{"id":"816df1ce8de199aa50964d39a9a04818b2688bca7754d60720c2b73753764c50","title":"arXiv eess.SY: Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","argument":"In their paper, Zexin Li, Nikil Dutt and Cong Liu propose Orion, a system for managing memory as a device learns from incoming data. Online continual learning (OCL) means adapting a model as new data arrives, rather than training it once on a fixed dataset. The authors argue that memory pressure and changing workloads make fixed, offline memory settings unsuitable for this task on resource-limited devices.\n\nOrion uses URGE, a runtime indicator based on the idea that performance is limited by the scarcest resource, to reallocate memory among batch processing, replay buffers and optimization strategies. It also coordinates these choices at the operating-system and application levels, and uses data prefetching. The authors say they implemented a prototype with Avalanche-lib, evaluated it across OCL algorithms, benchmarks and hardware, and integrated it into an autonomous navigational robot.\n\nThe paper’s central claim is that dynamically managing memory can make on-device OCL more practical while balancing training latency, plasticity and stability. The authors report significant training speedups and minimal runtime, memory and energy overhead, but the abstract does not give the size of those gains. The evidence described here is the authors’ report of evaluations and a prototype; it does not provide the detailed results needed to assess performance across particular settings.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied evidence is the arXiv abstract, which reports the authors’ findings but does not include numerical speedups or overhead measurements, detailed evaluation results, or comparisons for specific settings.","speaker":{"name":"Li, Zexin, Dutt, Nikil, Liu, Cong","roleAtSource":"Article author"},"source":{"title":"Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","publisher":"arXiv eess.SY","format":"Article","url":"https://arxiv.org/abs/2605.26473","publishedAt":"2026-05-26","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors propose Orion to manage memory dynamically for on-device online continual learning, balancing training latency, plasticity and stability under memory constraints.","excerpt":"a holistic framework designed to co-optimize training latency, plasticity, and stability of state-of-the-art OCL models under strict memory constraints","url":"https://arxiv.org/abs/2605.26473","relationship":"Source-stated"},{"text":"Orion uses URGE to reallocate memory across OCL components and coordinates batch processing, replay buffers and optimization strategies at both operating-system and application levels.","excerpt":"dynamically reallocate memory across OCL components by jointly coordinating batch processing, replay buffers, and optimization strategies at both the OS and application level","url":"https://arxiv.org/abs/2605.26473","relationship":"Source-stated"},{"text":"The authors report evaluating an Orion prototype across OCL algorithms, benchmarks and hardware, and integrating it into an autonomous navigational robot.","excerpt":"evaluated across a diverse range of OCL algorithms, benchmarks, and hardware platforms commonly used in autonomous robotic applications.","url":"https://arxiv.org/abs/2605.26473","relationship":"Source-stated"},{"text":"The authors report significant training speedups with minimal runtime, memory and energy overhead.","excerpt":"achieves significant training speedups while maintaining balanced performance and effectively adapting to various scenarios, all with minimal runtime, memory, and energy overhead","url":"https://arxiv.org/abs/2605.26473","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2605.26473","checkedAt":"2026-11-10T10:50:07.644Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What speedups and runtime, memory and energy overheads did Orion achieve in the evaluated settings?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv eess.SY: Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","summary":"In their paper, Zexin Li, Nikil Dutt and Cong Liu propose Orion, a system for managing memory as a device learns from incoming data. Online continual learning (OCL) means adapting a model as new data arrives, rather than training it once on a fixed dataset. The authors argue that memory pressure and changing workloads make fixed, offline memory settings unsuitable for this task on resource-limited devices.\n\nOrion uses URGE, a runtime indicator based on the idea that performance is limited by the scarcest resource, to reallocate memory among batch processing, replay buffers and optimization strategies. It also coordinates these choices at the operating-system and application levels, and uses data prefetching. The authors say they implemented a prototype with Avalanche-lib, evaluated it across OCL algorithms, benchmarks and hardware, and integrated it into an autonomous navigational robot.\n\nThe paper’s central claim is that dynamically managing memory can make on-device OCL more practical while balancing training latency, plasticity and stability. The authors report significant training speedups and minimal runtime, memory and energy overhead, but the abstract does not give the size of those gains. The evidence described here is the authors’ report of evaluations and a prototype; it does not provide the detailed results needed to assess performance across particular settings.","question":"What speedups and runtime, memory and energy overheads did Orion achieve in the evaluated settings?","claims":[{"text":"The authors propose Orion to manage memory dynamically for on-device online continual learning, balancing training latency, plasticity and stability under memory constraints.","excerpt":"a holistic framework designed to co-optimize training latency, plasticity, and stability of state-of-the-art OCL models under strict memory constraints","support":"source-stated"},{"text":"Orion uses URGE to reallocate memory across OCL components and coordinates batch processing, replay buffers and optimization strategies at both operating-system and application levels.","excerpt":"dynamically reallocate memory across OCL components by jointly coordinating batch processing, replay buffers, and optimization strategies at both the OS and application level","support":"source-stated"},{"text":"The authors report evaluating an Orion prototype across OCL algorithms, benchmarks and hardware, and integrating it into an autonomous navigational robot.","excerpt":"evaluated across a diverse range of OCL algorithms, benchmarks, and hardware platforms commonly used in autonomous robotic applications.","support":"source-stated"},{"text":"The authors report significant training speedups with minimal runtime, memory and energy overhead.","excerpt":"achieves significant training speedups while maintaining balanced performance and effectively adapting to various scenarios, all with minimal runtime, memory, and energy overhead","support":"source-stated"}],"limitations":["The supplied evidence is the arXiv abstract, which reports the authors’ findings but does not include numerical speedups or overhead measurements, detailed evaluation results, or comparisons for specific settings."],"mediaId":"9046b562174a46d7bcaa04109733dd90c74e0289a9195067f8ea196385b902da","correctionReason":null,"id":"816df1ce8de199aa50964d39a9a04818b2688bca7754d60720c2b73753764c50","publicationId":"d618867b-d871-4ed1-8060-3770ebe14632","version":1,"sourceUrl":"https://arxiv.org/abs/2605.26473","sourceTitle":"Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","publisher":"arXiv eess.SY","publishedAt":null,"originalPublished":{"value":"2026-05-26","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-03T08:58:49.837Z","createdAt":"2026-11-10T09:46:12.064Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"9046b562174a46d7bcaa04109733dd90c74e0289a9195067f8ea196385b902da","kind":"article","url":"https://arxiv.org/abs/2605.26473","sourceUrl":"https://arxiv.org/abs/2605.26473","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-04T01:46:23.796Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2605.26473","checkedAt":"2026-11-10T10:50:07.644Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"9046b562174a46d7bcaa04109733dd90c74e0289a9195067f8ea196385b902da","confidence":0.75,"provider":"typesafe","model":"jev-1.13.0","inputHash":"8c1dbc0af61efbd208e6ec2d59110ae4b99c0f5b02ffb41069f6be6ecebfb4b7","decidedAt":"2026-11-04T01:46:24.325Z","redecided":{"from":"stream-admission-v5","under":"stream-admission-v5+confidence-0.5"}},"provenance":{"textHash":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","observationId":"ee96c0eb9edeafd26775bdf53837c54274a659dabdc2cefbf7ef9162c54def29","retrievedAt":"2026-11-04T01:46:23.736Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"7c9e2c46-6797-436f-81d3-3fc0e5e22417"},"reviewer":{"model":"gpt-6-luna","runId":"e398f34e-32f3-4182-8e19-c4e3ea4e8a33"}},"audit":{"status":"passed","reviewedAt":"2026-11-10T09:46:12.064Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Li, Zexin","url":null,"kind":"unknown"},{"name":"Dutt, Nikil","url":null,"kind":"unknown"},{"name":"Liu, Cong","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Li, Zexin","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Li, Zexin","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.26473","locator":"head/meta[name=author|citation_author]"},"assertionId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T10:50:07.644Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081","sourceUrl":"https://arxiv.org/abs/2605.26473","contentSha256":"fd53d53b2357e7e8da394b8790465f0cb6104bbeb611a5016a00342eef4d2257","textHash":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","observedAt":"2026-11-10T10:50:07.644Z"}},{"actor":{"name":"Dutt, Nikil","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Dutt, Nikil","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.26473","locator":"head/meta[name=author|citation_author]"},"assertionId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T10:50:07.644Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081","sourceUrl":"https://arxiv.org/abs/2605.26473","contentSha256":"fd53d53b2357e7e8da394b8790465f0cb6104bbeb611a5016a00342eef4d2257","textHash":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","observedAt":"2026-11-10T10:50:07.644Z"}},{"actor":{"name":"Liu, Cong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Liu, Cong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.26473","locator":"head/meta[name=author|citation_author]"},"assertionId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T10:50:07.644Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081","sourceUrl":"https://arxiv.org/abs/2605.26473","contentSha256":"fd53d53b2357e7e8da394b8790465f0cb6104bbeb611a5016a00342eef4d2257","textHash":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","observedAt":"2026-11-10T10:50:07.644Z"}}],"participationCoverage":{"total":3,"returned":3,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2605.26473","snapshotId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081","contentSha256":"fd53d53b2357e7e8da394b8790465f0cb6104bbeb611a5016a00342eef4d2257","textHash":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","observedAt":"2026-11-10T10:50:07.644Z","basis":"publisher-page-metadata","sourceLinks":44,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/816df1ce8de199aa50964d39a9a04818b2688bca7754d60720c2b73753764c50/source"}},"canonicalUrl":"https://escapemeridian.com/readings/arxiv-eess-sy-orion-enabling-self-adaptive-memory-management-816df1ce","formats":{"html":"/readings/arxiv-eess-sy-orion-enabling-self-adaptive-memory-management-816df1ce","markdown":"/readings/arxiv-eess-sy-orion-enabling-self-adaptive-memory-management-816df1ce.md","json":"/readings/arxiv-eess-sy-orion-enabling-self-adaptive-memory-management-816df1ce.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-continual-learning","name":"Continual learning","explanation":"A model improving from new experience after deployment, the way a new employee learns on the job, without full retraining.","referenceUrl":"https://en.wikipedia.org/wiki/Incremental_learning"},{"id":"term-benchmark","name":"Benchmark","explanation":"A defined set of tests used to compare systems. Results depend on the tasks and measures selected.","referenceUrl":"https://arxiv.org/abs/2211.09110"},{"id":"term-latency","name":"Latency","explanation":"The delay between sending a request and receiving the result.","referenceUrl":"https://developers.google.com/machine-learning/glossary#latency"}],"entities":[{"id":"arxiv","name":"arXiv","kind":"organization","context":"arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.","referenceUrls":["https://arxiv.org/","https://info.arxiv.org/help/api/index.html"],"path":"/entities/arxiv","json":"/entities/arxiv.json","markdown":"/entities/arxiv.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"816df1ce8de199aa50964d39a9a04818b2688bca7754d60720c2b73753764c50","revision":"d618867b-d871-4ed1-8060-3770ebe14632","revisionKind":"publication-id","title":"arXiv eess.SY: Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","url":"https://escapemeridian.com/readings/arxiv-eess-sy-orion-enabling-self-adaptive-memory-management-816df1ce"},"source":{"type":"source","id":"https://arxiv.org/abs/2605.26473","revision":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","revisionKind":"sha256-source-text","title":"Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","url":"https://arxiv.org/abs/2605.26473"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"816df1ce8de199aa50964d39a9a04818b2688bca7754d60720c2b73753764c50","title":"arXiv eess.SY: Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","argument":"In their paper, Zexin Li, Nikil Dutt and Cong Liu propose Orion, a system for managing memory as a device learns from incoming data. Online continual learning (OCL) means adapting a model as new data arrives, rather than training it once on a fixed dataset. The authors argue that memory pressure and changing workloads make fixed, offline memory settings unsuitable for this task on resource-limited devices.\n\nOrion uses URGE, a runtime indicator based on the idea that performance is limited by the scarcest resource, to reallocate memory among batch processing, replay buffers and optimization strategies. It also coordinates these choices at the operating-system and application levels, and uses data prefetching. The authors say they implemented a prototype with Avalanche-lib, evaluated it across OCL algorithms, benchmarks and hardware, and integrated it into an autonomous navigational robot.\n\nThe paper’s central claim is that dynamically managing memory can make on-device OCL more practical while balancing training latency, plasticity and stability. The authors report significant training speedups and minimal runtime, memory and energy overhead, but the abstract does not give the size of those gains. The evidence described here is the authors’ report of evaluations and a prototype; it does not provide the detailed results needed to assess performance across particular settings.","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The supplied evidence is the arXiv abstract, which reports the authors’ findings but does not include numerical speedups or overhead measurements, detailed evaluation results, or comparisons for specific settings.","speaker":{"name":"Li, Zexin, Dutt, Nikil, Liu, Cong","roleAtSource":"Article author"},"source":{"title":"Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","publisher":"arXiv eess.SY","format":"Article","url":"https://arxiv.org/abs/2605.26473","publishedAt":"2026-05-26","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors propose Orion to manage memory dynamically for on-device online continual learning, balancing training latency, plasticity and stability under memory constraints.","excerpt":"a holistic framework designed to co-optimize training latency, plasticity, and stability of state-of-the-art OCL models under strict memory constraints","url":"https://arxiv.org/abs/2605.26473","relationship":"Source-stated"},{"text":"Orion uses URGE to reallocate memory across OCL components and coordinates batch processing, replay buffers and optimization strategies at both operating-system and application levels.","excerpt":"dynamically reallocate memory across OCL components by jointly coordinating batch processing, replay buffers, and optimization strategies at both the OS and application level","url":"https://arxiv.org/abs/2605.26473","relationship":"Source-stated"},{"text":"The authors report evaluating an Orion prototype across OCL algorithms, benchmarks and hardware, and integrating it into an autonomous navigational robot.","excerpt":"evaluated across a diverse range of OCL algorithms, benchmarks, and hardware platforms commonly used in autonomous robotic applications.","url":"https://arxiv.org/abs/2605.26473","relationship":"Source-stated"},{"text":"The authors report significant training speedups with minimal runtime, memory and energy overhead.","excerpt":"achieves significant training speedups while maintaining balanced performance and effectively adapting to various scenarios, all with minimal runtime, memory, and energy overhead","url":"https://arxiv.org/abs/2605.26473","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2605.26473","checkedAt":"2026-11-10T10:50:07.644Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What speedups and runtime, memory and energy overheads did Orion achieve in the evaluated settings?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"arXiv eess.SY: Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","summary":"In their paper, Zexin Li, Nikil Dutt and Cong Liu propose Orion, a system for managing memory as a device learns from incoming data. Online continual learning (OCL) means adapting a model as new data arrives, rather than training it once on a fixed dataset. The authors argue that memory pressure and changing workloads make fixed, offline memory settings unsuitable for this task on resource-limited devices.\n\nOrion uses URGE, a runtime indicator based on the idea that performance is limited by the scarcest resource, to reallocate memory among batch processing, replay buffers and optimization strategies. It also coordinates these choices at the operating-system and application levels, and uses data prefetching. The authors say they implemented a prototype with Avalanche-lib, evaluated it across OCL algorithms, benchmarks and hardware, and integrated it into an autonomous navigational robot.\n\nThe paper’s central claim is that dynamically managing memory can make on-device OCL more practical while balancing training latency, plasticity and stability. The authors report significant training speedups and minimal runtime, memory and energy overhead, but the abstract does not give the size of those gains. The evidence described here is the authors’ report of evaluations and a prototype; it does not provide the detailed results needed to assess performance across particular settings.","question":"What speedups and runtime, memory and energy overheads did Orion achieve in the evaluated settings?","claims":[{"text":"The authors propose Orion to manage memory dynamically for on-device online continual learning, balancing training latency, plasticity and stability under memory constraints.","excerpt":"a holistic framework designed to co-optimize training latency, plasticity, and stability of state-of-the-art OCL models under strict memory constraints","support":"source-stated"},{"text":"Orion uses URGE to reallocate memory across OCL components and coordinates batch processing, replay buffers and optimization strategies at both operating-system and application levels.","excerpt":"dynamically reallocate memory across OCL components by jointly coordinating batch processing, replay buffers, and optimization strategies at both the OS and application level","support":"source-stated"},{"text":"The authors report evaluating an Orion prototype across OCL algorithms, benchmarks and hardware, and integrating it into an autonomous navigational robot.","excerpt":"evaluated across a diverse range of OCL algorithms, benchmarks, and hardware platforms commonly used in autonomous robotic applications.","support":"source-stated"},{"text":"The authors report significant training speedups with minimal runtime, memory and energy overhead.","excerpt":"achieves significant training speedups while maintaining balanced performance and effectively adapting to various scenarios, all with minimal runtime, memory, and energy overhead","support":"source-stated"}],"limitations":["The supplied evidence is the arXiv abstract, which reports the authors’ findings but does not include numerical speedups or overhead measurements, detailed evaluation results, or comparisons for specific settings."],"mediaId":"9046b562174a46d7bcaa04109733dd90c74e0289a9195067f8ea196385b902da","correctionReason":null,"id":"816df1ce8de199aa50964d39a9a04818b2688bca7754d60720c2b73753764c50","publicationId":"d618867b-d871-4ed1-8060-3770ebe14632","version":1,"sourceUrl":"https://arxiv.org/abs/2605.26473","sourceTitle":"Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","publisher":"arXiv eess.SY","publishedAt":null,"originalPublished":{"value":"2026-05-26","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-03T08:58:49.837Z","createdAt":"2026-11-10T09:46:12.064Z","historical":true,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"9046b562174a46d7bcaa04109733dd90c74e0289a9195067f8ea196385b902da","kind":"article","url":"https://arxiv.org/abs/2605.26473","sourceUrl":"https://arxiv.org/abs/2605.26473","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-04T01:46:23.796Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2605.26473","checkedAt":"2026-11-10T10:50:07.644Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"9046b562174a46d7bcaa04109733dd90c74e0289a9195067f8ea196385b902da","confidence":0.75,"provider":"typesafe","model":"jev-1.13.0","inputHash":"8c1dbc0af61efbd208e6ec2d59110ae4b99c0f5b02ffb41069f6be6ecebfb4b7","decidedAt":"2026-11-04T01:46:24.325Z","redecided":{"from":"stream-admission-v5","under":"stream-admission-v5+confidence-0.5"}},"provenance":{"textHash":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","observationId":"ee96c0eb9edeafd26775bdf53837c54274a659dabdc2cefbf7ef9162c54def29","retrievedAt":"2026-11-04T01:46:23.736Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"7c9e2c46-6797-436f-81d3-3fc0e5e22417"},"reviewer":{"model":"gpt-6-luna","runId":"e398f34e-32f3-4182-8e19-c4e3ea4e8a33"}},"audit":{"status":"passed","reviewedAt":"2026-11-10T09:46:12.064Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Li, Zexin","url":null,"kind":"unknown"},{"name":"Dutt, Nikil","url":null,"kind":"unknown"},{"name":"Liu, Cong","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Li, Zexin","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Li, Zexin","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.26473","locator":"head/meta[name=author|citation_author]"},"assertionId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T10:50:07.644Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081","sourceUrl":"https://arxiv.org/abs/2605.26473","contentSha256":"fd53d53b2357e7e8da394b8790465f0cb6104bbeb611a5016a00342eef4d2257","textHash":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","observedAt":"2026-11-10T10:50:07.644Z"}},{"actor":{"name":"Dutt, Nikil","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Dutt, Nikil","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.26473","locator":"head/meta[name=author|citation_author]"},"assertionId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T10:50:07.644Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081","sourceUrl":"https://arxiv.org/abs/2605.26473","contentSha256":"fd53d53b2357e7e8da394b8790465f0cb6104bbeb611a5016a00342eef4d2257","textHash":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","observedAt":"2026-11-10T10:50:07.644Z"}},{"actor":{"name":"Liu, Cong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Liu, Cong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2605.26473","locator":"head/meta[name=author|citation_author]"},"assertionId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-10T10:50:07.644Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081","sourceUrl":"https://arxiv.org/abs/2605.26473","contentSha256":"fd53d53b2357e7e8da394b8790465f0cb6104bbeb611a5016a00342eef4d2257","textHash":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","observedAt":"2026-11-10T10:50:07.644Z"}}],"participationCoverage":{"total":3,"returned":3,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2605.26473","snapshotId":"3c93fe18d187dc8f97af684492b7aa8e9fef1b530dd88cb9d7aa83067f74f081","contentSha256":"fd53d53b2357e7e8da394b8790465f0cb6104bbeb611a5016a00342eef4d2257","textHash":"f1b8fe0196072fde07f793d28e6e7dcd945859f975236236a7156426fcdf9b7d","observedAt":"2026-11-10T10:50:07.644Z","basis":"publisher-page-metadata","sourceLinks":44,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/816df1ce8de199aa50964d39a9a04818b2688bca7754d60720c2b73753764c50/source"}}}],"audits":[{"publicationId":"d618867b-d871-4ed1-8060-3770ebe14632","verdict":"passed","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading correctly attributes the paper and its claims to Zexin Li, Nikil Dutt and Cong Liu. It frames results as the authors' reports, not established facts.","excerpt":"Authors:Zexin Li, Nikil Dutt, Cong Liu"},{"dimension":"summary","status":"pass","explanation":"The summary leads with the actor and the proposal, defines online continual learning in plain language, and explains URGE's scarcest-resource idea. It also keeps the abstract's robotics motivation. Plasticity, stability and replay buffers are kept as exact technical terms and should be covered by term-library tooltips; this is not blocking. The title and question are faithful, and the question asks about a real gap in the abstract.","excerpt":"Online continual learning (OCL) enables real-time adaptation to new data, making it crucial for dynamic robotic applications."},{"dimension":"date","status":"pass","explanation":"originalPublished is 2026-05-26, which matches the v1 submission date. It is not the 2 Nov 2026 revision date or the feed date.","excerpt":"Submitted on 26 May 2026 (v1), last revised 2 Nov 2026 (this version, v2)"},{"dimension":"support","status":"pass","explanation":"Each claim matches the abstract, including URGE, the OS- and application-level coordination, prefetching, the Avalanche-lib prototype, the robot integration, and the reported speedups with minimal overhead. The reading says the size of the gains is not given and invents no benefits or costs.","excerpt":"Orion is integrated into a realistic autonomous navigational robot powered by OCL"},{"dimension":"media","status":"pass","explanation":"The selected media is the verified arXiv abstract page for this paper, so it gives direct access to the original source.","excerpt":null}],"createdAt":"2026-11-10T09:46:12.064Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}