---
title: "Expanding the Category of Classifiers with LLM Supervision"
authors:
  - "Derui Lyu"
  - "Xiangyu Wang"
  - "Taiyu Ban"
  - "Lyuzhou Chen"
  - "Xiren Zhou"
  - "Huanhuan Chen"
doi: "10.24963/ijcai.2025/657"
url: "https://ijcai.org/proceedings/2025/657"
pdf: "https://ijcai.org/proceedings/2025/0657.pdf"
venue: "Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence"
year: 2025
pages: "5905-5913"
track: "Main Track"
keywords:
  - "Machine Learning: ML: Incremental learning"
  - "Machine Learning: ML: Applications"
  - "Natural Language Processing: NLP: Language models"
  - "Machine Learning: ML: Multi-modal learning"
---

# Expanding the Category of Classifiers with LLM Supervision

Derui Lyu, Xiangyu Wang, Taiyu Ban, Lyuzhou Chen, Xiren Zhou, Huanhuan Chen

*Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence*, pages 5905-5913. DOI: [10.24963/ijcai.2025/657](https://doi.org/10.24963/ijcai.2025/657)

## Abstract

Zero-shot learning has shown significant potential for creating cost-effective and flexible systems to expand classifiers to new categories. However, existing methods still rely on manually created attributes designed by domain experts. Motivated by the widespread success of large language models (LLMs), we introduce an LLM-driven framework for class-incremental learning that removes the need for human intervention, termed Classifier Expansion with Multi-vIew LLM knowledge (CEMIL). In CEMIL, an LLM agent autonomously generates detailed textual multi-view descriptions for unseen classes, offering richer and more flexible class representations than traditional expert-constructed vectorized attributes. These LLM-derived textual descriptions are integrated through a contextual filtering attention mechanism to produce discriminative class embeddings. Subsequently, a weight injection module maps the class embeddings to classifier weights, enabling seamless expansion to new classes. Experimental results show that CEMIL outperforms existing methods using expert-constructed attributes, demonstrating its effectiveness for fully automated classifier expansion without human participation.