{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:TGP673A7GZ5SHYCF7C6PUSPMVJ","short_pith_number":"pith:TGP673A7","schema_version":"1.0","canonical_sha256":"999fefec1f367b23e045f8bcfa49ecaa74463100861d899aa0e8bdf06ab3c30a","source":{"kind":"arxiv","id":"2303.06652","version":1},"attestation_state":"computed","paper":{"title":"Interpreting Hidden Semantics in the Intermediate Layers of 3D Point Cloud Classification Neural Network","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cheng Wang, Minghao Liu, Shijun Zheng, Weiquan Liu","submitted_at":"2023-03-12T12:54:29Z","abstract_excerpt":"Although 3D point cloud classification neural network models have been widely used, the in-depth interpretation of the activation of the neurons and layers is still a challenge. We propose a novel approach, named Relevance Flow, to interpret the hidden semantics of 3D point cloud classification neural networks. It delivers the class Relevance to the activated neurons in the intermediate layers in a back-propagation manner, and associates the activation of neurons with the input points to visualize the hidden semantics of each layer. Specially, we reveal that the 3D point cloud classification n"},"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":"2303.06652","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2023-03-12T12:54:29Z","cross_cats_sorted":[],"title_canon_sha256":"f80cb3abb72a1eba1fd35c205bbe497b814f930d19da2bbf5c2658426c8b5e7e","abstract_canon_sha256":"907c15b8d31b84a4993556c8ec3bf512b356176d9276841187eb3b15c0c59edb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:50:21.514937Z","signature_b64":"YvcY48GiOsMwXj+GDaxlmw/NSfb8G3UwzzDvgh3Qj+AWLANV2zUY5p5E3HxS7vSreya+FFGjs/I9ogA9XlpkAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"999fefec1f367b23e045f8bcfa49ecaa74463100861d899aa0e8bdf06ab3c30a","last_reissued_at":"2026-07-05T05:50:21.514594Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:50:21.514594Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Interpreting Hidden Semantics in the Intermediate Layers of 3D Point Cloud Classification Neural Network","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cheng Wang, Minghao Liu, Shijun Zheng, Weiquan Liu","submitted_at":"2023-03-12T12:54:29Z","abstract_excerpt":"Although 3D point cloud classification neural network models have been widely used, the in-depth interpretation of the activation of the neurons and layers is still a challenge. We propose a novel approach, named Relevance Flow, to interpret the hidden semantics of 3D point cloud classification neural networks. It delivers the class Relevance to the activated neurons in the intermediate layers in a back-propagation manner, and associates the activation of neurons with the input points to visualize the hidden semantics of each layer. Specially, we reveal that the 3D point cloud classification n"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.06652","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/2303.06652/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":"2303.06652","created_at":"2026-07-05T05:50:21.514650+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.06652v1","created_at":"2026-07-05T05:50:21.514650+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.06652","created_at":"2026-07-05T05:50:21.514650+00:00"},{"alias_kind":"pith_short_12","alias_value":"TGP673A7GZ5S","created_at":"2026-07-05T05:50:21.514650+00:00"},{"alias_kind":"pith_short_16","alias_value":"TGP673A7GZ5SHYCF","created_at":"2026-07-05T05:50:21.514650+00:00"},{"alias_kind":"pith_short_8","alias_value":"TGP673A7","created_at":"2026-07-05T05:50:21.514650+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2504.15796","citing_title":"Locating and Mitigating Gradient Conflicts in Point Cloud Domain Adaptation via Saliency Map Skewness","ref_index":36,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TGP673A7GZ5SHYCF7C6PUSPMVJ","json":"https://pith.science/pith/TGP673A7GZ5SHYCF7C6PUSPMVJ.json","graph_json":"https://pith.science/api/pith-number/TGP673A7GZ5SHYCF7C6PUSPMVJ/graph.json","events_json":"https://pith.science/api/pith-number/TGP673A7GZ5SHYCF7C6PUSPMVJ/events.json","paper":"https://pith.science/paper/TGP673A7"},"agent_actions":{"view_html":"https://pith.science/pith/TGP673A7GZ5SHYCF7C6PUSPMVJ","download_json":"https://pith.science/pith/TGP673A7GZ5SHYCF7C6PUSPMVJ.json","view_paper":"https://pith.science/paper/TGP673A7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.06652&json=true","fetch_graph":"https://pith.science/api/pith-number/TGP673A7GZ5SHYCF7C6PUSPMVJ/graph.json","fetch_events":"https://pith.science/api/pith-number/TGP673A7GZ5SHYCF7C6PUSPMVJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TGP673A7GZ5SHYCF7C6PUSPMVJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TGP673A7GZ5SHYCF7C6PUSPMVJ/action/storage_attestation","attest_author":"https://pith.science/pith/TGP673A7GZ5SHYCF7C6PUSPMVJ/action/author_attestation","sign_citation":"https://pith.science/pith/TGP673A7GZ5SHYCF7C6PUSPMVJ/action/citation_signature","submit_replication":"https://pith.science/pith/TGP673A7GZ5SHYCF7C6PUSPMVJ/action/replication_record"}},"created_at":"2026-07-05T05:50:21.514650+00:00","updated_at":"2026-07-05T05:50:21.514650+00:00"}