Datasets:
Tasks:
Summarization
Modalities:
Text
Formats:
json
Languages:
English
Size:
10K - 100K
ArXiv:
License:
File size: 1,068 Bytes
cef8013 2117d40 cef8013 a203d91 78a8583 a203d91 78a8583 7156b2c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
---
license: apache-2.0
task_categories:
- summarization
language:
- en
pretty_name: PeerSum
size_categories:
- 10K<n<100K
---
This is PeerSum, a multi-document summarization dataset in the peer-review domain. More details can be found in the paper accepted at EMNLP 2023, [Summarizing Multiple Documents with Conversational Structure for Meta-review Generation](https://arxiv.org/abs/2305.01498). The original code and datasets are public on [GitHub](https://github.com/oaimli/PeerSum).
Please use the following code to download the dataset with the datasets library from Huggingface.
```python
from datasets import load_dataset
peersum_train = load_dataset("oaimli/PeerSum", split="train")
peersum_validation = load_dataset("oaimli/PeerSum", split="validation")
peersum_test = load_dataset("oaimli/PeerSum", split="test")
```
You can also download the raw data from [Google Drive](https://drive.google.com/drive/folders/1SGYvxY1vOZF2MpDn3B-apdWHCIfpN2uB?usp=sharing). The raw data comprises more information and it can be used for other analysis for peer reviews. |