{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:5YGJCJSHDBNFVPP3GOIVEAMNOP","short_pith_number":"pith:5YGJCJSH","schema_version":"1.0","canonical_sha256":"ee0c912647185a5abdfb339152018d73df79ef83b451a74546d90ca8d3a04d6b","source":{"kind":"arxiv","id":"2006.15078","version":1},"attestation_state":"computed","paper":{"title":"Continual Learning from the Perspective of Compression","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NE","stat.ML"],"primary_cat":"cs.LG","authors_text":"Min Lin, Xu He","submitted_at":"2020-06-26T16:15:49Z","abstract_excerpt":"Connectionist models such as neural networks suffer from catastrophic forgetting. In this work, we study this problem from the perspective of information theory and define forgetting as the increase of description lengths of previous data when they are compressed with a sequentially learned model. In addition, we show that continual learning approaches based on variational posterior approximation and generative replay can be considered as approximations to two prequential coding methods in compression, namely, the Bayesian mixture code and maximum likelihood (ML) plug-in code. We compare these"},"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":"2006.15078","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-06-26T16:15:49Z","cross_cats_sorted":["cs.NE","stat.ML"],"title_canon_sha256":"53ee90295dceeb6a6b3783ca4e815856f7a8fd085a5554a93a3453e69277f838","abstract_canon_sha256":"26611fc622b7d113df37136253dd186aff80694f0c2e58b4d914b39e14a6daea"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:13:49.442449Z","signature_b64":"rttnyhMlDFg9outbWuExeNsDUY0DEjPjhQXqmIMsSkYZeZrGnOlfsjClh/W4LXog3Wzyx6Yxu98SZqx1l897Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ee0c912647185a5abdfb339152018d73df79ef83b451a74546d90ca8d3a04d6b","last_reissued_at":"2026-07-05T01:13:49.442093Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:13:49.442093Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Continual Learning from the Perspective of Compression","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NE","stat.ML"],"primary_cat":"cs.LG","authors_text":"Min Lin, Xu He","submitted_at":"2020-06-26T16:15:49Z","abstract_excerpt":"Connectionist models such as neural networks suffer from catastrophic forgetting. In this work, we study this problem from the perspective of information theory and define forgetting as the increase of description lengths of previous data when they are compressed with a sequentially learned model. In addition, we show that continual learning approaches based on variational posterior approximation and generative replay can be considered as approximations to two prequential coding methods in compression, namely, the Bayesian mixture code and maximum likelihood (ML) plug-in code. We compare these"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2006.15078","kind":"arxiv","version":1},"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/2006.15078/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":"2006.15078","created_at":"2026-07-05T01:13:49.442144+00:00"},{"alias_kind":"arxiv_version","alias_value":"2006.15078v1","created_at":"2026-07-05T01:13:49.442144+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2006.15078","created_at":"2026-07-05T01:13:49.442144+00:00"},{"alias_kind":"pith_short_12","alias_value":"5YGJCJSHDBNF","created_at":"2026-07-05T01:13:49.442144+00:00"},{"alias_kind":"pith_short_16","alias_value":"5YGJCJSHDBNFVPP3","created_at":"2026-07-05T01:13:49.442144+00:00"},{"alias_kind":"pith_short_8","alias_value":"5YGJCJSH","created_at":"2026-07-05T01:13:49.442144+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.10861","citing_title":"Dynamic Continual Learning: Harnessing Parameter Uncertainty for Improved Network Adaptation","ref_index":22,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5YGJCJSHDBNFVPP3GOIVEAMNOP","json":"https://pith.science/pith/5YGJCJSHDBNFVPP3GOIVEAMNOP.json","graph_json":"https://pith.science/api/pith-number/5YGJCJSHDBNFVPP3GOIVEAMNOP/graph.json","events_json":"https://pith.science/api/pith-number/5YGJCJSHDBNFVPP3GOIVEAMNOP/events.json","paper":"https://pith.science/paper/5YGJCJSH"},"agent_actions":{"view_html":"https://pith.science/pith/5YGJCJSHDBNFVPP3GOIVEAMNOP","download_json":"https://pith.science/pith/5YGJCJSHDBNFVPP3GOIVEAMNOP.json","view_paper":"https://pith.science/paper/5YGJCJSH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2006.15078&json=true","fetch_graph":"https://pith.science/api/pith-number/5YGJCJSHDBNFVPP3GOIVEAMNOP/graph.json","fetch_events":"https://pith.science/api/pith-number/5YGJCJSHDBNFVPP3GOIVEAMNOP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5YGJCJSHDBNFVPP3GOIVEAMNOP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5YGJCJSHDBNFVPP3GOIVEAMNOP/action/storage_attestation","attest_author":"https://pith.science/pith/5YGJCJSHDBNFVPP3GOIVEAMNOP/action/author_attestation","sign_citation":"https://pith.science/pith/5YGJCJSHDBNFVPP3GOIVEAMNOP/action/citation_signature","submit_replication":"https://pith.science/pith/5YGJCJSHDBNFVPP3GOIVEAMNOP/action/replication_record"}},"created_at":"2026-07-05T01:13:49.442144+00:00","updated_at":"2026-07-05T01:13:49.442144+00:00"}