{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:N76ISPAOT3XPA43A5XGLQPLGI3","short_pith_number":"pith:N76ISPAO","schema_version":"1.0","canonical_sha256":"6ffc893c0e9eeef07360edccb83d6646d16476ec7530a391490932d81169859b","source":{"kind":"arxiv","id":"2605.15216","version":2},"attestation_state":"computed","paper":{"title":"Hardware-Software Co-Design of Scalable, Energy-Efficient Analog Recurrent Computations","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AR","authors_text":"Arthur Fyon, Damien Ernst, Guillaume Drion, Jean-Michel Redout\\'e, Julien Brandoit, Loris Mendolia","submitted_at":"2026-05-12T09:44:32Z","abstract_excerpt":"Always-on AI applications, from environmental sensors to biomedical implants, require ultra-low power consumption. Analog circuits offer a path to sub-microwatt inference, yet existing analog implementations are limited to feedforward architectures: extending them to recurrent dynamics has been considered impractical due to noise accumulation through temporal feedback. We demonstrate that this barrier can be overcome through hardware-software co-design. Specifically, we identify that Bistable Memory Recurrent Units (BMRUs), a class of Recurrent Neural Networks (RNNs) with discrete-valued outpu"},"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":"2605.15216","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.AR","submitted_at":"2026-05-12T09:44:32Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"fce78fea8c85348420bb216f0fa86371d291c32af80768476827d19ea2cf9f9f","abstract_canon_sha256":"8e226218cabade65f9f9f8c22c6b6bc711c9e0fd06a68744430ab8c2b78d9d82"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:04:37.714469Z","signature_b64":"zAkPb2JMct+9nIz42GguLKJ/E7EB047B1byX0lW8b6BU1/IOHQuLerB0j08f/5Rj+rhOcArQrjoP4jXkNuoVDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ffc893c0e9eeef07360edccb83d6646d16476ec7530a391490932d81169859b","last_reissued_at":"2026-05-20T00:04:37.713585Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:04:37.713585Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hardware-Software Co-Design of Scalable, Energy-Efficient Analog Recurrent Computations","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AR","authors_text":"Arthur Fyon, Damien Ernst, Guillaume Drion, Jean-Michel Redout\\'e, Julien Brandoit, Loris Mendolia","submitted_at":"2026-05-12T09:44:32Z","abstract_excerpt":"Always-on AI applications, from environmental sensors to biomedical implants, require ultra-low power consumption. Analog circuits offer a path to sub-microwatt inference, yet existing analog implementations are limited to feedforward architectures: extending them to recurrent dynamics has been considered impractical due to noise accumulation through temporal feedback. We demonstrate that this barrier can be overcome through hardware-software co-design. Specifically, we identify that Bistable Memory Recurrent Units (BMRUs), a class of Recurrent Neural Networks (RNNs) with discrete-valued outpu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.15216","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/2605.15216/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":"2605.15216","created_at":"2026-05-20T00:04:37.713734+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.15216v2","created_at":"2026-05-20T00:04:37.713734+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15216","created_at":"2026-05-20T00:04:37.713734+00:00"},{"alias_kind":"pith_short_12","alias_value":"N76ISPAOT3XP","created_at":"2026-05-20T00:04:37.713734+00:00"},{"alias_kind":"pith_short_16","alias_value":"N76ISPAOT3XPA43A","created_at":"2026-05-20T00:04:37.713734+00:00"},{"alias_kind":"pith_short_8","alias_value":"N76ISPAO","created_at":"2026-05-20T00:04:37.713734+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.07936","citing_title":"A Fully Tunable Ultra-Low Power Current-Mode Memory Cell in Standard CMOS Technology","ref_index":31,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/N76ISPAOT3XPA43A5XGLQPLGI3","json":"https://pith.science/pith/N76ISPAOT3XPA43A5XGLQPLGI3.json","graph_json":"https://pith.science/api/pith-number/N76ISPAOT3XPA43A5XGLQPLGI3/graph.json","events_json":"https://pith.science/api/pith-number/N76ISPAOT3XPA43A5XGLQPLGI3/events.json","paper":"https://pith.science/paper/N76ISPAO"},"agent_actions":{"view_html":"https://pith.science/pith/N76ISPAOT3XPA43A5XGLQPLGI3","download_json":"https://pith.science/pith/N76ISPAOT3XPA43A5XGLQPLGI3.json","view_paper":"https://pith.science/paper/N76ISPAO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.15216&json=true","fetch_graph":"https://pith.science/api/pith-number/N76ISPAOT3XPA43A5XGLQPLGI3/graph.json","fetch_events":"https://pith.science/api/pith-number/N76ISPAOT3XPA43A5XGLQPLGI3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N76ISPAOT3XPA43A5XGLQPLGI3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N76ISPAOT3XPA43A5XGLQPLGI3/action/storage_attestation","attest_author":"https://pith.science/pith/N76ISPAOT3XPA43A5XGLQPLGI3/action/author_attestation","sign_citation":"https://pith.science/pith/N76ISPAOT3XPA43A5XGLQPLGI3/action/citation_signature","submit_replication":"https://pith.science/pith/N76ISPAOT3XPA43A5XGLQPLGI3/action/replication_record"}},"created_at":"2026-05-20T00:04:37.713734+00:00","updated_at":"2026-05-20T00:04:37.713734+00:00"}