{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:VEB5UKV64NK3VLTIHKTQ4L2ODE","short_pith_number":"pith:VEB5UKV6","schema_version":"1.0","canonical_sha256":"a903da2abee355baae683aa70e2f4e19052f6a7069dd10b4702404972d954cd6","source":{"kind":"arxiv","id":"2412.17737","version":6},"attestation_state":"computed","paper":{"title":"Contextual Feedback Loops: Amplifying Deep Reasoning with Iterative Top-Down Feedback","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jacob Fein-Ashley, Rajgopal Kannan, Viktor Prasanna","submitted_at":"2024-12-23T17:36:51Z","abstract_excerpt":"Conventional deep networks rely on one-way backpropagation that overlooks reconciling high-level predictions with lower-level representations. We propose \\emph{Contextual Feedback Loops} (CFLs), a lightweight mechanism that re-injects top-down context into earlier layers for iterative refinement. Concretely, CFLs map the network's prediction to a compact \\emph{context vector}, which is fused back into each layer via gating adapters. Unrolled over multiple feedback steps, CFLs unify feed-forward and feedback-driven inference, letting top-level outputs continually refine lower-level features. De"},"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":"2412.17737","kind":"arxiv","version":6},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-12-23T17:36:51Z","cross_cats_sorted":[],"title_canon_sha256":"4dddee81dca52f2faaba81f7df3d8dec8c7e92e3dbb03b0d11c10d962b58e883","abstract_canon_sha256":"e5c26b4b437cc1586652850b02ea563d60e000e1e07a907186fa73c25e2b2bec"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:55:28.586065Z","signature_b64":"Wbtb3K3qC1GzhOPubMPm9cCgafgoAF6hkTAgDYzecJpmRKvrsUcZyH/xu1jRWHvsktTNd5y5GldwnTs7kG2TCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a903da2abee355baae683aa70e2f4e19052f6a7069dd10b4702404972d954cd6","last_reissued_at":"2026-07-05T10:55:28.585557Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:55:28.585557Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Contextual Feedback Loops: Amplifying Deep Reasoning with Iterative Top-Down Feedback","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jacob Fein-Ashley, Rajgopal Kannan, Viktor Prasanna","submitted_at":"2024-12-23T17:36:51Z","abstract_excerpt":"Conventional deep networks rely on one-way backpropagation that overlooks reconciling high-level predictions with lower-level representations. We propose \\emph{Contextual Feedback Loops} (CFLs), a lightweight mechanism that re-injects top-down context into earlier layers for iterative refinement. Concretely, CFLs map the network's prediction to a compact \\emph{context vector}, which is fused back into each layer via gating adapters. Unrolled over multiple feedback steps, CFLs unify feed-forward and feedback-driven inference, letting top-level outputs continually refine lower-level features. De"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.17737","kind":"arxiv","version":6},"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/2412.17737/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":"2412.17737","created_at":"2026-07-05T10:55:28.585617+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.17737v6","created_at":"2026-07-05T10:55:28.585617+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.17737","created_at":"2026-07-05T10:55:28.585617+00:00"},{"alias_kind":"pith_short_12","alias_value":"VEB5UKV64NK3","created_at":"2026-07-05T10:55:28.585617+00:00"},{"alias_kind":"pith_short_16","alias_value":"VEB5UKV64NK3VLTI","created_at":"2026-07-05T10:55:28.585617+00:00"},{"alias_kind":"pith_short_8","alias_value":"VEB5UKV6","created_at":"2026-07-05T10:55:28.585617+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.03787","citing_title":"Iterate to Accelerate: A Unified Framework for Iterative Reasoning and Feedback Convergence","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VEB5UKV64NK3VLTIHKTQ4L2ODE","json":"https://pith.science/pith/VEB5UKV64NK3VLTIHKTQ4L2ODE.json","graph_json":"https://pith.science/api/pith-number/VEB5UKV64NK3VLTIHKTQ4L2ODE/graph.json","events_json":"https://pith.science/api/pith-number/VEB5UKV64NK3VLTIHKTQ4L2ODE/events.json","paper":"https://pith.science/paper/VEB5UKV6"},"agent_actions":{"view_html":"https://pith.science/pith/VEB5UKV64NK3VLTIHKTQ4L2ODE","download_json":"https://pith.science/pith/VEB5UKV64NK3VLTIHKTQ4L2ODE.json","view_paper":"https://pith.science/paper/VEB5UKV6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.17737&json=true","fetch_graph":"https://pith.science/api/pith-number/VEB5UKV64NK3VLTIHKTQ4L2ODE/graph.json","fetch_events":"https://pith.science/api/pith-number/VEB5UKV64NK3VLTIHKTQ4L2ODE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VEB5UKV64NK3VLTIHKTQ4L2ODE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VEB5UKV64NK3VLTIHKTQ4L2ODE/action/storage_attestation","attest_author":"https://pith.science/pith/VEB5UKV64NK3VLTIHKTQ4L2ODE/action/author_attestation","sign_citation":"https://pith.science/pith/VEB5UKV64NK3VLTIHKTQ4L2ODE/action/citation_signature","submit_replication":"https://pith.science/pith/VEB5UKV64NK3VLTIHKTQ4L2ODE/action/replication_record"}},"created_at":"2026-07-05T10:55:28.585617+00:00","updated_at":"2026-07-05T10:55:28.585617+00:00"}