Optical Character Recognition2020en
howsumm-step
Dataset from Papers With Code
Metrics:accuracy, cer, wer, f1
rouge-1
| # | Model | Score | Paper / Code | Date |
|---|---|---|---|---|
| 1 | LexRank (query: step title) | 39.6 | Oct 2021 | |
| 2 | CES (query: step title) | 39.3 | Oct 2021 | |
| 3 | CES (query: step + method titles) | 38.3 | Oct 2021 | |
| 4 | LexRank (query: step + method titles) | 38.2 | Oct 2021 | |
| 5 | CES (query: step + method + article titles) | 37 | Oct 2021 | |
| 6 | LexRank (query: step + method + article titles) | 36.3 | Oct 2021 | |
| 7 | GreedyRel (query: step + method titles) | 30.3 | Oct 2021 | |
| 8 | GreedyRel (query: step title) | 30.1 | Oct 2021 | |
| 9 | BM25-HierSumm (query: step + method titles) | 23 | Oct 2021 | |
| 10 | BM25-HierSumm (query: step title) | 22.3 | Oct 2021 | |
| 11 | BM25-HierSumm (query: step + method + article titles) | 21.9 | Oct 2021 |
Related Papers1
HowSumm: A Multi-Document Summarization Dataset Derived from WikiHow Articles
Oct 2021Models: LexRank (query: step title), CES (query: step title), CES (query: step + method titles) +8 more