{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:CEBYBTOC2NSMKREY3TJH4TJOY3","short_pith_number":"pith:CEBYBTOC","schema_version":"1.0","canonical_sha256":"110380cdc2d364c54498dcd27e4d2ec6dccd6bbd610b691a673442e98778cb5e","source":{"kind":"arxiv","id":"2402.09236","version":2},"attestation_state":"computed","paper":{"title":"Learning Interpretable Concepts: Unifying Causal Representation Learning and Foundation Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Bernhard Sch\\\"olkopf, Bryon Aragam, Goutham Rajendran, Pradeep Ravikumar, Simon Buchholz","submitted_at":"2024-02-14T15:23:59Z","abstract_excerpt":"To build intelligent machine learning systems, there are two broad approaches. One approach is to build inherently interpretable models, as endeavored by the growing field of causal representation learning. The other approach is to build highly-performant foundation models and then invest efforts into understanding how they work. In this work, we relate these two approaches and study how to learn human-interpretable concepts from data. Weaving together ideas from both fields, we formally define a notion of concepts and show that they can be provably recovered from diverse data. Experiments on "},"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":"2402.09236","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-14T15:23:59Z","cross_cats_sorted":["cs.AI","math.ST","stat.ML","stat.TH"],"title_canon_sha256":"c2fbff902c7cd8a68dc5e298a04df46243937fffdd454c509dacc280ffcd34a7","abstract_canon_sha256":"750a7c983fb2f1d5e1eeec55cf384b3c0112703c3b9c1e791ba32e636d4aa43c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:45:44.466055Z","signature_b64":"XKR8ac+XG3PEiH2Vvy1qMwD9zloO54g8krfRNuCiJhvOMoJeuyH82fdg1+CUFjePKM6cpM/yTPW4qaGfk+eNBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"110380cdc2d364c54498dcd27e4d2ec6dccd6bbd610b691a673442e98778cb5e","last_reissued_at":"2026-07-05T09:45:44.465595Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:45:44.465595Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning Interpretable Concepts: Unifying Causal Representation Learning and Foundation Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Bernhard Sch\\\"olkopf, Bryon Aragam, Goutham Rajendran, Pradeep Ravikumar, Simon Buchholz","submitted_at":"2024-02-14T15:23:59Z","abstract_excerpt":"To build intelligent machine learning systems, there are two broad approaches. One approach is to build inherently interpretable models, as endeavored by the growing field of causal representation learning. The other approach is to build highly-performant foundation models and then invest efforts into understanding how they work. In this work, we relate these two approaches and study how to learn human-interpretable concepts from data. Weaving together ideas from both fields, we formally define a notion of concepts and show that they can be provably recovered from diverse data. Experiments on "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.09236","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/2402.09236/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":"2402.09236","created_at":"2026-07-05T09:45:44.465653+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.09236v2","created_at":"2026-07-05T09:45:44.465653+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.09236","created_at":"2026-07-05T09:45:44.465653+00:00"},{"alias_kind":"pith_short_12","alias_value":"CEBYBTOC2NSM","created_at":"2026-07-05T09:45:44.465653+00:00"},{"alias_kind":"pith_short_16","alias_value":"CEBYBTOC2NSMKREY","created_at":"2026-07-05T09:45:44.465653+00:00"},{"alias_kind":"pith_short_8","alias_value":"CEBYBTOC","created_at":"2026-07-05T09:45:44.465653+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18284","citing_title":"Breaking the Solver Bottleneck: Training Task Generators at the Learnable Frontier","ref_index":120,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05966","citing_title":"Causal Scaffolding for Physical Reasoning: A Benchmark for Causally-Informed Physical World Understanding in VLMs","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02640","citing_title":"Trustworthy AI Suffers from Invariance Conflicts and Causality is The Solution","ref_index":42,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CEBYBTOC2NSMKREY3TJH4TJOY3","json":"https://pith.science/pith/CEBYBTOC2NSMKREY3TJH4TJOY3.json","graph_json":"https://pith.science/api/pith-number/CEBYBTOC2NSMKREY3TJH4TJOY3/graph.json","events_json":"https://pith.science/api/pith-number/CEBYBTOC2NSMKREY3TJH4TJOY3/events.json","paper":"https://pith.science/paper/CEBYBTOC"},"agent_actions":{"view_html":"https://pith.science/pith/CEBYBTOC2NSMKREY3TJH4TJOY3","download_json":"https://pith.science/pith/CEBYBTOC2NSMKREY3TJH4TJOY3.json","view_paper":"https://pith.science/paper/CEBYBTOC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.09236&json=true","fetch_graph":"https://pith.science/api/pith-number/CEBYBTOC2NSMKREY3TJH4TJOY3/graph.json","fetch_events":"https://pith.science/api/pith-number/CEBYBTOC2NSMKREY3TJH4TJOY3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CEBYBTOC2NSMKREY3TJH4TJOY3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CEBYBTOC2NSMKREY3TJH4TJOY3/action/storage_attestation","attest_author":"https://pith.science/pith/CEBYBTOC2NSMKREY3TJH4TJOY3/action/author_attestation","sign_citation":"https://pith.science/pith/CEBYBTOC2NSMKREY3TJH4TJOY3/action/citation_signature","submit_replication":"https://pith.science/pith/CEBYBTOC2NSMKREY3TJH4TJOY3/action/replication_record"}},"created_at":"2026-07-05T09:45:44.465653+00:00","updated_at":"2026-07-05T09:45:44.465653+00:00"}