Scholarlink

<p>Hierarchical classification of biotechnology R&amp;D projects is critical for strategic planning and national R&amp;D resource allocation, yet remains challenging due to fine-grained hierarchical taxonomies, severe class imbalance, and the cost of large language model (LLM) inference at scale. We propose <b>EnergyRoute</b>, an uncertainty-aware routing framework that distributes computation across three tiers: (i)&#xa0;a fine-tuned transformer encoder for confident predictions, (ii)&#xa0;retrieval-augmented <i>k</i>-nearest-neighbor fusion for moderately uncertain cases, and (iii)&#xa0;evidence-grounded LLM classification for the most difficult cases. Routing decisions are guided by the Helmholtz free energy of the classifier’s logit distribution, which serves as a calibration-free uncertainty measure. On a Korean biotechnology project dataset comprising 82,316 training instances and 301 leaf classes with a <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(964\times\)</EquationSource> </InlineEquation> class-imbalance ratio, EnergyRoute achieves a leaf-level micro-F1 of 0.862 and a macro-F1 of 0.778—a statistically significant improvement over the fine-tuned encoder baseline—while routing only 20% of samples to the LLM. Cross-dataset evaluation on the English Web of Science benchmark (142 classes) provides evidence of cross-dataset generalizability. Compared with existing hierarchical text classification methods and confidence-based cascading, EnergyRoute yields the best performance–cost trade-off, reducing LLM usage approximately fivefold relative to full LLM pipelines.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

EnergyRoute: energy-based uncertainty routing for selective retrieval and large language model assistance in the hierarchical classification of biotechnology R&D projects

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月28日

查看全文
  • Gwangseon Jang,
  • Moo Woong Kim,
  • Youn-jeong Nam,
  • Yewon Yong,
  • Kyong-Ha Lee,
  • Chanuk Lim

摘要

Hierarchical classification of biotechnology R&D projects is critical for strategic planning and national R&D resource allocation, yet remains challenging due to fine-grained hierarchical taxonomies, severe class imbalance, and the cost of large language model (LLM) inference at scale. We propose EnergyRoute, an uncertainty-aware routing framework that distributes computation across three tiers: (i) a fine-tuned transformer encoder for confident predictions, (ii) retrieval-augmented k-nearest-neighbor fusion for moderately uncertain cases, and (iii) evidence-grounded LLM classification for the most difficult cases. Routing decisions are guided by the Helmholtz free energy of the classifier’s logit distribution, which serves as a calibration-free uncertainty measure. On a Korean biotechnology project dataset comprising 82,316 training instances and 301 leaf classes with a \(964\times\) class-imbalance ratio, EnergyRoute achieves a leaf-level micro-F1 of 0.862 and a macro-F1 of 0.778—a statistically significant improvement over the fine-tuned encoder baseline—while routing only 20% of samples to the LLM. Cross-dataset evaluation on the English Web of Science benchmark (142 classes) provides evidence of cross-dataset generalizability. Compared with existing hierarchical text classification methods and confidence-based cascading, EnergyRoute yields the best performance–cost trade-off, reducing LLM usage approximately fivefold relative to full LLM pipelines.