Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.04997 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914450413453312 |
|---|---|
| author | Lu, Rong Liu, Hao Hou, Song |
| author_facet | Lu, Rong Liu, Hao Hou, Song |
| contents | This work presents a comparative analysis of embedding-based and generative models for classifying geoscience technical documents. Using a multi-disciplinary benchmark dataset, we evaluated the trade-offs between model accuracy, stability, and computational cost. We find that generative Vision-Language Models (VLMs) like Qwen2.5-VL, enhanced with Chain-of-Thought (CoT) prompting, achieve superior zero-shot accuracy (82%) compared to state-of-the-art multimodal embedding models like QQMM (63%). We also demonstrate that while supervised fine-tuning (SFT) can improve VLM performance, it is sensitive to training data imbalance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_04997 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Evaluation of Embedding-Based and Generative Methods for LLM-Driven Document Classification: Opportunities and Challenges Lu, Rong Liu, Hao Hou, Song Information Retrieval Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition Machine Learning This work presents a comparative analysis of embedding-based and generative models for classifying geoscience technical documents. Using a multi-disciplinary benchmark dataset, we evaluated the trade-offs between model accuracy, stability, and computational cost. We find that generative Vision-Language Models (VLMs) like Qwen2.5-VL, enhanced with Chain-of-Thought (CoT) prompting, achieve superior zero-shot accuracy (82%) compared to state-of-the-art multimodal embedding models like QQMM (63%). We also demonstrate that while supervised fine-tuning (SFT) can improve VLM performance, it is sensitive to training data imbalance. |
| title | Evaluation of Embedding-Based and Generative Methods for LLM-Driven Document Classification: Opportunities and Challenges |
| topic | Information Retrieval Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2604.04997 |