With the development of large language models (LLMs), zero-shot learning has attracted much attention for various NLP tasks. Different from prior works that generate training data with billion-scale natural language generation (NLG) models, we propose a retrieval-enhanced framework to create training data from a general-domain unlabeled corpus. To realize this, we first conduct contrastive pretraining to learn an unsupervised dense retriever for extracting the most relevant documents using class-descriptive verbalizers. We then further propose two simple strategies, namely Verbalizer Augmentation with Demonstrations and Self-consistency Guided Filtering to improve the topic coverage of the dataset while removing noisy examples. Experiments on nine datasets demonstrate that REGEN achieves 4.3% gain over the strongest baselines and saves around 70% of the time compared to baselines using large NLG models. Besides, REGEN can be naturally integrated with recently proposed large language models to boost performance.
Listening to long video/audio recordings from video conferencing and online courses for acquiring information is extremely inefficient. Even after ASR systems transcribe recordings into long-form spoken language documents, reading ASR transcripts only partly speeds up seeking information. It has been observed that a range of NLP applications, such as keyphrase extraction, topic segmentation, and summarization, significantly improve users' efficiency in grasping important information. The meeting scenario is among the most valuable scenarios for deploying these spoken language processing (SLP) capabilities. However, the lack of large-scale public meeting datasets annotated for these SLP tasks severely hinders their advancement. To prompt SLP advancement, we establish a large-scale general Meeting Understanding and Generation Benchmark (MUG) to benchmark the performance of a wide range of SLP tasks, including topic segmentation, topic-level and session-level extractive summarization and topic title generation, keyphrase extraction, and action item detection. To facilitate the MUG benchmark, we construct and release a large-scale meeting dataset for comprehensive long-form SLP development, the AliMeeting4MUG Corpus, which consists of 654 recorded Mandarin meeting sessions with diverse topic coverage, with manual annotations for SLP tasks on manual transcripts of meeting recordings. To the best of our knowledge, the AliMeeting4MUG Corpus is so far the largest meeting corpus in scale and facilitates most SLP tasks. In this paper, we provide a detailed introduction of this corpus, SLP tasks and evaluation methods, baseline systems and their performance.
Crowdsourced dialogue corpora are usually limited in scale and topic coverage due to the expensive cost of data curation. This would hinder the generalization of downstream dialogue models to open-domain topics. In this work, we leverage large language models for dialogue augmentation in the task of emotional support conversation (ESC). By treating dialogue augmentation as a dialogue completion task, we prompt a fine-tuned language model to complete full dialogues from available dialogue posts of various topics, which are then postprocessed based on heuristics. Applying this approach, we construct AugESC, an augmented dataset for the ESC task, which largely extends the scale and topic coverage of the crowdsourced ESConv corpus. Through comprehensive human evaluation, we demonstrate that our approach is superior to strong baselines of dialogue augmentation and that AugESC has comparable dialogue quality to the crowdsourced corpus. We also conduct human interactive evaluation and prove that post-training on AugESC improves downstream dialogue models' generalization ability to open-domain topics. These results suggest the utility of AugESC and highlight the potential of large language models in improving data-scarce dialogue generation tasks.
Automatic chat summarization can help people quickly grasp important information from numerous chat messages. Unlike conventional documents, chat logs usually have fragmented and evolving topics. In addition, these logs contain a quantity of elliptical and interrogative sentences, which make the chat summarization highly context dependent. In this work, we propose a novel unsupervised framework called RankAE to perform chat summarization without employing manually labeled data. RankAE consists of a topic-oriented ranking strategy that selects topic utterances according to centrality and diversity simultaneously, as well as a denoising auto-encoder that is carefully designed to generate succinct but context-informative summaries based on the selected utterances. To evaluate the proposed method, we collect a large-scale dataset of chat logs from a customer service environment and build an annotated set only for model evaluation. Experimental results show that RankAE significantly outperforms other unsupervised methods and is able to generate high-quality summaries in terms of relevance and topic coverage.
Topic models are widely used unsupervised models capable of learning topics - weighted lists of words and documents - from large collections of text documents. When topic models are used for discovery of topics in text collections, a question that arises naturally is how well the model-induced topics correspond to topics of interest to the analyst. In this paper we revisit and extend a so far neglected approach to topic model evaluation based on measuring topic coverage - computationally matching model topics with a set of reference topics that models are expected to uncover. The approach is well suited for analyzing models' performance in topic discovery and for large-scale analysis of both topic models and measures of model quality. We propose new measures of coverage and evaluate, in a series of experiments, different types of topic models on two distinct text domains for which interest for topic discovery exists. The experiments include evaluation of model quality, analysis of coverage of distinct topic categories, and the analysis of the relationship between coverage and other methods of topic model evaluation. The paper contributes a new supervised measure of coverage, and the first unsupervised measure of coverage. The supervised measure achieves topic matching accuracy close to human agreement. The unsupervised measure correlates highly with the supervised one (Spearman's $\rho \geq 0.95$). Other contributions include insights into both topic models and different methods of model evaluation, and the datasets and code for facilitating future research on topic coverage.