{"paper":{"title":"Scalable neuromorphic computing from autonomous spiking dynamics in a clockless reconfigurable chip","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Clockless asynchronous circuits on standard FPGAs generate autonomous spiking dynamics that solve machine-learning tasks at competitive accuracy with low power.","cross_cats":["cs.LG"],"primary_cat":"cs.NE","authors_text":"Damien Rontani, Eric Oliveira Gomes","submitted_at":"2026-05-15T15:58:38Z","abstract_excerpt":"We propose a scalable neuromorphic architecture based on spiking dynamics emerging from the autonomous time-continuous evolution of clockless (asynchronous) digital circuits. Implemented on commercially available field-programmable gate arrays (FPGAs), our system implements networks of interacting Boolean spiking neurons with configurable excitatory and inhibitory synaptic weights. A complete processing pipeline enables efficient handling of spike-encoded data for solving machine-learning tasks. We demonstrate competitive performance for an audio classification task with spike-based encoding a"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We demonstrate competitive performance for an audio classification task with spike-based encoding and high-speed processing. Power consumption is significantly lower than traditional digital implementations; this makes our approach an efficient alternative that bridges the gap to dedicated analog neuromorphic systems without the need for specialized hardware design.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That autonomous spiking dynamics arising from the time-continuous evolution of clockless digital circuits on commercial FPGAs can be configured via synaptic weights to solve machine-learning tasks at competitive accuracy without hidden costs in scalability or stability.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Clockless FPGA circuits produce autonomous spiking neuron networks that achieve competitive audio classification accuracy with significantly lower power than conventional digital implementations.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Clockless asynchronous circuits on standard FPGAs generate autonomous spiking dynamics that solve machine-learning tasks at competitive accuracy with low power.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4b1820e8acb03d87b6ef47e33e1707fce39bdaece956cf0827c83be4dea6b8a2"},"source":{"id":"2605.16114","kind":"arxiv","version":1},"verdict":{"id":"3dcd6ee0-560e-43da-b22d-43fdfe43fdf6","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T17:51:36.469262Z","strongest_claim":"We demonstrate competitive performance for an audio classification task with spike-based encoding and high-speed processing. Power consumption is significantly lower than traditional digital implementations; this makes our approach an efficient alternative that bridges the gap to dedicated analog neuromorphic systems without the need for specialized hardware design.","one_line_summary":"Clockless FPGA circuits produce autonomous spiking neuron networks that achieve competitive audio classification accuracy with significantly lower power than conventional digital implementations.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That autonomous spiking dynamics arising from the time-continuous evolution of clockless digital circuits on commercial FPGAs can be configured via synaptic weights to solve machine-learning tasks at competitive accuracy without hidden costs in scalability or stability.","pith_extraction_headline":"Clockless asynchronous circuits on standard FPGAs generate autonomous spiking dynamics that solve machine-learning tasks at competitive accuracy with low power."},"integrity":{"clean":false,"summary":{"advisory":0,"critical":1,"by_detector":{"doi_compliance":{"total":1,"advisory":0,"critical":1,"informational":0}},"informational":0},"endpoint":"/pith/2605.16114/integrity.json","findings":[{"note":"Identifier '10.1016/s0361-9230(99' is syntactically valid but the DOI registry (doi.org) returned 404, and Crossref / OpenAlex / internal corpus also have no record. The cited work could not be located through any authoritative source.","detector":"doi_compliance","severity":"critical","ref_index":45,"audited_at":"2026-05-19T18:00:33.653433Z","detected_doi":"10.1016/s0361-9230(99","finding_type":"unresolvable_identifier","verdict_class":"cross_source","detected_arxiv_id":null}],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T18:01:18.534654Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T18:00:33.653433Z","status":"completed","version":"1.0.0","findings_count":1},{"name":"ai_meta_artifact","ran_at":"2026-05-19T17:33:33.855176Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T16:41:55.478506Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"c68be3b5df378492151e9c5c5a8e7630d29a423e6bd7936f793d3d9401991676"},"references":{"count":58,"sample":[{"doi":"10.3389/fnins.2022.873935","year":2022,"title":"Ostrau, C., Klarhorst, C., Thies, M. & Rückert, U. Benchmarking Neuromorphic Hardware and Its Energy Expenditure. Front. Neurosci.16, 873935, DOI: 10.3389/fnins.2022.873935 (2022)","work_id":"ec33aaa2-5429-4db2-ba50-9443afd73cc6","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1038/s41467-024-46397-3","year":2024,"title":"Pal, A.et al.An ultra energy-efficient hardware platform for neuromorphic computing enabled by 2D-TMD tunnel-FETs. Nat. Commun.15, 3392, DOI: 10.1038/s41467-024-46397-3 (2024)","work_id":"4e77fd13-ee1e-482b-93e4-2633d71116c1","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1126/science.1091277","year":2004,"title":"Jaeger, H. & Haas, H. Harnessing Nonlinearity: Predicting Chaotic Systems and Saving Energy in Wireless Communication. Science304, 78–80, DOI: 10.1126/science.1091277 (2004)","work_id":"2bd1679a-b5bf-4e31-b0be-dc2ade003925","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1162/089976602760407955","year":2002,"title":"Real-time computing without stable states: A new framework for neural computation based on perturbations.Neural Computation, 14(11): 2531–2560","work_id":"7acfab49-7aca-427f-835b-6767d9596840","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1016/j.cosrev.2009.03.005","year":2009,"title":"Reservoir computing approaches to recurrent neural network training","work_id":"7718cee7-cd49-4d02-a16c-8c5252702120","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":58,"snapshot_sha256":"e7a245c37dda24fbddbc9c65b617be53ec962eb0faba7e0ff34759aba3723ac3","internal_anchors":5},"formal_canon":{"evidence_count":3,"snapshot_sha256":"ee786577867f95f0e1c10a53953d2552c48697dd00b1dd1482fb36b20051a8f6"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}