Abstractive Meeting Summarization: A Survey (2022-08-08T00:00:00.000000Z)

TL;DR

An overview of the challenges raised by the task of abstractive meeting summarization and of the data sets, models, and evaluation metrics that have been used to tackle the problems is provided.

Abstract

Abstract A system that could reliably identify and sum up the most important points of a conversation would be valuable in a wide variety of real-world contexts, from business meetings to medical consultations to customer service calls. Recent advances in deep learning, and especially the invention of encoder-decoder architectures, has significantly improved language generation systems, opening the door to improved forms of abstractive summarization—a form of summarization particularly well-suited for multi-party conversation. In this paper, we provide an overview of the challenges raised by the task of abstractive meeting summarization and of the data sets, models, and evaluation metrics that have been used to tackle the problems.

Authors

M. Vazirgiannis

8 papers

Virgile Rennard

1 papers

Guokan Shang

3 papers

TL;DR

Abstract

Authors

References158 items

BLOOM: A 176B-Parameter Open-Access Multilingual Language Model

Questioning the Validity of Summarization Datasets and Improving Their Factual Consistency

DATScore: Evaluating Translation with Data Augmented Translations

News Summarization and Evaluation in the Era of GPT-3

Unsupervised Abstractive Dialogue Summarization with Word Graphs and POV Conversion

PaLM: Scaling Language Modeling with Pathways

Training language models to follow instructions with human feedback

Symbolic and Statistical Learning Approaches to Speech Summarization: A Scoping Review

CONFIT: Toward Faithful Dialogue Summarization with Linguistically-Informed Contrastive Fine-tuning

Scaling Language Models: Methods, Analysis & Insights from Training Gopher

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

FrugalScore: Learning Cheaper, Lighter and Faster Evaluation Metrics for Automatic Text Generation

Multitask Prompted Training Enables Zero-Shot Task Generalization

Who speaks like a style of Vitamin: Towards Syntax-Aware Dialogue Summarization using Multi-task Learning

Low-Resource Dialogue Summarization with Domain-Agnostic Multi-Source Pretraining

DialogLM: Pre-trained Model for Long Dialogue Understanding and Summarization

Finetuned Language Models Are Zero-Shot Learners

A Survey on Dialogue Summarization: Recent Advances and New Frontiers

BARTScore: Evaluating Generated Text as Text Generation

Improving Unsupervised Dialogue Topic Segmentation with Utterance-Pair Coherence Scoring

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

Language Model as an Annotator: Exploring DialoGPT for Dialogue Summarization

The Factual Inconsistency Problem in Abstractive Text Summarization: A Survey

A Sliding-Window Approach to Automatic Creation of Meeting Minutes

DADgraph: A Discourse-aware Dialogue Graph Neural Network for Multiparty Dialogue Machine Reading Comprehension

OVERLOADING IN LOCKDOWN: EFFECTS OF SOCIAL, INFORMATION AND COMMUNICATION OVERLOADS IN COVID-19 DAYS

Hierarchical Learning for Generation with Long Source Sequences

QMSum: A New Benchmark for Query-based Multi-domain Meeting Summarization

Zoom Exhaustion & Fatigue Scale

CSKG: The CommonSense Knowledge Graph

Dialogue Discourse-Aware Graph Model and Data Augmentation for Meeting Summarization

Metrics also Disagree in the Low Scoring Range: Revisiting Summarization Evaluation Metrics

How Domain Terminology Affects Meeting Summarization Performance

BARThez: a Skilled Pretrained French Sequence-to-Sequence Model

Incorporating Commonsense Knowledge into Abstractive Dialogue Summarization via Heterogeneous Graph Networks

Re-evaluating Evaluation in Text Summarization

Multi-View Sequence-to-Sequence Models with Conversational Structure for Abstractive Dialogue Summarization

Learning to summarize from human feedback

A Survey of Evaluation Metrics Used for NLG Systems

Language Models are Few-Shot Learners

Generating SOAP Notes from Doctor-Patient Conversations Using Modular Summarization Techniques

Don’t Stop Pretraining: Adapt Language Models to Domains and Tasks

Longformer: The Long-Document Transformer

Speaker-change Aware CRF for Dialogue Act Classification

A Hierarchical Network for Abstractive Meeting Summarization with Cross-Domain Pretraining

Fill in the BLANC: Human-free quality estimation of document summaries

SAMSum Corpus: A Human-annotated Dialogue Dataset for Abstractive Summarization

DIALOGPT : Large-Scale Generative Pre-training for Conversational Response Generation

BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension

Evaluating the Factual Consistency of Abstractive Text Summarization

Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer

Topic-Aware Pointer-Generator Networks for Summarizing Spoken Conversations

CTRL: A Conditional Transformer Language Model for Controllable Generation

Neural Text Summarization: A Critical Evaluation

MoverScore: Text Generation Evaluating with Contextualized Embeddings and Earth Mover Distance

Automatic Dialogue Summary Generation for Customer Service

Exploring Methods for Predicting Important Utterances Contributing to Meeting Summarization

Discourse Analysis and Its Applications

Keep Meeting Summaries on Topic: Abstractive Multi-Modal Meeting Summarization

Studying Summarization Evaluation Metrics in the Appropriate Scoring Range

Abstractive Meeting Summarization via Hierarchical Adaptive Segmental Network Learning

BERTScore: Evaluating Text Generation with BERT

Energy-based Self-attentive Learning of Abstractive Communities for Spoken Language Understanding

Restructuring Conversations using Discourse Relations for Zero-shot Abstractive Dialogue Summarization.

Transformer-XL: Attentive Language Models beyond a Fixed-Length Context

A Deep Sequential Model for Discourse Parsing on Multi-Party Dialogues

Do We Really Need Another Meeting? The Science of Workplace Meetings

Fusing Verbal and Nonverbal Information for Extractive Meeting Summarization

Abstractive Dialogue Summarization with Sentence-Gated Modeling Optimized by Dialogue Acts

Unsupervised Abstractive Sentence Summarization using Length Controlled Variational Autoencoder

Commonsense Knowledge Aware Conversation Generation with Graph Attention

Unsupervised Abstractive Meeting Summarization with Multi-Sentence Compression and Budgeted Submodular Maximization

Faithful to the Original: Fact Aware Neural Abstractive Summarization

Graph Attention Networks

Combining Graph Degeneracy and Submodularity for Unsupervised Extractive Summarization

Attention is All you Need