Text Summarization
文本摘要(Text Summarization)是自然语言处理的一项任务,旨在将长篇文档压缩成更简短精炼的版本,同时保留原文的核心信息与意义。其目标是生成能够准确反映原始内容的概要,以便用户快速获取关键信息。该任务包括抽取式方法和生成式方法,前者通过识别并提取重要句子或短语,后者则基于原文内容生成新的文本。文本摘要在新闻报道、科研文献、商业报告等领域具有重要应用价值。
ACI-Bench
CriSPO 3-shot
AMI
Arxiv HEP-TH citation graph
arXiv
BigBird-Pegasus
arXiv Summarization Dataset
PRIMER
BBC XSum
MatchSum
BigPatent
LongT5
BillSum
Longformer Encoder Decoder
BookSum
Echoes-Extractive-Abstractive
CL-SciSumm
CNN / Daily Mail (Anonymized)
CORD-19
DialogSum
InstructDS
DUC 2004 Task 1
Transformer+WDrop
EurekaAlert
Gazeta
Finetuned mBART
GigaWord
BART-RXF
GigaWord-10k
ERNIE-GENLARGE (large-scale text corpora)
GovReport
FactorSum
How2
Klexikon
Luhn's algorithm (25 sentences)
LCSTS
LSTM-seq2seq
MediaSum
SRformer-BART
MeetingBank
MentSum
MeQSum
BiomedGPT
MTEB
OrangeSum
mBARThez (OrangeSum abstract)
Pubmed
QMSum
BART-LS
Reddit TIFU
S2ORC
GenCompareSum
SAMSum
Webis-Snippet-20 Corpus
Anchor-context + Query biased
WikiHow
BertSum
X-Sum
Selfmem
XSum
SRformer-BART