{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NH6DEQ4PHHQFCQASODCS5RZR55","short_pith_number":"pith:NH6DEQ4P","schema_version":"1.0","canonical_sha256":"69fc32438f39e051401270c52ec731ef726c2e783a6971dd84b437ab2ed49468","source":{"kind":"arxiv","id":"2411.07483","version":2},"attestation_state":"computed","paper":{"title":"Quantifying Knowledge Distillation Using Partial Information Decomposition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","cs.IT","cs.LG","eess.IV","math.IT"],"primary_cat":"stat.ML","authors_text":"Barproda Halder, Faisal Hamman, Ilia Sucholutsky, Pasan Dissanayake, Qiuyi Zhang, Sanghamitra Dutta","submitted_at":"2024-11-12T02:12:41Z","abstract_excerpt":"Knowledge distillation deploys complex machine learning models in resource-constrained environments by training a smaller student model to emulate internal representations of a complex teacher model. However, the teacher's representations can also encode nuisance or additional information not relevant to the downstream task. Distilling such irrelevant information can actually impede the performance of a capacity-limited student model. This observation motivates our primary question: What are the information-theoretic limits of knowledge distillation? To this end, we leverage Partial Informatio"},"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":"2411.07483","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2024-11-12T02:12:41Z","cross_cats_sorted":["cs.CV","cs.IT","cs.LG","eess.IV","math.IT"],"title_canon_sha256":"b14ea8a551d0f7a9e6c3fa42d3578dee5605accad93ca8e83b074fb3457817f6","abstract_canon_sha256":"77ba69ce483fa6b2672ec2e8f654b7364f9372f88551c662707d79e2528b43ab"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:44:19.774143Z","signature_b64":"kcGLmMCQQO5sj0C/tEWEFMYFs0R4bf2e2eMAXhG0D2/mzgiw43Nz2rlG7F/33ujcrn/TZ7EmSTKAOcqufn7zCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"69fc32438f39e051401270c52ec731ef726c2e783a6971dd84b437ab2ed49468","last_reissued_at":"2026-07-05T10:44:19.773713Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:44:19.773713Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Quantifying Knowledge Distillation Using Partial Information Decomposition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","cs.IT","cs.LG","eess.IV","math.IT"],"primary_cat":"stat.ML","authors_text":"Barproda Halder, Faisal Hamman, Ilia Sucholutsky, Pasan Dissanayake, Qiuyi Zhang, Sanghamitra Dutta","submitted_at":"2024-11-12T02:12:41Z","abstract_excerpt":"Knowledge distillation deploys complex machine learning models in resource-constrained environments by training a smaller student model to emulate internal representations of a complex teacher model. However, the teacher's representations can also encode nuisance or additional information not relevant to the downstream task. Distilling such irrelevant information can actually impede the performance of a capacity-limited student model. This observation motivates our primary question: What are the information-theoretic limits of knowledge distillation? To this end, we leverage Partial Informatio"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.07483","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/2411.07483/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":"2411.07483","created_at":"2026-07-05T10:44:19.773774+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.07483v2","created_at":"2026-07-05T10:44:19.773774+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.07483","created_at":"2026-07-05T10:44:19.773774+00:00"},{"alias_kind":"pith_short_12","alias_value":"NH6DEQ4PHHQF","created_at":"2026-07-05T10:44:19.773774+00:00"},{"alias_kind":"pith_short_16","alias_value":"NH6DEQ4PHHQFCQAS","created_at":"2026-07-05T10:44:19.773774+00:00"},{"alias_kind":"pith_short_8","alias_value":"NH6DEQ4P","created_at":"2026-07-05T10:44:19.773774+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.04555","citing_title":"Decomposing Multivariate Information Rates in Networks of Random Processes","ref_index":23,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NH6DEQ4PHHQFCQASODCS5RZR55","json":"https://pith.science/pith/NH6DEQ4PHHQFCQASODCS5RZR55.json","graph_json":"https://pith.science/api/pith-number/NH6DEQ4PHHQFCQASODCS5RZR55/graph.json","events_json":"https://pith.science/api/pith-number/NH6DEQ4PHHQFCQASODCS5RZR55/events.json","paper":"https://pith.science/paper/NH6DEQ4P"},"agent_actions":{"view_html":"https://pith.science/pith/NH6DEQ4PHHQFCQASODCS5RZR55","download_json":"https://pith.science/pith/NH6DEQ4PHHQFCQASODCS5RZR55.json","view_paper":"https://pith.science/paper/NH6DEQ4P","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.07483&json=true","fetch_graph":"https://pith.science/api/pith-number/NH6DEQ4PHHQFCQASODCS5RZR55/graph.json","fetch_events":"https://pith.science/api/pith-number/NH6DEQ4PHHQFCQASODCS5RZR55/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NH6DEQ4PHHQFCQASODCS5RZR55/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NH6DEQ4PHHQFCQASODCS5RZR55/action/storage_attestation","attest_author":"https://pith.science/pith/NH6DEQ4PHHQFCQASODCS5RZR55/action/author_attestation","sign_citation":"https://pith.science/pith/NH6DEQ4PHHQFCQASODCS5RZR55/action/citation_signature","submit_replication":"https://pith.science/pith/NH6DEQ4PHHQFCQASODCS5RZR55/action/replication_record"}},"created_at":"2026-07-05T10:44:19.773774+00:00","updated_at":"2026-07-05T10:44:19.773774+00:00"}