{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NBVLWZUQICLU357RZWQ4NDO7DC","short_pith_number":"pith:NBVLWZUQ","schema_version":"1.0","canonical_sha256":"686abb669040974df7f1cda1c68ddf18a9f9a11be8b6ff54996249ce67d5d9db","source":{"kind":"arxiv","id":"2605.31371","version":1},"attestation_state":"computed","paper":{"title":"Softsign: Smooth Sign in Your Optimizer For Better Parameter Heterogeneity Handling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Aleksandr Beznosikov, Amir Zainullin, Andrey Veprikov, Dmitrii Feoktistov, Timofey Belinsky","submitted_at":"2026-05-29T14:41:36Z","abstract_excerpt":"Sign-based and LMO-inspired optimizers have recently attracted substantial attention in deep learning due to their strong performance and low memory footprint. However, their fixed-magnitude updates can hurt terminal convergence: they decouple update mechanisms from gradient magnitudes and fail to account for parameter heterogeneity, often leading to oscillation rather than convergence. We propose SoftSignum, a smooth relaxation of sign-based optimization that replaces the hard sign map with a temperature-controlled soft-sign transformation, enabling a parameter-wise transition from sign-like "},"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":"2605.31371","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-29T14:41:36Z","cross_cats_sorted":[],"title_canon_sha256":"4815b067262f19534f93939bc8827840f69956320900bb63bd80a1ae507e20c1","abstract_canon_sha256":"1e601a57dbb5d1898d8fa6e0bab72760641951689af40477c15de75d66bfe897"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T02:04:01.482465Z","signature_b64":"QRHVTRO7jvtDpDgjr1ZIJip2c7XdwU/fcQfQrnvHbR/vWxE1Tzwp9Y6YXak/D3TNr+WoPxCS35XcWHznCh9DCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"686abb669040974df7f1cda1c68ddf18a9f9a11be8b6ff54996249ce67d5d9db","last_reissued_at":"2026-06-01T02:04:01.481451Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T02:04:01.481451Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Softsign: Smooth Sign in Your Optimizer For Better Parameter Heterogeneity Handling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Aleksandr Beznosikov, Amir Zainullin, Andrey Veprikov, Dmitrii Feoktistov, Timofey Belinsky","submitted_at":"2026-05-29T14:41:36Z","abstract_excerpt":"Sign-based and LMO-inspired optimizers have recently attracted substantial attention in deep learning due to their strong performance and low memory footprint. However, their fixed-magnitude updates can hurt terminal convergence: they decouple update mechanisms from gradient magnitudes and fail to account for parameter heterogeneity, often leading to oscillation rather than convergence. We propose SoftSignum, a smooth relaxation of sign-based optimization that replaces the hard sign map with a temperature-controlled soft-sign transformation, enabling a parameter-wise transition from sign-like "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.31371","kind":"arxiv","version":1},"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/2605.31371/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":"2605.31371","created_at":"2026-06-01T02:04:01.481599+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.31371v1","created_at":"2026-06-01T02:04:01.481599+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.31371","created_at":"2026-06-01T02:04:01.481599+00:00"},{"alias_kind":"pith_short_12","alias_value":"NBVLWZUQICLU","created_at":"2026-06-01T02:04:01.481599+00:00"},{"alias_kind":"pith_short_16","alias_value":"NBVLWZUQICLU357R","created_at":"2026-06-01T02:04:01.481599+00:00"},{"alias_kind":"pith_short_8","alias_value":"NBVLWZUQ","created_at":"2026-06-01T02:04:01.481599+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NBVLWZUQICLU357RZWQ4NDO7DC","json":"https://pith.science/pith/NBVLWZUQICLU357RZWQ4NDO7DC.json","graph_json":"https://pith.science/api/pith-number/NBVLWZUQICLU357RZWQ4NDO7DC/graph.json","events_json":"https://pith.science/api/pith-number/NBVLWZUQICLU357RZWQ4NDO7DC/events.json","paper":"https://pith.science/paper/NBVLWZUQ"},"agent_actions":{"view_html":"https://pith.science/pith/NBVLWZUQICLU357RZWQ4NDO7DC","download_json":"https://pith.science/pith/NBVLWZUQICLU357RZWQ4NDO7DC.json","view_paper":"https://pith.science/paper/NBVLWZUQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.31371&json=true","fetch_graph":"https://pith.science/api/pith-number/NBVLWZUQICLU357RZWQ4NDO7DC/graph.json","fetch_events":"https://pith.science/api/pith-number/NBVLWZUQICLU357RZWQ4NDO7DC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NBVLWZUQICLU357RZWQ4NDO7DC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NBVLWZUQICLU357RZWQ4NDO7DC/action/storage_attestation","attest_author":"https://pith.science/pith/NBVLWZUQICLU357RZWQ4NDO7DC/action/author_attestation","sign_citation":"https://pith.science/pith/NBVLWZUQICLU357RZWQ4NDO7DC/action/citation_signature","submit_replication":"https://pith.science/pith/NBVLWZUQICLU357RZWQ4NDO7DC/action/replication_record"}},"created_at":"2026-06-01T02:04:01.481599+00:00","updated_at":"2026-06-01T02:04:01.481599+00:00"}