Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
datasets:
|
| 3 |
+
- cyanic-selkie/aida-conll-yago-wikidata
|
| 4 |
+
language:
|
| 5 |
+
- en
|
| 6 |
+
metrics:
|
| 7 |
+
- accuracy
|
| 8 |
+
- f1
|
| 9 |
+
tags:
|
| 10 |
+
- entity linking
|
| 11 |
+
- entity disambiguation
|
| 12 |
+
- EL
|
| 13 |
+
- ReFinED
|
| 14 |
+
- RoBERTa
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
# Model Card for LLMAEL-ReFinED-FT
|
| 18 |
+
|
| 19 |
+
<p align="justify">
|
| 20 |
+
|
| 21 |
+
We introduce <b>LLMAEL</b> (<b>LLM</b>-<b>A</b>ugmented <b>E</b>ntity <b>L</b>inking), a pipeline method to enhance entity linking through LLM data augmentation.
|
| 22 |
+
We release our customly fine-tuned <b>LLMAEL-ReFinED-FT</b> model, which is fine-tuned from the <b>ReFinED</b> EL model using an <b>Llama-3-70b</b> augmented version of the <b>AIDA_train</b> dataset.
|
| 23 |
+
LLMAEL-ReFinED-FT yields new SOTA results across six standard EL benchmarks: AIDA_test, MSNBC, AQUAINT, ACE2004, WNED-CLUEWEB, and WNED-WIKIPEDIA, achieving an average 1.21% accuracy gain.
|
| 24 |
+
|
| 25 |
+
For more details, refer to our paper 📖 [LLMAEL: Large Language Models are Good Context Augmenters for Entity Linking](https://arxiv.org/abs/2407.04020)
|
| 26 |
+
</p>
|
| 27 |
+
|
| 28 |
+
|
| 29 |
+
### Model Description
|
| 30 |
+
|
| 31 |
+
- **Developed by:** Amy Xin, Yunjia Qi, Zijun Yao, Fangwei Zhu, Kaisheng Zeng, Bin Xu, Lei Hou, Juanzi Li
|
| 32 |
+
- **Model type:** Entity Linking Model
|
| 33 |
+
- **Language(s):** English
|
| 34 |
+
- **Finetuned from model [optional]:** [ReFinED](https://arxiv.org/abs/2207.04108)
|