Text Summarization
Text summarization is a task in natural language processing that aims to compress long documents into shorter, more concise versions while retaining the core information and meaning of the original text. Its goal is to produce summaries that accurately reflect the original content, enabling users to quickly grasp key information. This task encompasses both extractive and abstractive methods; the former identifies and extracts important sentences or phrases, while the latter generates new text based on the content of the original document. Text summarization has significant application value in areas such as news reporting, scientific literature, and business reports.
GigaWord
BART-RXF
Pubmed
Arxiv HEP-TH citation graph
MTEB
X-Sum
Selfmem
CNN / Daily Mail (Anonymized)
DUC 2004 Task 1
Transformer+WDrop
SAMSum
Reddit TIFU
arXiv Summarization Dataset
PRIMER
DialogSum
InstructDS
Klexikon
Luhn's algorithm (25 sentences)
BookSum
Echoes-Extractive-Abstractive
GigaWord-10k
ERNIE-GENLARGE (large-scale text corpora)
WikiHow
BertSum
BigPatent
BigBird-Pegasus
GovReport
FactorSum
How2
MeetingBank
OrangeSum
mBARThez (OrangeSum abstract)
ACI-Bench
CriSPO 3-shot
AMI
arXiv
BigBird-Pegasus
BBC XSum
MatchSum
BillSum
Longformer Encoder Decoder
CL-SciSumm
CORD-19
EurekaAlert
Gazeta
Finetuned mBART
LCSTS
LSTM-seq2seq
MediaSum
SRformer-BART
MentSum
MeQSum
BiomedGPT
QMSum
BART-LS
S2ORC
GenCompareSum
Webis-Snippet-20 Corpus
Anchor-context + Query biased
XSum
SRformer-BART