{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:H7XUMELKY2HENPRCDCWK7SNETS","short_pith_number":"pith:H7XUMELK","schema_version":"1.0","canonical_sha256":"3fef46116ac68e46be2218acafc9a49cbeb80379b564c743d5c6215587fd5f11","source":{"kind":"arxiv","id":"2602.07885","version":2},"attestation_state":"computed","paper":{"title":"MemFly: On-the-Fly Memory Optimization via Information Bottleneck","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Sirui Han, Wei Xue, Xianzhang Jia, Yike Guo, Zhenbo Song, Zhenyuan Zhang, Zhiqin Yang","submitted_at":"2026-02-08T09:37:25Z","abstract_excerpt":"Long-term memory enables large language model agents to tackle complex tasks through historical interactions. However, existing frameworks encounter a fundamental dilemma between compressing redundant information efficiently and maintaining precise retrieval for downstream tasks. To bridge this gap, we propose MemFly, a framework grounded in information bottleneck principles that facilitates on-the-fly memory evolution for LLMs. Our approach minimizes compression entropy while maximizing relevance entropy via a gradient-free optimizer, constructing a stratified memory structure for efficient s"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2602.07885","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-02-08T09:37:25Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"6d439f5bdf6fea41618b2ee011c4377a432dee4f643325d625f06d9a05e3c927","abstract_canon_sha256":"a5cd0b9add16200b5e008a8e792fa7c78d46d96d8031b154589864012dd04ee7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-06T01:38:02.646038Z","signature_b64":"0aHwG2z0YceW5IFKjX/M6JQOIkETfvoTvxsv8hpCJTww2Afj4ZYKh0VQbqGiLckcndOHQ0pu+TekYRqNX5/eDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3fef46116ac68e46be2218acafc9a49cbeb80379b564c743d5c6215587fd5f11","last_reissued_at":"2026-08-06T01:38:02.644476Z","signature_status":"signed_v1","first_computed_at":"2026-08-06T01:38:02.644476Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MemFly: On-the-Fly Memory Optimization via Information Bottleneck","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Sirui Han, Wei Xue, Xianzhang Jia, Yike Guo, Zhenbo Song, Zhenyuan Zhang, Zhiqin Yang","submitted_at":"2026-02-08T09:37:25Z","abstract_excerpt":"Long-term memory enables large language model agents to tackle complex tasks through historical interactions. However, existing frameworks encounter a fundamental dilemma between compressing redundant information efficiently and maintaining precise retrieval for downstream tasks. To bridge this gap, we propose MemFly, a framework grounded in information bottleneck principles that facilitates on-the-fly memory evolution for LLMs. Our approach minimizes compression entropy while maximizing relevance entropy via a gradient-free optimizer, constructing a stratified memory structure for efficient s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.07885","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2602.07885/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2602.07885","created_at":"2026-08-06T01:38:02.646285+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.07885v2","created_at":"2026-08-06T01:38:02.646285+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.07885","created_at":"2026-08-06T01:38:02.646285+00:00"},{"alias_kind":"pith_short_12","alias_value":"H7XUMELKY2HE","created_at":"2026-08-06T01:38:02.646285+00:00"},{"alias_kind":"pith_short_16","alias_value":"H7XUMELKY2HENPRC","created_at":"2026-08-06T01:38:02.646285+00:00"},{"alias_kind":"pith_short_8","alias_value":"H7XUMELK","created_at":"2026-08-06T01:38:02.646285+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.01480","citing_title":"Procedural Memory Distillation: Online Reflection for Self-Improving Language Models","ref_index":52,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/H7XUMELKY2HENPRCDCWK7SNETS","json":"https://pith.science/pith/H7XUMELKY2HENPRCDCWK7SNETS.json","graph_json":"https://pith.science/api/pith-number/H7XUMELKY2HENPRCDCWK7SNETS/graph.json","events_json":"https://pith.science/api/pith-number/H7XUMELKY2HENPRCDCWK7SNETS/events.json","paper":"https://pith.science/paper/H7XUMELK"},"agent_actions":{"view_html":"https://pith.science/pith/H7XUMELKY2HENPRCDCWK7SNETS","download_json":"https://pith.science/pith/H7XUMELKY2HENPRCDCWK7SNETS.json","view_paper":"https://pith.science/paper/H7XUMELK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.07885&json=true","fetch_graph":"https://pith.science/api/pith-number/H7XUMELKY2HENPRCDCWK7SNETS/graph.json","fetch_events":"https://pith.science/api/pith-number/H7XUMELKY2HENPRCDCWK7SNETS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/H7XUMELKY2HENPRCDCWK7SNETS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/H7XUMELKY2HENPRCDCWK7SNETS/action/storage_attestation","attest_author":"https://pith.science/pith/H7XUMELKY2HENPRCDCWK7SNETS/action/author_attestation","sign_citation":"https://pith.science/pith/H7XUMELKY2HENPRCDCWK7SNETS/action/citation_signature","submit_replication":"https://pith.science/pith/H7XUMELKY2HENPRCDCWK7SNETS/action/replication_record"}},"created_at":"2026-08-06T01:38:02.646285+00:00","updated_at":"2026-08-06T01:38:02.646285+00:00"}