Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -10,4 +10,46 @@ tags:
|
|
| 10 |
- nli
|
| 11 |
size_categories:
|
| 12 |
- 1K<n<10K
|
| 13 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
- nli
|
| 11 |
size_categories:
|
| 12 |
- 1K<n<10K
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
# USB: A Unified Summarization Benchmark Across Tasks and Domains
|
| 17 |
+
|
| 18 |
+
This benchmark contains labeled datasets for 8 text summarization based tasks given below.
|
| 19 |
+
The labeled datasets are created by collecting manual annotations on top of Wikipedia articles from 6 different domains.
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
|Task |Description |Code snippet |
|
| 24 |
+
|----------------|-------------------------------|-----------------------------|
|
| 25 |
+
| Extractive Summarization | Highlight important sentences in the source article | `load_dataset("kundank/usb","extractive_summarization")` |
|
| 26 |
+
| Abstractive Summarization | Generate a summary of the source | `load_dataset("kundank/usb","abstractive_summarization")` |
|
| 27 |
+
| Topic-based Summarization | Generate a summary of the source focusing on the given topic | `load_dataset("kundank/usb","topicbased_summarization")` |
|
| 28 |
+
| Multi-sentence Compression | Compress selected sentences into a one-line summary | `load_dataset("kundank/usb","multisentence_compression")` |
|
| 29 |
+
| Evidence Extraction | Surface evidence from the source for a summary sentence | `load_dataset("kundank/usb","evidence_extraction")` |
|
| 30 |
+
| Factuality Classification | Predict the factual accuracy of a summary sentence with respect to provided evidence | `load_dataset("kundank/usb","factuality_classification")` |
|
| 31 |
+
| Unsupported Span Prediction | Identify spans in a summary sentence which are not substantiated by the provided evidence | `load_dataset("kundank/usb","unsupported_span_prediction")` |
|
| 32 |
+
| Fixing Factuality | Rewrite a summary sentence to remove any factual errors or unsupported claims, with respect to provided evidence | `load_dataset("kundank/usb","fixing_factuality")` |
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
Additionally, to load the full set of collected annotations which were leveraged to make the labeled datasets for above tasks, use the command: ``load_dataset("kundank/usb","all_annotations")``
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
More details can be found in the paper: https://aclanthology.org/2023.findings-emnlp.592/
|
| 39 |
+
|
| 40 |
+
If you use this dataset, please cite it as below:
|
| 41 |
+
|
| 42 |
+
```
|
| 43 |
+
@inproceedings{krishna-etal-2023-usb,
|
| 44 |
+
title = "{USB}: A Unified Summarization Benchmark Across Tasks and Domains",
|
| 45 |
+
author = "Krishna, Kundan and
|
| 46 |
+
Gupta, Prakhar and
|
| 47 |
+
Ramprasad, Sanjana and
|
| 48 |
+
Wallace, Byron and
|
| 49 |
+
Bigham, Jeffrey and
|
| 50 |
+
Lipton, Zachary",
|
| 51 |
+
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
|
| 52 |
+
year = "2023",
|
| 53 |
+
pages = "8826--8845"
|
| 54 |
+
}
|
| 55 |
+
```
|