{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:JPSMKDCKL4R2ODSQMYZPR2DOLQ","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"806867118c0399c7bb6e2c7cce660173b78777d4da839bd3617d37e03fa008b9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-11-17T08:05:31Z","title_canon_sha256":"2d78b257e76acbfa33b1786eed5643fdad5f34dd38b4ae3aa304544e03a5c6c8"},"schema_version":"1.0","source":{"id":"2511.13108","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2511.13108","created_at":"2026-05-20T00:01:37Z"},{"alias_kind":"arxiv_version","alias_value":"2511.13108v4","created_at":"2026-05-20T00:01:37Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2511.13108","created_at":"2026-05-20T00:01:37Z"},{"alias_kind":"pith_short_12","alias_value":"JPSMKDCKL4R2","created_at":"2026-05-20T00:01:37Z"},{"alias_kind":"pith_short_16","alias_value":"JPSMKDCKL4R2ODSQ","created_at":"2026-05-20T00:01:37Z"},{"alias_kind":"pith_short_8","alias_value":"JPSMKDCK","created_at":"2026-05-20T00:01:37Z"}],"graph_snapshots":[{"event_id":"sha256:9703df49546256daea08a664e01a238f13dfd749fafab378e89f52870e7a3712","target":"graph","created_at":"2026-05-20T00:01:37Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"By projecting task gradients onto the orthogonal complement of harmful directions and aligning with beneficial ones distilled from a frozen CLIP encoder, DGS-Net achieves unified optimization of prior preservation and irrelevant suppression, outperforming state-of-the-art approaches by an average margin of 6.6% across 50 generative models."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the proposed gradient-space decomposition can reliably and consistently separate harmful from beneficial descent directions without introducing instability or requiring extensive hyperparameter tuning specific to each dataset."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"DGS-Net applies distillation-guided gradient surgery during CLIP fine-tuning to preserve pre-trained knowledge and suppress irrelevant features, reporting 6.6% average gains over prior methods on 50 generative models."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"By projecting task gradients onto the orthogonal complement of harmful directions and aligning with beneficial ones distilled from a frozen CLIP encoder, DGS-Net fine-tunes for AI-generated image detection without catastrophic forgetting."}],"snapshot_sha256":"3f9963f3ae5b797f1a281820499238d5cbbdb886f0f27461d9399c1327e5f523"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"01533e939b13e000cb3846fbaceea91b21b43ccd8f784f1e9a16b82fc82f1d39"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2511.13108/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The rapid progress of generative models such as GANs and diffusion models has led to the widespread proliferation of AI-generated images, raising concerns about misinformation, privacy violations, and trust erosion in digital media. Although large-scale multimodal models like CLIP offer strong transferable representations for detecting synthetic content, fine-tuning them often induces catastrophic forgetting, which degrades pre-trained priors and limits cross-domain generalization. To address this issue, we propose the Distillation-guided Gradient Surgery Network (DGS-Net), a novel framework t","authors_text":"Boyu Wang, Fan Wang, Jiazhen Yan, Zhangjie Fu, Ziqiang Li, Ziwen He","cross_cats":[],"headline":"By projecting task gradients onto the orthogonal complement of harmful directions and aligning with beneficial ones distilled from a frozen CLIP encoder, DGS-Net fine-tunes for AI-generated image detection without catastrophic forgetting.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-11-17T08:05:31Z","title":"DGS-Net: Distillation-Guided Gradient Surgery for CLIP Fine-Tuning in AI-Generated Image Detection"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2511.13108","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-17T22:07:21.866370Z","id":"b3fb5e10-f9a6-4f4d-b731-2806db10627e","model_set":{"reader":"grok-4.3"},"one_line_summary":"DGS-Net applies distillation-guided gradient surgery during CLIP fine-tuning to preserve pre-trained knowledge and suppress irrelevant features, reporting 6.6% average gains over prior methods on 50 generative models.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"By projecting task gradients onto the orthogonal complement of harmful directions and aligning with beneficial ones distilled from a frozen CLIP encoder, DGS-Net fine-tunes for AI-generated image detection without catastrophic forgetting.","strongest_claim":"By projecting task gradients onto the orthogonal complement of harmful directions and aligning with beneficial ones distilled from a frozen CLIP encoder, DGS-Net achieves unified optimization of prior preservation and irrelevant suppression, outperforming state-of-the-art approaches by an average margin of 6.6% across 50 generative models.","weakest_assumption":"That the proposed gradient-space decomposition can reliably and consistently separate harmful from beneficial descent directions without introducing instability or requiring extensive hyperparameter tuning specific to each dataset."}},"verdict_id":"b3fb5e10-f9a6-4f4d-b731-2806db10627e"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:495d0426018df62a58e558830764a886a6b61fbac98351e2115c7a6470ead98c","target":"record","created_at":"2026-05-20T00:01:37Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"806867118c0399c7bb6e2c7cce660173b78777d4da839bd3617d37e03fa008b9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-11-17T08:05:31Z","title_canon_sha256":"2d78b257e76acbfa33b1786eed5643fdad5f34dd38b4ae3aa304544e03a5c6c8"},"schema_version":"1.0","source":{"id":"2511.13108","kind":"arxiv","version":4}},"canonical_sha256":"4be4c50c4a5f23a70e506632f8e86e5c37e99067b3615799171250c76b69623a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4be4c50c4a5f23a70e506632f8e86e5c37e99067b3615799171250c76b69623a","first_computed_at":"2026-05-20T00:01:37.038776Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:01:37.038776Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"MzRPGrMibWgZnDAQzHm0lRMG/ZY4GmTmTn12w8x+80469D/QOD/QCh8sTmVy2xqu1zJmzL7gCrfDm9yl5ebyBw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:01:37.039537Z","signed_message":"canonical_sha256_bytes"},"source_id":"2511.13108","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:495d0426018df62a58e558830764a886a6b61fbac98351e2115c7a6470ead98c","sha256:9703df49546256daea08a664e01a238f13dfd749fafab378e89f52870e7a3712"],"state_sha256":"08758eba97f1c273a84e596c55d4203cd2ca9262957d2c3dd2baa53c7ed9cf06"}