ConvoSumm: Conversation Summarization Benchmark and Improved Abstractive Summarization with Argument Mining

Fabbri, Alexander R.; Rahman, Faiaz; Rizvi, Imad; Wang, Borui; Li, Haoran; Mehdad, Yashar; Radev, Dragomir

doi:10.18653/v1/2021.acl-long.535

Cited by 22 publications

(21 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previous works widely explore dialogue features explicitly and input them as known labels to enhance the dialogue understanding ability of summarization models. Features, including dialogue acts (Goo and Chen, 2018), topic transitions (Chen and Yang, 2020), discourse dependencies , coreference relations , argument graphs (Fabbri et al, 2021), semantic structures or slots (Lei et al, 2021;, etc. are carefully designed and collected by transferring tools pre-trained on other corpus or unsupervised methods with multiple hyper-parameters.…”

Section: A Related Workmentioning

confidence: 99%

Post-Training Dialogue Summarization using Pseudo-Paraphrasing

Jia¹,

Liu²,

Tang³

et al. 2022

Preprint

View full text Add to dashboard Cite

Previous dialogue summarization techniques adapt large language models pretrained on the narrative text by injecting dialogue-specific features into the models. These features either require additional knowledge to recognize or make the resulting models harder to tune. To bridge the format gap between dialogues and narrative summaries in dialogue summarization tasks, we propose to post-train pretrained language models (PLMs) to rephrase from dialogue to narratives. After that, the model is fine-tuned for dialogue summarization as usual. Comprehensive experiments show that our approach significantly improves vanilla PLMs on dialogue summarization and outperforms other SOTA models by the summary quality and implementation costs. 1

show abstract

Section: A Related Workmentioning

confidence: 99%

Post-Training Dialogue Summarization using Pseudo-Paraphrasing

Jia¹,

Liu²,

Tang³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…HMNet and HAT-BART (Rohde et al, 2021) leverage a two-level transformer-based model to obtain word level and sentence level representations. DialLM (Zhong et al, 2021a), Longformer-BART-arg (Fabbri et al, 2021) use Transformer models to incorporate the external knowledge while maintaining the accuracy of lengthy input via fine tuning or data augmentation.…”

Section: Related Workmentioning

confidence: 99%

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

Zhang¹,

Ni²,

Mao³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

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 QM-Sum, two long TV series datasets from Summ-Screen, and a newly proposed long document summarization dataset GovReport. Our data and code are available at https://github. com/chatc/Summ-N.

show abstract

“…Meeting such as AMI (Carletta et al 2005), MediaSum (Zhu et al 2021b), and QMSum (Zhong et al 2021) focus on the meeting scenario involving multiple people and topics. Multi-turn QA like CQASumm (Chowdhury and Chakraborty 2019) and ConvoSumm (Fabbri et al 2021) instead summarize user questions and system answers like Quora. Customer-service dialogues Yuan and Yu 2019;Zou et al 2021a,b) are recently proposed to address user issues about specific topics, such as the E-commerce after-sales service.…”

Section: Dialoguementioning

confidence: 99%

“…Since the meeting is essentially different from the dialogue in the language pattern, researchers further propose some practical conversational scenario datasets. They are either about the open-domain chitchats (Gliwa et al 2019; or the multi/single turn QA scenarios (Chowdhury and Chakraborty 2019;Fabbri et al 2021). Additionally, the studies on the customer-service dialogue summarization datasets, which is about the after-sales service or counseling, have attracted more attention and they belong to a taskoriented dialogue system.…”

Section: Related Workmentioning

confidence: 99%

TODSum: Task-Oriented Dialogue Summarization with State Tracking

Zhao¹,

Zheng²,

He³

et al. 2021

Preprint

View full text Add to dashboard Cite

Previous dialogue summarization datasets mainly focus on open-domain chitchat dialogues, while summarization datasets for the broadly used task-oriented dialogue haven't been explored yet. Automatically summarizing such taskoriented dialogues can help a business collect and review needs to improve the service. Besides, previous datasets pay more attention to generate good summaries with higher ROUGE scores, but they hardly understand the structured information of dialogues and ignore the factuality of summaries. In this paper, we introduce a large-scale public Task-Oriented Dialogue Summarization dataset, TODSum, which aims to summarize the key points of the agent completing certain tasks with the user. Compared to existing work, TODSum suffers from severe scattered information issues and requires strict factual consistency, which makes it hard to directly apply recent dialogue summarization models. Therefore, we introduce additional dialogue state knowledge for TODSum to enhance the faithfulness of generated summaries. We hope a better understanding of conversational content helps summarization models generate concise and coherent summaries. Meanwhile, we establish a comprehensive benchmark for TODSum and propose a state-aware structured dialogue summarization model to integrate dialogue state information and dialogue history. Exhaustive experiments and qualitative analysis prove the effectiveness of dialogue structure guidance. Finally, we discuss the current issues of TODSum and potential development directions for future work.

show abstract

ConvoSumm: Conversation Summarization Benchmark and Improved Abstractive Summarization with Argument Mining

Cited by 22 publications

References 43 publications

Post-Training Dialogue Summarization using Pseudo-Paraphrasing

Post-Training Dialogue Summarization using Pseudo-Paraphrasing

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

TODSum: Task-Oriented Dialogue Summarization with State Tracking

Contact Info

Product

Resources

About