Improve model card: Add `library_name`, update `license`, link paper & code, correct title and citation
#1
by
nielsr
HF Staff
- opened
README.md
CHANGED
|
@@ -1,19 +1,23 @@
|
|
| 1 |
---
|
| 2 |
language: en
|
| 3 |
-
license:
|
| 4 |
-
tags:
|
| 5 |
-
- finance
|
| 6 |
-
- risk-relation
|
| 7 |
-
- retrieval
|
| 8 |
-
- encoder
|
| 9 |
-
- feature-extraction
|
| 10 |
-
- stock-prediction
|
| 11 |
pipeline_tag: feature-extraction
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
---
|
| 13 |
|
| 14 |
-
# Financial Risk Identification through Dual-view Adaptation — Encoder
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
-
This repository hosts the pretrained encoder from the work **“Financial Risk Identification through Dual-view Adaptation.”**
|
| 17 |
The model is designed to uncover **inter-firm risk relations** from financial text, supporting downstream tasks such as **retrieval**, **relation mining**, and **stock-signal experiments** where relation strength acts as a feature.
|
| 18 |
|
| 19 |
> **Files**
|
|
@@ -28,7 +32,7 @@ The model is designed to uncover **inter-firm risk relations** from financial te
|
|
| 28 |
The model aligns two complementary “views” of firm relations and adapts them during training:
|
| 29 |
|
| 30 |
- **Lexical view (`lex`)** — focuses on token/phrase-level and domain terms common in 10-K and financial news.
|
| 31 |
-
- **Temporal view (`time`)
|
| 32 |
|
| 33 |
A **two-view combination (“Best”)** integrates both signals and yields stronger retrieval quality and more stable risk-relation estimates. Ablations (`lex`, `time`) are also supported for analysis.
|
| 34 |
|
|
@@ -82,9 +86,12 @@ print(similarity)
|
|
| 82 |
If you use this model or the dual-view methodology, please cite:
|
| 83 |
```bibtex
|
| 84 |
@misc{financial_risk_dualview_2025,
|
| 85 |
-
title = {Financial Risk Identification through Dual-view Adaptation},
|
| 86 |
author = {Chiu, Wei-Ning and collaborators},
|
| 87 |
year = {2025},
|
| 88 |
-
|
| 89 |
-
|
|
|
|
|
|
|
| 90 |
}
|
|
|
|
|
|
| 1 |
---
|
| 2 |
language: en
|
| 3 |
+
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
pipeline_tag: feature-extraction
|
| 5 |
+
tags:
|
| 6 |
+
- finance
|
| 7 |
+
- risk-relation
|
| 8 |
+
- retrieval
|
| 9 |
+
- encoder
|
| 10 |
+
- feature-extraction
|
| 11 |
+
- stock-prediction
|
| 12 |
+
library_name: transformers
|
| 13 |
---
|
| 14 |
|
| 15 |
+
# Financial Risk Relation Identification through Dual-view Adaptation — Encoder
|
| 16 |
+
|
| 17 |
+
**Paper:** [Financial Risk Relation Identification through Dual-view Adaptation](https://huggingface.co/papers/2509.18775)
|
| 18 |
+
**Code:** [https://github.com/cnclabs/codes.fin.relation](https://github.com/cnclabs/codes.fin.relation)
|
| 19 |
|
| 20 |
+
This repository hosts the pretrained encoder from the work **“Financial Risk Relation Identification through Dual-view Adaptation.”**
|
| 21 |
The model is designed to uncover **inter-firm risk relations** from financial text, supporting downstream tasks such as **retrieval**, **relation mining**, and **stock-signal experiments** where relation strength acts as a feature.
|
| 22 |
|
| 23 |
> **Files**
|
|
|
|
| 32 |
The model aligns two complementary “views” of firm relations and adapts them during training:
|
| 33 |
|
| 34 |
- **Lexical view (`lex`)** — focuses on token/phrase-level and domain terms common in 10-K and financial news.
|
| 35 |
+
- **Temporal view (`time`) — encourages stability/consistency of relations across reporting periods and evolving events.
|
| 36 |
|
| 37 |
A **two-view combination (“Best”)** integrates both signals and yields stronger retrieval quality and more stable risk-relation estimates. Ablations (`lex`, `time`) are also supported for analysis.
|
| 38 |
|
|
|
|
| 86 |
If you use this model or the dual-view methodology, please cite:
|
| 87 |
```bibtex
|
| 88 |
@misc{financial_risk_dualview_2025,
|
| 89 |
+
title = {Financial Risk Relation Identification through Dual-view Adaptation},
|
| 90 |
author = {Chiu, Wei-Ning and collaborators},
|
| 91 |
year = {2025},
|
| 92 |
+
eprint={2509.18775},
|
| 93 |
+
archivePrefix={arXiv},
|
| 94 |
+
primaryClass={cs.CL},
|
| 95 |
+
url={https://arxiv.org/abs/2509.18775}
|
| 96 |
}
|
| 97 |
+
```
|