{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.22877", "kind": "paper", "url": "https://arxiv.org/abs/2606.22877"}, {"id": "gh:wenyaxie023/dynamicmem", "kind": "repo", "url": "https://github.com/wenyaxie023/DynamicMem"}, {"kind": "website", "url": "https://wenyaxie023.github.io/DynamicMem/"}], "authors": ["Wenya Xie", "Shengming Zhou", "Zelin Li", "Pouya Parsa", "Shuang Zhou", "Xinheng Ding", "Chinmay Arvind", "Guanchu Wang", "Vladimir Braverman", "Ali Payani", "Yantao Zheng", "Zirui Liu"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.22877", "source_url": "https://arxiv.org/abs/2606.22877"}, "categories": ["General AI", "Language & Knowledge", "cs.CL"], "description": {"en": "DynamicMem evaluates long-horizon memory in LLM agents through a synthetic benchmark with 15 months of multi-app activity per user across 16 applications, including attributes, habits, and preferences that evolve and must be inferred from scattered evidence. Scoring occurs at quarterly checkpoints, tracking performance as history grows."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.22877", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.22877", "title": "DynamicMem"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["University of Minnesota", "University of North Carolina at Charlotte", "Johns Hopkins University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.22877", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.22877", "source_url": "https://openalex.org/W7165636227"}, "key": "claire-radar:2606.22877", "modality": null, "name": "DynamicMem", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "unknown", "origin_record_id": "bm_dynamicmem_faf51af5", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "53750beb48d9ad7296ff527855678b7306ce3434ccbfcc3bcd3ca760e8c32733", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.22877"}, "publisher": null, "released": "2026-06-22", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.22877", "source_url": "https://arxiv.org/abs/2606.22877"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-22877", "source": "claire_radar", "source_benchmark_id": "2606.22877", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Machine Generated", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/wenyaxie023/DynamicMem", "githubScope": "benchmark_repo", "githubStars": 19, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2606.22877", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "state": "unavailable"}}, "status": "fresh"}, "capabilities": [], "confidence": "High", "construction": "Original Synthetic", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "a12059d7f5c24799fcc4e6e7d2f7faa87292b2a9544de5f4f2694b6e6459213a", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "DynamicMem evaluates long-horizon memory in LLM agents through a synthetic benchmark with 15 months of multi-app activity per user across 16 applications, including attributes, habits, and preferences that evolve and must be inferred from scattered evidence. Scoring occurs at quarterly checkpoints, tracking performance as history grows.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact coined title identity tied to benchmark evidence", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We introduce DynamicMem, a synthetic benchmark that constructs 15 months of activity per user, providing long-term multi-app data that real users' privacy keeps out of reach."}, "familyId": "bmf_bbff9299325a", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [{"evidence": [{"basis": "linked-from-project", "linkedUrl": "https://github.com/wenyaxie023/DynamicMem", "url": "https://wenyaxie023.github.io/DynamicMem/"}], "scope": "benchmark_repo", "url": "https://github.com/wenyaxie023/DynamicMem", "verification": "source-linked"}], "status": "linked"}, "heat": null, "id": "bm_dynamicmem_faf51af5", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": "https://github.com/wenyaxie023/DynamicMem", "data": null, "hfPaper": "https://huggingface.co/papers/2606.22877", "paper": "https://arxiv.org/abs/2606.22877", "pdf": "https://arxiv.org/pdf/2606.22877", "project": "https://wenyaxie023.github.io/DynamicMem/", "report": "https://arxiv.org/abs/2606.22877"}, "metrics": [], "motivation": "LLM agents increasingly act as personal assistants that must remember a user's profile over months: who they are (attributes), what they routinely do (habits), and what they prefer (preferences), and keep it updated as jobs, routines, and tastes drift.", "name": "DynamicMem", "oneLine": "DynamicMem evaluates long-horizon memory in LLM agents through a synthetic benchmark with 15 months of multi-app activity per user across 16 applications, including attributes, habits, and preferences that evolve and must be inferred from scattered evidence. Scoring occurs at quarterly checkpoints, tracking performance as history grows.", "paperTitle": "DynamicMem: A Long-Horizon Memory Benchmark in Real-World Settings", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.22877", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {}, "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-06-22", "source": {"authors": ["Wenya Xie", "Shengming Zhou", "Zelin Li", "Pouya Parsa", "Shuang Zhou", "Xinheng Ding", "Chinmay Arvind", "Guanchu Wang", "Vladimir Braverman", "Ali Payani", "Yantao Zheng", "Zirui Liu"], "categories": ["cs.CL"], "id": "2606.22877", "title": "DynamicMem: A Long-Horizon Memory Benchmark in Real-World Settings", "type": "arxiv", "url": "https://arxiv.org/abs/2606.22877"}, "sourceUpdatedAt": "2026-06-22T05:41:20Z", "topics": ["cs.CL"], "venueAttempts": [], "whyItMatters": "Existing memory benchmarks use short, simplified interactions, missing real-world complexity. DynamicMem provides a long-horizon, multi-app evaluation that yields insights into memory failures, such as degradation with history length and retrieval-driven errors, which can guide improvements in memory systems for personal assistants."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
