{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:U6SLNXK3HAKNVPBPUFECOIQAVO","short_pith_number":"pith:U6SLNXK3","schema_version":"1.0","canonical_sha256":"a7a4b6dd5b3814dabc2fa148272200ab895bbb493e1445bbde10b7890069c6b7","source":{"kind":"arxiv","id":"2410.06405","version":2},"attestation_state":"computed","paper":{"title":"Tackling the Abstraction and Reasoning Corpus with Vision Transformers: the Importance of 2D Representation, Positions, and Objects","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Elias Boutros Khalil, Scott Sanner, Wenhao Li, Yudong Xu","submitted_at":"2024-10-08T22:25:34Z","abstract_excerpt":"The Abstraction and Reasoning Corpus (ARC) is a popular benchmark focused on visual reasoning in the evaluation of Artificial Intelligence systems. In its original framing, an ARC task requires solving a program synthesis problem over small 2D images using a few input-output training pairs. In this work, we adopt the recently popular data-driven approach to the ARC and ask whether a Vision Transformer (ViT) can learn the implicit mapping, from input image to output image, that underlies the task. We show that a ViT -- otherwise a state-of-the-art model for images -- fails dramatically on most "},"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":"2410.06405","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-10-08T22:25:34Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9f5fb59235ae6e5ef174ea4fb82abe54d584ea7a1c49304e8911357e33259321","abstract_canon_sha256":"09f1322f86274c638541df9d57828f2cdff1daec662831249c5fc92f8369f72d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:37:42.586346Z","signature_b64":"s/BUJ5OMOHpVwMW/rojVZZ9lBSEmVu4FhOGiKSYJrZrsR8xEFLDFivtIYxGEDVf0JZeXGfZzh4w3BgPkQm0hCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a7a4b6dd5b3814dabc2fa148272200ab895bbb493e1445bbde10b7890069c6b7","last_reissued_at":"2026-07-05T11:37:42.585753Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:37:42.585753Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Tackling the Abstraction and Reasoning Corpus with Vision Transformers: the Importance of 2D Representation, Positions, and Objects","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Elias Boutros Khalil, Scott Sanner, Wenhao Li, Yudong Xu","submitted_at":"2024-10-08T22:25:34Z","abstract_excerpt":"The Abstraction and Reasoning Corpus (ARC) is a popular benchmark focused on visual reasoning in the evaluation of Artificial Intelligence systems. In its original framing, an ARC task requires solving a program synthesis problem over small 2D images using a few input-output training pairs. In this work, we adopt the recently popular data-driven approach to the ARC and ask whether a Vision Transformer (ViT) can learn the implicit mapping, from input image to output image, that underlies the task. We show that a ViT -- otherwise a state-of-the-art model for images -- fails dramatically on most "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.06405","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/2410.06405/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":"2410.06405","created_at":"2026-07-05T11:37:42.585817+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.06405v2","created_at":"2026-07-05T11:37:42.585817+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.06405","created_at":"2026-07-05T11:37:42.585817+00:00"},{"alias_kind":"pith_short_12","alias_value":"U6SLNXK3HAKN","created_at":"2026-07-05T11:37:42.585817+00:00"},{"alias_kind":"pith_short_16","alias_value":"U6SLNXK3HAKNVPBP","created_at":"2026-07-05T11:37:42.585817+00:00"},{"alias_kind":"pith_short_8","alias_value":"U6SLNXK3","created_at":"2026-07-05T11:37:42.585817+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.17482","citing_title":"From Reasoning to Generalization: Knowledge-Augmented LLMs for ARC Benchmark","ref_index":41,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/U6SLNXK3HAKNVPBPUFECOIQAVO","json":"https://pith.science/pith/U6SLNXK3HAKNVPBPUFECOIQAVO.json","graph_json":"https://pith.science/api/pith-number/U6SLNXK3HAKNVPBPUFECOIQAVO/graph.json","events_json":"https://pith.science/api/pith-number/U6SLNXK3HAKNVPBPUFECOIQAVO/events.json","paper":"https://pith.science/paper/U6SLNXK3"},"agent_actions":{"view_html":"https://pith.science/pith/U6SLNXK3HAKNVPBPUFECOIQAVO","download_json":"https://pith.science/pith/U6SLNXK3HAKNVPBPUFECOIQAVO.json","view_paper":"https://pith.science/paper/U6SLNXK3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.06405&json=true","fetch_graph":"https://pith.science/api/pith-number/U6SLNXK3HAKNVPBPUFECOIQAVO/graph.json","fetch_events":"https://pith.science/api/pith-number/U6SLNXK3HAKNVPBPUFECOIQAVO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/U6SLNXK3HAKNVPBPUFECOIQAVO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/U6SLNXK3HAKNVPBPUFECOIQAVO/action/storage_attestation","attest_author":"https://pith.science/pith/U6SLNXK3HAKNVPBPUFECOIQAVO/action/author_attestation","sign_citation":"https://pith.science/pith/U6SLNXK3HAKNVPBPUFECOIQAVO/action/citation_signature","submit_replication":"https://pith.science/pith/U6SLNXK3HAKNVPBPUFECOIQAVO/action/replication_record"}},"created_at":"2026-07-05T11:37:42.585817+00:00","updated_at":"2026-07-05T11:37:42.585817+00:00"}