TabClustPFN tops 44 real tabular clustering benchmarks, moving clustering into pretraining
jiqizhixin · x · 2026-11-06
Researchers from Renmin University of China, Rutgers, NUS and MBZUAI present TabClustPFN at NeurIPS 2026, shifting dataset-specific clustering work into pretraining. Across 44 real tabular datasets it ranks first by median ARI when given the true cluster count, with its automatic-count variant sharing second place.
- A Partition Inference Network proposes assignments for candidate cluster counts; a Cardinality Inference Network summarizes those partitions and predicts a distribution over cluster numbers.
- The SoftARI objective treats equivalent relabelings as the same partition, avoiding arbitrary cluster-name conventions; training tasks come from synthetic priors.
- At inference the checkpoint stays fixed—new tables need no per-dataset optimization. Benchmark datasets have at most 1,000 samples, 64 features and ten clusters.
More from Research
- Project AURORA launches at Harvard: AI cancer co-scientists backed by CRUK and NCI — zakkohane · 2026-11-06
- Google DeepMind Open-Endedness team hiring Student Researchers in Toronto — robertarail · 2026-11-06
- VisMatch v1.4.0 adds RaCo, UPAL, REDI-Match: 50+ image matchers in 5 lines — rsasaki0109 · 2026-11-06
- Trainable [DUMP] Prefix Tokens Reduce Representational Interference in Retrieval Embeddings — nlpai-lab · 2026-11-06
- Nace AI open-sources Drex decision models: 27B rivals frontier LLMs at 183ms — omarsar0 · 2026-11-06
- MIT's David Bau Challenges Gates, Hinton, Bengio: Show Me One Unthinkable Thought — davidbau · 2026-11-06