{"paper":{"title":"Beyond Single-Model Optimization: Preserving Plasticity in Continual Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Maintaining archives of diverse policies in a shared latent space preserves plasticity in continual reinforcement learning.","cross_cats":["cs.AI","cs.NE"],"primary_cat":"cs.LG","authors_text":"Lute Lillo, Nick Cheney","submitted_at":"2026-04-16T17:06:54Z","abstract_excerpt":"Continual reinforcement learning must balance retention with adaptation, yet many methods still rely on \\emph{single-model preservation}, committing to one evolving policy as the main reusable solution across tasks. Even when a previously successful policy is retained, it may no longer provide a reliable starting point for rapid adaptation after interference, reflecting a form of \\emph{loss of plasticity} that single-policy preservation cannot address. Inspired by quality-diversity methods, we introduce \\textsc{TeLAPA} (Transfer-Enabled Latent-Aligned Policy Archives), a continual RL framework"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"In our MiniGrid CL setting, TeLAPA learns more tasks successfully, recovers competence faster on revisited tasks after interference, and retains higher performance across a sequence of tasks. Our analyses show that source-optimal policies are often not transfer-optimal, even within a local competent neighborhood, and that effective reuse depends on retaining and selecting among multiple nearby alternatives rather than collapsing them to one representative.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That organizing policies into behaviorally diverse neighborhoods via a shared latent space will reliably preserve plasticity and transfer better than single-model preservation across a wide range of continual RL domains and interference patterns.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"TeLAPA maintains archives of behaviorally diverse yet competent policies aligned in a shared latent space to preserve plasticity and enable faster recovery after interference in continual reinforcement learning.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Maintaining archives of diverse policies in a shared latent space preserves plasticity in continual reinforcement learning.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"8257ab599890b7d64ca5e74285966f095502f11f70a030cab81649447d8d2bbd"},"source":{"id":"2604.15414","kind":"arxiv","version":2},"verdict":{"id":"c9dee478-1f4a-4b6b-8206-e391795a0c0b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T10:57:09.649445Z","strongest_claim":"In our MiniGrid CL setting, TeLAPA learns more tasks successfully, recovers competence faster on revisited tasks after interference, and retains higher performance across a sequence of tasks. Our analyses show that source-optimal policies are often not transfer-optimal, even within a local competent neighborhood, and that effective reuse depends on retaining and selecting among multiple nearby alternatives rather than collapsing them to one representative.","one_line_summary":"TeLAPA maintains archives of behaviorally diverse yet competent policies aligned in a shared latent space to preserve plasticity and enable faster recovery after interference in continual reinforcement learning.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That organizing policies into behaviorally diverse neighborhoods via a shared latent space will reliably preserve plasticity and transfer better than single-model preservation across a wide range of continual RL domains and interference patterns.","pith_extraction_headline":"Maintaining archives of diverse policies in a shared latent space preserves plasticity in continual reinforcement learning."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.15414/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}