{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:YXE2NXWCWDLAWL3MAQDYX56EOR","short_pith_number":"pith:YXE2NXWC","schema_version":"1.0","canonical_sha256":"c5c9a6dec2b0d60b2f6c04078bf7c4744da89556939880a0eee0739dc97966fe","source":{"kind":"arxiv","id":"2001.02610","version":1},"attestation_state":"computed","paper":{"title":"iDLG: Improved Deep Leakage from Gradients","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Bo Zhao, Hakan Bilen, Konda Reddy Mopuri","submitted_at":"2020-01-08T16:45:09Z","abstract_excerpt":"It is widely believed that sharing gradients will not leak private training data in distributed learning systems such as Collaborative Learning and Federated Learning, etc. Recently, Zhu et al. presented an approach which shows the possibility to obtain private training data from the publicly shared gradients. In their Deep Leakage from Gradient (DLG) method, they synthesize the dummy data and corresponding labels with the supervision of shared gradients. However, DLG has difficulty in convergence and discovering the ground-truth labels consistently. In this paper, we find that sharing gradien"},"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":"2001.02610","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-01-08T16:45:09Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"dcc2fd990008e5c9cb08ab7ca801b4cfe19035a371c128317826bd125c6d3dc4","abstract_canon_sha256":"49f41d086d870278c754f4d6884a618fe21b4f0dcc0c1c1de72de7d386553e91"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:32:19.439111Z","signature_b64":"D1OMdndpbCxHnFt3CvrwI6KWv6q2ff3ts2Yo+Hor9aleh+eH47iicfXoWMyqV2DzLP6mBOuQbDKexVFrs9s1Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c5c9a6dec2b0d60b2f6c04078bf7c4744da89556939880a0eee0739dc97966fe","last_reissued_at":"2026-07-05T00:32:19.438697Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:32:19.438697Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"iDLG: Improved Deep Leakage from Gradients","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Bo Zhao, Hakan Bilen, Konda Reddy Mopuri","submitted_at":"2020-01-08T16:45:09Z","abstract_excerpt":"It is widely believed that sharing gradients will not leak private training data in distributed learning systems such as Collaborative Learning and Federated Learning, etc. Recently, Zhu et al. presented an approach which shows the possibility to obtain private training data from the publicly shared gradients. In their Deep Leakage from Gradient (DLG) method, they synthesize the dummy data and corresponding labels with the supervision of shared gradients. However, DLG has difficulty in convergence and discovering the ground-truth labels consistently. In this paper, we find that sharing gradien"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2001.02610","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/2001.02610/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":"2001.02610","created_at":"2026-07-05T00:32:19.438754+00:00"},{"alias_kind":"arxiv_version","alias_value":"2001.02610v1","created_at":"2026-07-05T00:32:19.438754+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2001.02610","created_at":"2026-07-05T00:32:19.438754+00:00"},{"alias_kind":"pith_short_12","alias_value":"YXE2NXWCWDLA","created_at":"2026-07-05T00:32:19.438754+00:00"},{"alias_kind":"pith_short_16","alias_value":"YXE2NXWCWDLAWL3M","created_at":"2026-07-05T00:32:19.438754+00:00"},{"alias_kind":"pith_short_8","alias_value":"YXE2NXWC","created_at":"2026-07-05T00:32:19.438754+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":22,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.07314","citing_title":"FedCVESA: Taking Away Training Data in Federated Learning via Correlation Value Encoding and Segmented Aggregation","ref_index":43,"is_internal_anchor":true},{"citing_arxiv_id":"2606.25627","citing_title":"TL++: Accuracy and Privacy Preserving Traversal Learning for Distributed Intelligent Systems","ref_index":67,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22928","citing_title":"HADES: Privacy-Preserving Federated Learning via Selective Feature Encryption and Hybrid Model Fusion","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2606.23276","citing_title":"Exposing the Illusion of Erasure in Knowledge Editing for LLMs","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20553","citing_title":"From Efficiency to Leakage -- Privacy Backdoor in Federated Language Model Fine-Tuning","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18312","citing_title":"TIGER: Inverting Transformer Gradients via Embedding-Subspace Distance Optimization","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2606.17110","citing_title":"Loss Landscape Poisoning: Targeted Extraction of Unseen Training Data from LLMs","ref_index":91,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10780","citing_title":"Secure Aggregation with Top-K Sparsification in Decentralized Federated Learning","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04399","citing_title":"DPDL: Towards Differential Privacy Preservation in Decentralized Stochastic Learning on Non-IID Data","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00986","citing_title":"Profiling Privacy Preservation Against Gradient Inversion Attacks in Tabular Federated Learning","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.29845","citing_title":"Local Differential Privacy via Dynamic Quantization in Distributed Online Stochastic Optimization","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22898","citing_title":"FIRMA: FIbonacci Ring Model Aggregation for Privacy-preserving Federated Learning","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2308.06822","citing_title":"Approximate and Weighted Data Reconstruction Attack in Federated Learning","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2406.10861","citing_title":"Knowledge Distillation in Federated Learning: a Survey on Long Lasting Challenges and New Solutions","ref_index":187,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15573","citing_title":"Response-Conditioned Parallel-to-Sequential Orchestration for Multi-Agent Systems","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13265","citing_title":"LightSplit: Practical Privacy-Preserving Split Learning via Orthogonal Projections","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12168","citing_title":"On What We Can Learn from Low-Resolution Data","ref_index":110,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08288","citing_title":"UMEDA: Unified Multi-modal Efficient Data Fusion for Privacy-Preserving Graph Federated Learning via Spectral-Gated Attention and Diffusion-Based Operator Alignment","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07125","citing_title":"Scalable and Private Federated Learning Using Distributed Differential Privacy and Secure Aggregation","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2604.06297","citing_title":"FedSpy-LLM: Towards Scalable and Generalizable Data Reconstruction Attacks from Gradients on LLMs","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14833","citing_title":"Federated User Behavior Modeling for Privacy-Preserving LLM Recommendation","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16606","citing_title":"SafeLM: Unified Privacy-Aware Optimization for Trustworthy Federated Large Language Models","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YXE2NXWCWDLAWL3MAQDYX56EOR","json":"https://pith.science/pith/YXE2NXWCWDLAWL3MAQDYX56EOR.json","graph_json":"https://pith.science/api/pith-number/YXE2NXWCWDLAWL3MAQDYX56EOR/graph.json","events_json":"https://pith.science/api/pith-number/YXE2NXWCWDLAWL3MAQDYX56EOR/events.json","paper":"https://pith.science/paper/YXE2NXWC"},"agent_actions":{"view_html":"https://pith.science/pith/YXE2NXWCWDLAWL3MAQDYX56EOR","download_json":"https://pith.science/pith/YXE2NXWCWDLAWL3MAQDYX56EOR.json","view_paper":"https://pith.science/paper/YXE2NXWC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2001.02610&json=true","fetch_graph":"https://pith.science/api/pith-number/YXE2NXWCWDLAWL3MAQDYX56EOR/graph.json","fetch_events":"https://pith.science/api/pith-number/YXE2NXWCWDLAWL3MAQDYX56EOR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YXE2NXWCWDLAWL3MAQDYX56EOR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YXE2NXWCWDLAWL3MAQDYX56EOR/action/storage_attestation","attest_author":"https://pith.science/pith/YXE2NXWCWDLAWL3MAQDYX56EOR/action/author_attestation","sign_citation":"https://pith.science/pith/YXE2NXWCWDLAWL3MAQDYX56EOR/action/citation_signature","submit_replication":"https://pith.science/pith/YXE2NXWCWDLAWL3MAQDYX56EOR/action/replication_record"}},"created_at":"2026-07-05T00:32:19.438754+00:00","updated_at":"2026-07-05T00:32:19.438754+00:00"}