QuestEval is a reference-less metric used in text-to-text tasks, that compares the generated summaries directly to the source text, by automatically asking and answering questions. Its adaptation to Data-to-Text tasks is not straightforward, as it requires multimodal Question Generation and Answering systems on the considered tasks, which are seldom available. To this purpose, we propose a method to build synthetic multimodal corpora enabling to train multimodal components for a data-QuestEval metric. The resulting metric is reference-less and multimodal; it obtains state-of-the-art correlations with human judgment on the WebNLG and WikiBio benchmarks. We make data-QuestEval's code and models available for reproducibility purpose, as part of the QuestEval project.

Data-QuestEval: A Referenceless Metric for Data to Text Semantic Evaluation

Models pre-trained with a language modeling objective possess ample world knowledge and language skills, but are known to struggle in tasks that require reasoning. In this work, we propose to leverage semi-structured tables, and automatically generate at scale question-paragraph pairs, where answering the question requires reasoning over multiple facts in the paragraph. We add a pre-training step over this synthetic data, which includes examples that require 16 different reasoning skills such as number comparison, conjunction, and fact composition. To improve data efficiency, we propose sampling strategies that focus training on reasoning skills the model is currently lacking. We evaluate our approach on three reading comprehension datasets that are focused on reasoning, and show that our model, PReasM, substantially outperforms T5, a popular pre-trained encoder-decoder model. Moreover, sampling examples based on current model errors leads to faster training and higher overall performance.

Turning Tables: Generating Examples from Semi-structured Tables for Endowing Language Models with Reasoning Skills

Long document summarization poses obstacles to current generative transformer-based models because of the broad context to process and understand. Indeed, detecting long-range dependencies is still challenging for today’s state-of-the-art solutions, usually requiring model expansion at the cost of an unsustainable demand for computing and memory capacities. This paper introduces Emma, a novel efficient memory-enhanced transformer-based architecture. By segmenting a lengthy input into multiple text fragments, our model stores and compares the current chunk with previous ones, gaining the capability to read and comprehend the entire context over the whole document with a fixed amount of GPU memory. This method enables the model to deal with theoretically infinitely long documents, using less than 18 and 13 GB of memory for training and inference, respectively. We conducted extensive performance analyses and demonstrate that Emma achieved competitive results on two datasets of different domains while consuming significantly less GPU memory than competitors do, even in low-resource settings.

https://www.mdpi.com/1424-8220/23/7/3542/pdf?version=1680073054

Efficient Memory-Enhanced Transformer for Long-Document Summarization in Low-Resource Regimes

Models pre-trained with a language modeling objective possess ample world knowledge and language skills, but are known to struggle in tasks that require reasoning. In this work, we propose to leverage semi-structured tables, and automatically generate at scale question-paragraph pairs, where answering the question requires reasoning over multiple facts in the paragraph. We add a pre-training step over this synthetic data, which includes examples that require 16 different reasoning skills such as number comparison, conjunction, and fact composition. To improve data efficiency, we sample examples from reasoning skills where the model currently errs. We evaluate our approach on three reasoning-focused reading comprehension datasets, and show that our model, PReasM, substantially outperforms T5, a popular pre-trained encoder-decoder model. Moreover, sampling examples based on model errors leads to faster training and higher performance. 

Query-focused meeting summarization (QFMS) aims to generate suimnaries from meeting transcripts in response to a given query. Previous works typically concatenate the query with meeting transcripts and implicitly model the query relevance only at the token level with attention mechanism. However, due to the dilution of key query-relevant infonnation caused by long meeting transcripts, the original transformer-based model is insufficient to highlight the key parts related to the query. In this paper, we propose a query-aware framework with joint modeling token and utterance based on Query-Utterance Attention. It calculates the utterance-level relevance to the query with a dense retrieval module. Then both token-level query relevance and utterance-level query relevance are combined and incorporated into the generation process with attention mechanism explicitly. We show that the query relevance of different granularities contributes to generating a suimnary more related to the query. Experimental results on the QMSum dataset show that the proposed model achieves new state-of-the-art perfonnance. 

Query-Utterance Attention With Joint Modeuing For Query-Focused Meeting Summarization

Transformer-based models have achieved state-of-the-art performance on short text summarization. However, they still struggle with long-input summarization. In this paper, we present a new approach for long-input summarization: Dynamic Latent Extraction for Abstractive Summarization. We jointly train an extractor with an abstractor and treat the extracted text snippets as the latent variable. We propose extractive oracles to provide the extractor with a strong learning signal. We introduce consistency loss, which encourages the extractor to approximate the averaged dynamic weights predicted by the generator. We conduct extensive tests on two long-input summarization datasets, GovReport (document) and QMSum (dialogue). Our model significantly outperforms the current state-of-the-art, including a 6.21 ROUGE-2 improvement on GovReport and a 2.13 ROUGE-1 improvement on QMSum. Further analysis shows that the dynamic weights make our generation process highly interpretable. Our code will be publicly available upon publication.

DYLE: Dynamic Latent Extraction for Abstractive Long-Input Summarization.

Existing table question answering datasets contain abundant factual questions that primarily evaluate the query and schema comprehension capability of a system, but they fail to include questions that require complex reasoning and integration of information due to the constraint of the associated short-form answers. To address these issues and to demonstrate the full challenge of table question answering, we introduce FeTaQA, a new dataset with 10K Wikipedia-based {table, question, free-form answer, supporting table cells} pairs. FeTaQA yields a more challenging table question answering setting because it requires generating free-form text answers after retrieval, inference, and integration of multiple discontinuous facts from a structured knowledge source. Unlike datasets of generative QA over text in which answers are prevalent with copies of short text spans from the source, answers in our dataset are human-generated explanations involving entities and their high-level relations. We provide two benchmark methods for the proposed task: a pipeline method based on semantic-parsing-based QA systems and an end-to-end method based on large pretrained text generation models, and show that FeTaQA poses a challenge for both methods.

FeTaQA: Free-form Table Question Answering.

Text summarization is an essential task to help readers capture salient information from documents, news, interviews, and meetings. However, most state-of-the-art pretrained language models are unable to efficiently process long text commonly seen in the summarization problem domain. In this paper, we propose Summ^N, a simple, flexible, and effective multi-stage framework for input texts that are longer than the maximum context lengths of typical pretrained LMs. Summ^N first generates the coarse summary in multiple stages and then produces the final fine-grained summary based on them. The framework can process input text of arbitrary length by adjusting the number of stages while keeping the LM context size fixed. Moreover, it can deal with both documents and dialogues and can be used on top of any underlying backbone abstractive summarization model. Our experiments demonstrate that Summ^N significantly outperforms previous state-of-the-art methods by improving ROUGE scores on three long meeting summarization datasets AMI, ICSI, and QMSum, two long TV series datasets from SummScreen, and a newly proposed long document summarization dataset GovReport. Our data and code are available at https://github.com/chatc/Summ-N.

Summ^N: A Multi-Stage Summarization Framework for Long Input Dialogues and Documents.

Current pre-trained models applied to summarization are prone to factual inconsistencies which either misrepresent the source text or introduce extraneous information. Thus, comparing the factual consistency of summaries is necessary as we develop improved models. However, the optimal human evaluation setup for factual consistency has not been standardized. To address this issue, we crowdsourced evaluations for factual consistency using the rating-based Likert scale and ranking-based Best-Worst Scaling protocols, on 100 articles from each of the CNN-Daily Mail and XSum datasets over four state-of-the-art models, to determine the most reliable evaluation framework. We find that ranking-based protocols offer a more reliable measure of summary quality across datasets, while the reliability of Likert ratings depends on the target dataset and the evaluation design. Our crowdsourcing templates and summary evaluations will be publicly available to facilitate future research on factual consistency in summarization.

Ziming Mao

Papers

DYLE: Dynamic Latent Extraction for Abstractive Long-Input Summarization.

FeTaQA: Free-form Table Question Answering.

Summ^N: A Multi-Stage Summarization Framework for Long Input Dialogues and Documents.

Investigating Crowdsourcing Protocols for Evaluatingthe Factual Consistency of Summaries

Investigating Crowdsourcing Protocols for Evaluating the Factual Consistency of Summaries.