Loading
Life sciences · Preprint
arXiv · September 9, 2026
Posted before peer review. The findings may change or fail to hold.
This is a preprint proposing CoGe-GCD, a machine learning module for generalized category discovery that combines compositional perception and structure-preserving induction. The work is computational and methodological, lacking peer review and any clinical or medical application, and therefore falls outside the scope of evidence for clinical practice.
Preprint.
CoGe-GCD improves all-class accuracy, unknown-class number estimation, and geometric quality on standard benchmarks. The method operates as a pluggable inductive-bias module with marginal computational overhead and compatibility with diverse GCD frameworks. Compositional Perception structures patch tokens via competitive token-primitive assignment; Generalizing Induction applies structure-preserving calibration over spatial relations.
Safety was not reported in the material analysed. Check the source before drawing any conclusion about harm.
The source did not state who this applies to in practice.
This is an unrefereed arXiv preprint describing a machine learning method for generalized category discovery; it presents algorithmic innovation and benchmark results but lacks peer review and clinical/medical relevance.
Graded across the dimensions that decide whether you should act, each from what the source actually supports. There is no single score, and where a dimension was not assessed it says so.
What is missing. This record has no reported figures. That is a gap in the analysis, not a judgement about the study.
Generalized Category Discovery (GCD) assigns unlabeled instances, mixed with labeled data, to known or novel categories, requiring human-like compositional reasoning: reusing primitives learned from known classes and deciding when new combinations imply new categories. Existing GCD methods operate on unstructured token features and struggle to extrapolate to novel compositions. We propose CoGe-GCD, which rethinks GCD through compositional generalization with two coupled stages. (i) Compositional Perception structures patch tokens by mapping them to a small vocabulary of primitives and refining token embeddings via competitive token-primitive assignment and information passing, yielding coherent groups for discovery. (ii) Generalizing Induction exploits the induced geometric structure and applies a structure-preserving calibration over spatial relations, maintaining probabilistic semantics while improving extrapolation to unseen primitive combinations. CoGe-GCD is implemented as an inductive-bias module between backbone and projection head, without modifying heads or losses, and can be plugged into diverse GCD frameworks. On standard benchmarks, it consistently improves all-class accuracy, unknown-class number estimation, and geometric quality, with marginal computational overhead. Code is available at https://github.com/lytang63/CoGe-GCD.
Taken from the source record, never inferred. Follow any of these and new work involving them reaches your briefing.