Update README.md
Browse files
README.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
-
base_model: Qwen/
|
| 4 |
tags:
|
| 5 |
- boolean-queries
|
| 6 |
- systematic-review
|
|
@@ -20,7 +20,7 @@ This model is part of the **AutoBool** framework, a reinforcement learning appro
|
|
| 20 |
|
| 21 |
This variant uses the **objective method** grounded in domain expertise and structured logic. The model simulates a relevant article and extracts key terms to construct the Boolean query, following a systematic 6-step process.
|
| 22 |
|
| 23 |
-
- **Base Model:** Qwen/
|
| 24 |
- **Training Method:** GRPO (Group Relative Policy Optimization) with LoRA fine-tuning
|
| 25 |
- **Prompt Strategy:** Objective method (hypothetical article simulation)
|
| 26 |
- **Step 1:** Simulate a concise title and abstract (2-3 sentences) of a relevant and focused article
|
|
@@ -37,7 +37,7 @@ This variant uses the **objective method** grounded in domain expertise and stru
|
|
| 37 |
The model was trained using:
|
| 38 |
- **Optimization:** GRPO (Group Relative Policy Optimization)
|
| 39 |
- **Fine-tuning:** LoRA (Low-Rank Adaptation)
|
| 40 |
-
- **Dataset:**
|
| 41 |
- **Reward Function:** Combines syntactic validity, format correctness, and retrieval effectiveness
|
| 42 |
- **Learning Approach:** Example-based pattern recognition
|
| 43 |
|
|
@@ -141,10 +141,10 @@ Query:", query)
|
|
| 141 |
If you use this model, please cite:
|
| 142 |
|
| 143 |
```bibtex
|
| 144 |
-
@inproceedings{
|
| 145 |
title={AutoBool: Reinforcement Learning for Boolean Query Generation in Systematic Reviews},
|
| 146 |
-
author={[]},
|
| 147 |
-
booktitle={Proceedings of the
|
| 148 |
year={2025}
|
| 149 |
}
|
| 150 |
```
|
|
@@ -152,7 +152,7 @@ If you use this model, please cite:
|
|
| 152 |
## More Information
|
| 153 |
|
| 154 |
- **GitHub Repository:** [https://github.com/ielab/AutoBool](https://github.com/ielab/AutoBool)
|
| 155 |
-
- **Paper:** Accepted at EACL
|
| 156 |
|
| 157 |
## License
|
| 158 |
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
+
base_model: Qwen/Qwen3-4B
|
| 4 |
tags:
|
| 5 |
- boolean-queries
|
| 6 |
- systematic-review
|
|
|
|
| 20 |
|
| 21 |
This variant uses the **objective method** grounded in domain expertise and structured logic. The model simulates a relevant article and extracts key terms to construct the Boolean query, following a systematic 6-step process.
|
| 22 |
|
| 23 |
+
- **Base Model:** Qwen/Qwen3-4B
|
| 24 |
- **Training Method:** GRPO (Group Relative Policy Optimization) with LoRA fine-tuning
|
| 25 |
- **Prompt Strategy:** Objective method (hypothetical article simulation)
|
| 26 |
- **Step 1:** Simulate a concise title and abstract (2-3 sentences) of a relevant and focused article
|
|
|
|
| 37 |
The model was trained using:
|
| 38 |
- **Optimization:** GRPO (Group Relative Policy Optimization)
|
| 39 |
- **Fine-tuning:** LoRA (Low-Rank Adaptation)
|
| 40 |
+
- **Dataset:** wshuai190/pubmed-pmc-sr-filtered
|
| 41 |
- **Reward Function:** Combines syntactic validity, format correctness, and retrieval effectiveness
|
| 42 |
- **Learning Approach:** Example-based pattern recognition
|
| 43 |
|
|
|
|
| 141 |
If you use this model, please cite:
|
| 142 |
|
| 143 |
```bibtex
|
| 144 |
+
@inproceedings{autobool2026,
|
| 145 |
title={AutoBool: Reinforcement Learning for Boolean Query Generation in Systematic Reviews},
|
| 146 |
+
author={[Shuai Wang, Harrisen Scells, Bevan Koopman, Guido Zuccon]},
|
| 147 |
+
booktitle={Proceedings of the 2026 Conference of the European Chapter of the Association for Computational Linguistics (EACL)},
|
| 148 |
year={2025}
|
| 149 |
}
|
| 150 |
```
|
|
|
|
| 152 |
## More Information
|
| 153 |
|
| 154 |
- **GitHub Repository:** [https://github.com/ielab/AutoBool](https://github.com/ielab/AutoBool)
|
| 155 |
+
- **Paper:** Accepted at EACL 2026
|
| 156 |
|
| 157 |
## License
|
| 158 |
|