Text Summarization
Text Summarization is a natural language processing (NLP) task that involves condensing a lengthy text document into a shorter, more compact version while still retaining the most important information and meaning. The goal is to produce a summary that accurately represents the content of the original text in a concise form.
There are different approaches to text summarization, including extractive methods that identify and extract important sentences or phrases from the text, and abstractive methods that generate new text based on the content of the original text.
Papers
Showing 1–10 of 1340 papers
All datasetsGigaWordPubmedArxiv HEP-TH citation graphX-SumCNN / Daily Mail (Anonymized)DUC 2004 Task 1SAMSumReddit TIFUarXiv Summarization DatasetDialogSumKlexikonBookSum
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | OpenAI/o3-mini | ROUGE-1 | 60.12 | — | Unverified |
| 2 | Riple/Saanvi-v0.1 | ROUGE-1 | 52.21 | — | Unverified |
| 3 | Pegasus+DotProd | ROUGE-1 | 40.6 | — | Unverified |
| 4 | BART-RXF | ROUGE-1 | 40.45 | — | Unverified |
| 5 | MUPPET BART Large | ROUGE-1 | 40.4 | — | Unverified |
| 6 | OFA | ROUGE-1 | 39.81 | — | Unverified |
| 7 | Transformer+Rep(Uni) | ROUGE-1 | 39.81 | — | Unverified |
| 8 | Transformer+Wdrop | ROUGE-1 | 39.66 | — | Unverified |
| 9 | ProphetNet | ROUGE-1 | 39.51 | — | Unverified |
| 10 | ERNIE-GENLARGE (large-scale text corpora) | ROUGE-1 | 39.46 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Top Down Transformer (AdaPool) (464M) | ROUGE-1 | 51.05 | — | Unverified |
| 2 | eyeglaxs | ROUGE-1 | 50.34 | — | Unverified |
| 3 | BART-LS | ROUGE-1 | 50.3 | — | Unverified |
| 4 | LongT5 | ROUGE-1 | 50.23 | — | Unverified |
| 5 | GoSum (extractive) | ROUGE-1 | 49.83 | — | Unverified |
| 6 | Lodoss-full-large (extractive) | ROUGE-1 | 49.38 | — | Unverified |
| 7 | MemSum (extractive) | ROUGE-1 | 49.25 | — | Unverified |
| 8 | Lodoss-full-base (extractive) | ROUGE-1 | 48.93 | — | Unverified |
| 9 | HAT-BART | ROUGE-1 | 48.25 | — | Unverified |
| 10 | GRETEL | ROUGE-1 | 48.2 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Top Down Transformer (AdaPool) (464M) | ROUGE-1 | 50.95 | — | Unverified |
| 2 | BART-LS | ROUGE-1 | 50.2 | — | Unverified |
| 3 | Pegasus-X | ROUGE-1 | 50 | — | Unverified |
| 4 | ExtSum + oracle segmentation (extractive) | ROUGE-1 | 49.49 | — | Unverified |
| 5 | FactorSum | ROUGE-1 | 49.32 | — | Unverified |
| 6 | ExtSum + supervised segmentation (extractive) | ROUGE-1 | 49.11 | — | Unverified |
| 7 | Lodoss-full-large (extractive) | ROUGE-1 | 48.45 | — | Unverified |
| 8 | MemSum (extractive) | ROUGE-1 | 48.42 | — | Unverified |
| 9 | LongT5 | ROUGE-1 | 48.35 | — | Unverified |
| 10 | Lodoss-full-base (extractive) | ROUGE-1 | 48.2 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Selfmem | ROUGE-1 | 50.3 | — | Unverified |
| 2 | BRIO | ROUGE-1 | 49.07 | — | Unverified |
| 3 | PEGASUS + SummaReranker | ROUGE-1 | 48.12 | — | Unverified |
| 4 | PEGASUS + SimCLS | ROUGE-1 | 47.61 | — | Unverified |
| 5 | PEGASUSLARGE | ROUGE-1 | 47.21 | — | Unverified |
| 6 | HAT-BART | ROUGE-1 | 45.92 | — | Unverified |
| 7 | BART | ROUGE-1 | 45.14 | — | Unverified |
| 8 | BertSumExtAbs | ROUGE-1 | 38.81 | — | Unverified |
| 9 | T-ConvS2S | ROUGE-1 | 31.89 | — | Unverified |
| 10 | Baseline : Extractive Oracle | ROUGE-1 | 29.79 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | HSSAS | ROUGE-1 | 42.3 | — | Unverified |
| 2 | SWAP-NET | ROUGE-1 | 41.6 | — | Unverified |
| 3 | RNES w/o coherence | ROUGE-1 | 41.25 | — | Unverified |
| 4 | ML+RL ROUGE+Novel, with LM | ROUGE-1 | 40.02 | — | Unverified |
| 5 | GAN | ROUGE-1 | 39.92 | — | Unverified |
| 6 | ML+RL, with intra-attention | ROUGE-1 | 39.87 | — | Unverified |
| 7 | rnn-ext + abs + RL + rerank | ROUGE-1 | 39.66 | — | Unverified |
| 8 | SummaRuNNer | ROUGE-1 | 39.6 | — | Unverified |
| 9 | Lead-3 baseline | ROUGE-1 | 39.2 | — | Unverified |
| 10 | KIGN+Prediction-guide | ROUGE-1 | 38.95 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Transformer+WDrop | ROUGE-1 | 33.06 | — | Unverified |
| 2 | Transformer+LRPE+PE+Re-ranking+Ensemble | ROUGE-1 | 32.85 | — | Unverified |
| 3 | Transformer+LRPE+PE+ALONE+Re-ranking | ROUGE-1 | 32.57 | — | Unverified |
| 4 | EndDec+WFE | ROUGE-1 | 32.28 | — | Unverified |
| 5 | DRGD | ROUGE-1 | 31.79 | — | Unverified |
| 6 | Reinforced-Topic-ConvS2S | ROUGE-1 | 31.15 | — | Unverified |
| 7 | Seq2seq + selective + MTL + ERAM | ROUGE-1 | 29.33 | — | Unverified |
| 8 | SEASS | ROUGE-1 | 29.21 | — | Unverified |
| 9 | RAS-Elman | ROUGE-1 | 28.97 | — | Unverified |
| 10 | words-lvt5k-1sent | ROUGE-1 | 28.61 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | OmniVec2 | ROUGE-1 | 59.1 | — | Unverified |
| 2 | InstructDS | ROUGE-1 | 55.3 | — | Unverified |
| 3 | MoCa | ROUGE-1 | 55.13 | — | Unverified |
| 4 | PEGASUS 2B + SliC | ROUGE-1 | 54.37 | — | Unverified |
| 5 | ConDigSum | ROUGE-1 | 54.3 | — | Unverified |
| 6 | SICK | ROUGE-1 | 53.73 | — | Unverified |
| 7 | HAT-CNNDM | ROUGE-1 | 53.01 | — | Unverified |
| 8 | BART-Large | ROUGE-1 | 52.6 | — | Unverified |
| 9 | HAT-CNNDM RL | ROUGE-L | 48.84 | — | Unverified |
| 10 | CriSPO 3-shot | ROUGE-1 | 47.2 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | PEGASUS 2B + SLiC | ROUGE-1 | 32.03 | — | Unverified |
| 2 | BART+R3F | ROUGE-1 | 30.31 | — | Unverified |
| 3 | MUPPET BART Large | ROUGE-1 | 30.3 | — | Unverified |
| 4 | PEGASUS + SummaReranker | ROUGE-1 | 29.83 | — | Unverified |
| 5 | MatchSum | ROUGE-1 | 25.09 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | PRIMER | ROUGE-1 | 47.6 | — | Unverified |
| 2 | DeepPyramidion | ROUGE-1 | 47.15 | — | Unverified |
| 3 | Blockwise (baseline) | ROUGE-1 | 46.85 | — | Unverified |
| 4 | Claude Instant + SigExt | ROUGE-1 | 45.2 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | InstructDS | Rouge1 | 47.8 | — | Unverified |
| 2 | OmniVec2 | Rouge1 | 47.6 | — | Unverified |
| 3 | OmniVec | Rouge1 | 46.91 | — | Unverified |
| 4 | SICK | Rouge1 | 46.26 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Luhn's algorithm (25 sentences) | ROUGE-1 | 32 | — | Unverified |
| 2 | Lead-k | ROUGE-1 | 25 | — | Unverified |
| 3 | Lead-3 | ROUGE-1 | 17.5 | — | Unverified |
| 4 | Full article | ROUGE-1 | 16.98 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Echoes-Extractive-Abstractive | ROUGE | 42.13 | — | Unverified |
| 2 | BART-LS | ROUGE | 38.5 | — | Unverified |
| 3 | Top Down Transformer (AdaPool) (464M) | ROUGE | 38.3 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | ERNIE-GENLARGE (large-scale text corpora) | ROUGE-L | 33.23 | — | Unverified |
| 2 | ERNIE-GENLARGE | ROUGE-L | 32.5 | — | Unverified |
| 3 | ERNIE-GENBASE | ROUGE-L | 31.35 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | BertSum | ROUGE-1 | 35.91 | — | Unverified |
| 2 | MatchSum (BERT-base) | ROUGE-1 | 31.85 | — | Unverified |
| 3 | Pointer-generator + coverage | ROUGE-1 | 28.53 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | LongT5 | ROUGE-1 | 76.87 | — | Unverified |
| 2 | BigBird-Pegasus | ROUGE-1 | 60.64 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Ground-truth transcript + Action with Hierarchical Attn | Content F1 | 48.9 | — | Unverified |
| 2 | BertSum | Content F1 | 36.4 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | CriSPO 3-shot | ROUGE-L | 54.1 | — | Unverified |
| 2 | Claude Instant + SigExt | ROUGE-L | 31.9 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | MiniLM-L12-multilingual | Spearman Correlation | 30.67 | — | Unverified |
| 2 | Contriever | Spearman Correlation | 30.36 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | mBARThez (OrangeSum abstract) | ROUGE-1 | 32.67 | — | Unverified |
| 2 | BARThez (OrangeSum abstract) | ROUGE-1 | 31.44 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | CriSPO 3-shot | ROUGE-1 | 63.1 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | HAT-CNNDM | ROUGE-1 | 52.27 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | BigBird-Pegasus | ROUGE-1 | 46.63 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | MatchSum | ROUGE-1 | 24.86 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Longformer Encoder Decoder | rouge1 | 38.65 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | GCN Hybrid | ROUGE-2 | 33.88 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | GenCompareSum | ROUGE-1 | 41.02 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | SATS | Avg. Test Rouge1 | 34.24 | — | Unverified |