain settings in a wide range of NLP tasks. We publicly release our code.

SciNLI: A Corpus for Natural Language Inference on Scientific Text

Existing Natural Language Inference (NLI) datasets, while being instrumental in the advancement of Natural Language Understanding (NLU) research, are not related to scientific text. In this paper, we introduce SciNLI, a large dataset for NLI that captures the formality in scientific text and contains 107,412 sentence pairs extracted from scholarly papers on NLP and computational linguistics. Given that the text used in scientific literature differs vastly from the text used in everyday language both in terms of vocabulary and sentence structure, our dataset is well suited to serve as a benchmark for the evaluation of scientific NLU models. Our experiments show that SciNLI is harder to classify than the existing NLI datasets. Our best performing model with XLNet achieves a Macro F1 score of only 78.18% and an accuracy of 78.23% showing that there is substantial room for improvement.

On the Calibration of Pre-trained Language Models using Mixup Guided by Area Under the Margin and Saliency

A well-calibrated neural model produces confidence (probability outputs) closely approximated by the expected accuracy. While prior studies have shown that mixup training as a data augmentation technique can improve model calibration on image classification tasks, little is known about using mixup for model calibration on natural language understanding (NLU) tasks. In this paper, we explore mixup for model calibration on several NLU tasks and propose a novel mixup strategy for pre-trained language models that improves model calibration further. Our proposed mixup is guided by both the Area Under the Margin (AUM) statistic (Pleiss et al., 2020) and the saliency map of each sample (Simonyan et al.,2013). Moreover, we combine our mixup strategy with model miscalibration correction techniques (i.e., label smoothing and temperature scaling) and provide detailed analyses of their impact on our proposed mixup. We focus on systematically designing experiments on three NLU tasks: natural language inference, paraphrase detection, and commonsense reasoning. Our method achieves the lowest expected calibration error compared to strong baselines on both in-domain and out-of-domain test samples while maintaining competitive accuracy.

On the Evaluation of Answer-Agnostic Paragraph-level Multi-Question Generation

We study the task of predicting a set of salient questions from a given paragraph without any prior knowledge of the precise answer. We make two main contributions. First, we propose a new method to evaluate a set of predicted questions against the set of references by using the Hungarian algorithm to assign predicted questions to references before scoring the assigned pairs. We show that our proposed evaluation strategy has better theoretical and practical properties compared to prior methods because it can properly account for the coverage of references. Second, we compare different strategies to utilize a pre-trained seq2seq model to generate and select a set of questions related to a given paragraph. The code is available.

Keyphrase Generation Beyond the Boundaries of Title and Abstract

Keyphrase generation aims at generating phrases (keyphrases) that best describe a given document. In scholarly domains, current approaches to this task are neural approaches and have largely worked with only the title and abstract of the articles. In this work, we explore whether the integration of additional data from semantically similar articles or from the full text of the given article can be helpful for a neural keyphrase generation model. We discover that adding sentences from the full text particularly in the form of summary of the article can significantly improve the generation of both types of keyphrases that are either present or absent from the title and abstract. The experimental results on the three acclaimed models along with one of the latest transformer models suitable for longer documents, Longformer Encoder-Decoder (LED) validate the observation. We also present a new large-scale scholarly dataset FullTextKP for keyphrase generation, which we use for our experiments. Unlike prior large-scale datasets, FullTextKP includes the full text of the articles alongside title and abstract. We will release the source code to stimulate research on the proposed ideas.

KPDrop: An Approach to Improving Absent Keyphrase Generation

Keyphrase generation is the task of generating phrases (keyphrases) that summarize the main topics of a given document. The generated keyphrases can be either present or absent from the text of the given document. While the extraction of present keyphrases has received much attention in the past, only recently a stronger focus has been placed on the generation of absent keyphrases. However, generating absent keyphrases is very challenging; even the best methods show only a modest degree of success. In this paper, we propose an approach, called keyphrase dropout (or KPDrop), to improve absent keyphrase generation. We randomly drop present keyphrases from the document and turn them into artificial absent keyphrases during training. We test our approach extensively and show that it consistently improves the absent performance of strong baselines in keyphrase generation.