The preprint presents a novel optimization technique, Energy-Aware Gradient Coordinator (EAGC), which addresses gradient entanglement in Generalized Category Discovery (GCD). Gradient entanglement distorts supervised gradients and weakens discrimination among known classes, and induces representation-subspace overlap between known and novel classes, reducing the separability of novel categories. EAGC comprises two components: Anchor-based Gradient Alignment (AGA) and Energy-aware Elastic Projection (EEP). AGA introduces a reference model to anchor the gradient directions of labeled samples, preserving the discriminative structure of known classes against the interference of unlabeled gradients. EEP softly projects unlabeled gradients onto the complement of the known-class subspace and derives an energy-based coefficient to adaptively scale the projection for each unlabeled sample according to its degree of alignment with the known subspace. Experiments show that EAGC consistently boosts existing methods and establishes new state-of-the-art results.
Energy-Aware Gradient Coordinator for Robust Generalized Category Discovery
A new optimization technique, EAGC, improves Generalized Category Discovery by addressing gradient entanglement and achieves state-of-the-art results.
External source stays available while the OJO article and comment thread stay local.
More in Artificial Intelligence & Machine Learning
view topicA new multimodal ML framework combines ECG and EHR features to classify LVEF, outperforming baselines and maintaining performance under temporal validation.
A novel framework for fault diagnosis in general aviation aircraft achieves 96.2% Macro-F1 using multi-fidelity digital twins and FMEA-driven fault injection.
A systematic study of weight matrix singular value spectra during transformer pretraining reveals three phenomena that fundamentally change how we understand transformer training.
A novel framework for adaptive and reproducible medical image processing addresses the limitations of current medical imaging research by introducing adaptability and reproducibility.
Comments load interactively on the live page.