{"ID":2921080,"CreatedAt":"2026-06-02T02:42:49.606572591Z","UpdatedAt":"2026-06-04T07:41:34.29888543Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2606.01863","arxiv_id":"2606.01863","title":"Continual Learning as a Multiphase Moving-Boundary Problem","abstract":"Continual learning struggles to balance retaining past knowledge with absorbing new tasks. Stefan-CL elegantly resolves this stability-plasticity dilemma through the physics of melting. It frames consolidated knowledge as a protected \"solid\" and unused capacity as an adaptable \"liquid.\" As the network learns, this boundary expands, governed by a \"latent heat\" tuning dial. By mathematically freezing the learned interior, Stefan-CL cuts forgetting to near zero, matching memory-heavy baselines without storing raw data, forging a beautiful, physics-grounded path for AI.","short_abstract":"Continual learning struggles to balance retaining past knowledge with absorbing new tasks. Stefan-CL elegantly resolves this stability-plasticity dilemma through the physics of melting. It frames consolidated knowledge as a protected \"solid\" and unused capacity as an adaptable \"liquid.\" As the network learns, this boun...","url_abs":"https://arxiv.org/abs/2606.01863","url_pdf":"https://arxiv.org/pdf/2606.01863v1","authors":"[\"Snigdha Chandan Khilar\"]","published":"2026-06-01T08:13:33Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"math-ph\"]","methods":"[\"Generative Adversarial Network\"]","has_code":false}