{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:XYC6SQIJB5K7GECSQTYRJRWXZY","short_pith_number":"pith:XYC6SQIJ","schema_version":"1.0","canonical_sha256":"be05e941090f55f3105284f114c6d7ce11f972419c2bfb4f70b07825d5976111","source":{"kind":"arxiv","id":"2006.08679","version":5},"attestation_state":"computed","paper":{"title":"Feature Space Saturation during Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.NE","stat.ML"],"primary_cat":"cs.LG","authors_text":"Anders Arpteg, Justin Shenk, Mats L. Richter, Mikael Huss, Wolf Byttner","submitted_at":"2020-06-15T18:28:21Z","abstract_excerpt":"We propose layer saturation - a simple, online-computable method for analyzing the information processing in neural networks. First, we show that a layer's output can be restricted to the eigenspace of its variance matrix without performance loss. We propose a computationally lightweight method for approximating the variance matrix during training. From the dimension of its lossless eigenspace we derive layer saturation - the ratio between the eigenspace dimension and layer width. We show that saturation seems to indicate which layers contribute to network performance. We demonstrate how to al"},"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":"2006.08679","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-06-15T18:28:21Z","cross_cats_sorted":["cs.CV","cs.NE","stat.ML"],"title_canon_sha256":"c4449d8502211f001638533cd85b7316a63c978d3c102c7b78edebe0d7115d35","abstract_canon_sha256":"1598cd6c626266a306cbb50ac6c0f592a8e384065c737d153c94741ba8544b50"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:33:37.782169Z","signature_b64":"vemq0EEER64a1BytnEGfkJt5Gg0jMR6xdtrH17KL8MMAAn6ZE6td65/csix9ZSFNalNqN7q3ed+IvCMt6EAbDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"be05e941090f55f3105284f114c6d7ce11f972419c2bfb4f70b07825d5976111","last_reissued_at":"2026-07-05T03:33:37.781677Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:33:37.781677Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Feature Space Saturation during Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.NE","stat.ML"],"primary_cat":"cs.LG","authors_text":"Anders Arpteg, Justin Shenk, Mats L. Richter, Mikael Huss, Wolf Byttner","submitted_at":"2020-06-15T18:28:21Z","abstract_excerpt":"We propose layer saturation - a simple, online-computable method for analyzing the information processing in neural networks. First, we show that a layer's output can be restricted to the eigenspace of its variance matrix without performance loss. We propose a computationally lightweight method for approximating the variance matrix during training. From the dimension of its lossless eigenspace we derive layer saturation - the ratio between the eigenspace dimension and layer width. We show that saturation seems to indicate which layers contribute to network performance. We demonstrate how to al"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2006.08679","kind":"arxiv","version":5},"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/2006.08679/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":"2006.08679","created_at":"2026-07-05T03:33:37.781738+00:00"},{"alias_kind":"arxiv_version","alias_value":"2006.08679v5","created_at":"2026-07-05T03:33:37.781738+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2006.08679","created_at":"2026-07-05T03:33:37.781738+00:00"},{"alias_kind":"pith_short_12","alias_value":"XYC6SQIJB5K7","created_at":"2026-07-05T03:33:37.781738+00:00"},{"alias_kind":"pith_short_16","alias_value":"XYC6SQIJB5K7GECS","created_at":"2026-07-05T03:33:37.781738+00:00"},{"alias_kind":"pith_short_8","alias_value":"XYC6SQIJ","created_at":"2026-07-05T03:33:37.781738+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/XYC6SQIJB5K7GECSQTYRJRWXZY","json":"https://pith.science/pith/XYC6SQIJB5K7GECSQTYRJRWXZY.json","graph_json":"https://pith.science/api/pith-number/XYC6SQIJB5K7GECSQTYRJRWXZY/graph.json","events_json":"https://pith.science/api/pith-number/XYC6SQIJB5K7GECSQTYRJRWXZY/events.json","paper":"https://pith.science/paper/XYC6SQIJ"},"agent_actions":{"view_html":"https://pith.science/pith/XYC6SQIJB5K7GECSQTYRJRWXZY","download_json":"https://pith.science/pith/XYC6SQIJB5K7GECSQTYRJRWXZY.json","view_paper":"https://pith.science/paper/XYC6SQIJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2006.08679&json=true","fetch_graph":"https://pith.science/api/pith-number/XYC6SQIJB5K7GECSQTYRJRWXZY/graph.json","fetch_events":"https://pith.science/api/pith-number/XYC6SQIJB5K7GECSQTYRJRWXZY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XYC6SQIJB5K7GECSQTYRJRWXZY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XYC6SQIJB5K7GECSQTYRJRWXZY/action/storage_attestation","attest_author":"https://pith.science/pith/XYC6SQIJB5K7GECSQTYRJRWXZY/action/author_attestation","sign_citation":"https://pith.science/pith/XYC6SQIJB5K7GECSQTYRJRWXZY/action/citation_signature","submit_replication":"https://pith.science/pith/XYC6SQIJB5K7GECSQTYRJRWXZY/action/replication_record"}},"created_at":"2026-07-05T03:33:37.781738+00:00","updated_at":"2026-07-05T03:33:37.781738+00:00"}