{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:JPSMKDCKL4R2ODSQMYZPR2DOLQ","short_pith_number":"pith:JPSMKDCK","schema_version":"1.0","canonical_sha256":"4be4c50c4a5f23a70e506632f8e86e5c37e99067b3615799171250c76b69623a","source":{"kind":"arxiv","id":"2511.13108","version":4},"attestation_state":"computed","paper":{"title":"DGS-Net: Distillation-Guided Gradient Surgery for CLIP Fine-Tuning in AI-Generated Image Detection","license":"http://creativecommons.org/licenses/by/4.0/","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.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Boyu Wang, Fan Wang, Jiazhen Yan, Zhangjie Fu, Ziqiang Li, Ziwen He","submitted_at":"2025-11-17T08:05:31Z","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"},"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":true},"canonical_record":{"source":{"id":"2511.13108","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-11-17T08:05:31Z","cross_cats_sorted":[],"title_canon_sha256":"2d78b257e76acbfa33b1786eed5643fdad5f34dd38b4ae3aa304544e03a5c6c8","abstract_canon_sha256":"806867118c0399c7bb6e2c7cce660173b78777d4da839bd3617d37e03fa008b9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:01:37.039537Z","signature_b64":"MzRPGrMibWgZnDAQzHm0lRMG/ZY4GmTmTn12w8x+80469D/QOD/QCh8sTmVy2xqu1zJmzL7gCrfDm9yl5ebyBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4be4c50c4a5f23a70e506632f8e86e5c37e99067b3615799171250c76b69623a","last_reissued_at":"2026-05-20T00:01:37.038776Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:01:37.038776Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DGS-Net: Distillation-Guided Gradient Surgery for CLIP Fine-Tuning in AI-Generated Image Detection","license":"http://creativecommons.org/licenses/by/4.0/","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.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Boyu Wang, Fan Wang, Jiazhen Yan, Zhangjie Fu, Ziqiang Li, Ziwen He","submitted_at":"2025-11-17T08:05:31Z","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"},"claims":{"count":4,"items":[{"kind":"strongest_claim","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.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","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.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","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.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","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.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"3f9963f3ae5b797f1a281820499238d5cbbdb886f0f27461d9399c1327e5f523"},"source":{"id":"2511.13108","kind":"arxiv","version":4},"verdict":{"id":"b3fb5e10-f9a6-4f4d-b731-2806db10627e","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T22:07:21.866370Z","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.","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","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.","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."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2511.13108/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":2,"snapshot_sha256":"01533e939b13e000cb3846fbaceea91b21b43ccd8f784f1e9a16b82fc82f1d39"},"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":"2511.13108","created_at":"2026-05-20T00:01:37.038891+00:00"},{"alias_kind":"arxiv_version","alias_value":"2511.13108v4","created_at":"2026-05-20T00:01:37.038891+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2511.13108","created_at":"2026-05-20T00:01:37.038891+00:00"},{"alias_kind":"pith_short_12","alias_value":"JPSMKDCKL4R2","created_at":"2026-05-20T00:01:37.038891+00:00"},{"alias_kind":"pith_short_16","alias_value":"JPSMKDCKL4R2ODSQ","created_at":"2026-05-20T00:01:37.038891+00:00"},{"alias_kind":"pith_short_8","alias_value":"JPSMKDCK","created_at":"2026-05-20T00:01:37.038891+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JPSMKDCKL4R2ODSQMYZPR2DOLQ","json":"https://pith.science/pith/JPSMKDCKL4R2ODSQMYZPR2DOLQ.json","graph_json":"https://pith.science/api/pith-number/JPSMKDCKL4R2ODSQMYZPR2DOLQ/graph.json","events_json":"https://pith.science/api/pith-number/JPSMKDCKL4R2ODSQMYZPR2DOLQ/events.json","paper":"https://pith.science/paper/JPSMKDCK"},"agent_actions":{"view_html":"https://pith.science/pith/JPSMKDCKL4R2ODSQMYZPR2DOLQ","download_json":"https://pith.science/pith/JPSMKDCKL4R2ODSQMYZPR2DOLQ.json","view_paper":"https://pith.science/paper/JPSMKDCK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2511.13108&json=true","fetch_graph":"https://pith.science/api/pith-number/JPSMKDCKL4R2ODSQMYZPR2DOLQ/graph.json","fetch_events":"https://pith.science/api/pith-number/JPSMKDCKL4R2ODSQMYZPR2DOLQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JPSMKDCKL4R2ODSQMYZPR2DOLQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JPSMKDCKL4R2ODSQMYZPR2DOLQ/action/storage_attestation","attest_author":"https://pith.science/pith/JPSMKDCKL4R2ODSQMYZPR2DOLQ/action/author_attestation","sign_citation":"https://pith.science/pith/JPSMKDCKL4R2ODSQMYZPR2DOLQ/action/citation_signature","submit_replication":"https://pith.science/pith/JPSMKDCKL4R2ODSQMYZPR2DOLQ/action/replication_record"}},"created_at":"2026-05-20T00:01:37.038891+00:00","updated_at":"2026-05-20T00:01:37.038891+00:00"}