AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

CATCH: A Controllable Theme Detection Framework with Contextualized Clustering and Hierarchical Generation

Rui Ke, Jiahui Xu, Shenghao Yang, Kuang Wang, Feng Jiang, Haizhou Li

AAAI 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add preference-enhanced clustering (PeC) (ablation)CATCH · DSTC-12 Financeclustering accuracy: 48.8 -> 55.8Improved
add PeC (ablation)CATCH · DSTC-12 FinanceLLM score and ROUGE-2: LLM 98.4 -> 97.3; Rouge-2 26.7 -> 24.5Mixed
add topic segmentation (TopSeg) context representation (ablation)CATCH · DSTC-12 Financeclustering accuracy: 41.6 -> 55.8Improved
add hierarchical theme generation (HieGen) instead of flat (ablation)CATCH · DSTC-12 Financeclustering accuracy: 36.8 -> 55.8Improved
add hierarchical generation instead of flat (ablation)CATCH · DSTC-12 FinanceROUGE-L: 19.6 -> 42.4Improved
add utterance-level adaptation of the topic segmenter (ablation)CATCH · DSTC-12 Financeclustering accuracy: 50.6 -> 55.8Improved
add conversation-level adaptation of segmenter (ablation)CATCH · DSTC-12 Financeclustering accuracy: 44.3 -> 50.6Improved
replace spectral clustering with ClusterLLM (ablation)CATCH · DSTC-12 Financeclustering accuracy: 55.8 -> 70.1Improved
replace LLaMA3-8B with GPT-4o for label generation (ablation)CATCH · DSTC-12 FinanceCosSim: 59.3 -> 61.1Improved
replace LLaMA3-8B with Qwen-3-8B for labels (ablation)CATCH · DSTC-12 FinanceCosSim: 59.3 -> 59.3No clear effect
oracle clustering (PeC=1) vs real clustering (ablation)CATCH · DSTC-12 FinanceROUGE-1: 42.4 -> 51.6Improved
CATCH vs GRM preference-reward baselineDSTC-12 Banking · DSTC-12 Bankingclustering accuracy: 46.9 -> 56.7Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy