{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QMFIYOEP5JVB7TZXYVGHPMTOKI","short_pith_number":"pith:QMFIYOEP","schema_version":"1.0","canonical_sha256":"830a8c388fea6a1fcf37c54c77b26e520224b3cf29afd242dc00c7f8a85e92fb","source":{"kind":"arxiv","id":"2602.04101","version":2},"attestation_state":"computed","paper":{"title":"Interfaze: The Future of AI is built on Task-Specific Small Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Harsha Vardhan Khurdula, Vineet Agarwal, Yoeven D Khemlani","submitted_at":"2026-02-04T00:36:37Z","abstract_excerpt":"We present Interfaze, a native hybrid model that fuses task-specific deep neural networks (CNNs and DNNs) directly into a transformer decoder through a shared embedding space. Specialized perceptual encoders handle optical character recognition (OCR) over complex multilingual PDFs, open-vocabulary object and graphical user interface (GUI) detection, and multilingual speech recognition with diarization. Each is exposed through a task-specific adapter and can be activated on its own, so a query touches only the parameters it needs. A built-in action foundation supplies a grounded external state:"},"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":"2602.04101","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-02-04T00:36:37Z","cross_cats_sorted":[],"title_canon_sha256":"57b17ed4a68508b37d1a3c4a4903a65ee1a631d02424db6ce5ee167e95ee002c","abstract_canon_sha256":"e915b4b904d8f29de4950a58500dae9a3c85e3fb755f7099413e3b958ef819d7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-04T01:08:12.888900Z","signature_b64":"Dlgw7RfIMxOx2YtZtRYjg20s0fjrL6nIa93Wv1L49NtaiM0J1uWT8bWRjR3H3c638MgP+oehobOo5rOJcoTKDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"830a8c388fea6a1fcf37c54c77b26e520224b3cf29afd242dc00c7f8a85e92fb","last_reissued_at":"2026-06-04T01:08:12.888129Z","signature_status":"signed_v1","first_computed_at":"2026-06-04T01:08:12.888129Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Interfaze: The Future of AI is built on Task-Specific Small Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Harsha Vardhan Khurdula, Vineet Agarwal, Yoeven D Khemlani","submitted_at":"2026-02-04T00:36:37Z","abstract_excerpt":"We present Interfaze, a native hybrid model that fuses task-specific deep neural networks (CNNs and DNNs) directly into a transformer decoder through a shared embedding space. Specialized perceptual encoders handle optical character recognition (OCR) over complex multilingual PDFs, open-vocabulary object and graphical user interface (GUI) detection, and multilingual speech recognition with diarization. Each is exposed through a task-specific adapter and can be activated on its own, so a query touches only the parameters it needs. A built-in action foundation supplies a grounded external state:"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.04101","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/2602.04101/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":"2602.04101","created_at":"2026-06-04T01:08:12.888228+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.04101v2","created_at":"2026-06-04T01:08:12.888228+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.04101","created_at":"2026-06-04T01:08:12.888228+00:00"},{"alias_kind":"pith_short_12","alias_value":"QMFIYOEP5JVB","created_at":"2026-06-04T01:08:12.888228+00:00"},{"alias_kind":"pith_short_16","alias_value":"QMFIYOEP5JVB7TZX","created_at":"2026-06-04T01:08:12.888228+00:00"},{"alias_kind":"pith_short_8","alias_value":"QMFIYOEP","created_at":"2026-06-04T01:08:12.888228+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.20795","citing_title":"Automatic Ontology Construction Using LLMs as an External Layer of Memory, Verification, and Planning for Hybrid Intelligent Systems","ref_index":16,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QMFIYOEP5JVB7TZXYVGHPMTOKI","json":"https://pith.science/pith/QMFIYOEP5JVB7TZXYVGHPMTOKI.json","graph_json":"https://pith.science/api/pith-number/QMFIYOEP5JVB7TZXYVGHPMTOKI/graph.json","events_json":"https://pith.science/api/pith-number/QMFIYOEP5JVB7TZXYVGHPMTOKI/events.json","paper":"https://pith.science/paper/QMFIYOEP"},"agent_actions":{"view_html":"https://pith.science/pith/QMFIYOEP5JVB7TZXYVGHPMTOKI","download_json":"https://pith.science/pith/QMFIYOEP5JVB7TZXYVGHPMTOKI.json","view_paper":"https://pith.science/paper/QMFIYOEP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.04101&json=true","fetch_graph":"https://pith.science/api/pith-number/QMFIYOEP5JVB7TZXYVGHPMTOKI/graph.json","fetch_events":"https://pith.science/api/pith-number/QMFIYOEP5JVB7TZXYVGHPMTOKI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QMFIYOEP5JVB7TZXYVGHPMTOKI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QMFIYOEP5JVB7TZXYVGHPMTOKI/action/storage_attestation","attest_author":"https://pith.science/pith/QMFIYOEP5JVB7TZXYVGHPMTOKI/action/author_attestation","sign_citation":"https://pith.science/pith/QMFIYOEP5JVB7TZXYVGHPMTOKI/action/citation_signature","submit_replication":"https://pith.science/pith/QMFIYOEP5JVB7TZXYVGHPMTOKI/action/replication_record"}},"created_at":"2026-06-04T01:08:12.888228+00:00","updated_at":"2026-06-04T01:08:12.888228+00:00"}