{"total":12,"items":[{"citing_arxiv_id":"2605.23879","ref_index":17,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"On the Stability of Spherical Hellinger-Kantorovich Flows and Their Implications for Differential Privacy","primary_cat":"stat.ML","submitted_at":"2026-05-22T17:38:20+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"Establishes stability bounds for SHK flows yielding dimension-free controls on log-likelihood ratios and divergences, then applies them to time-dependent Pure-DP and Approximate-DP certificates for exponential-mechanism samplers.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.23131","ref_index":31,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"When Determinants Are Not Enough: Private Rare Switching","primary_cat":"cs.LG","submitted_at":"2026-05-22T01:09:14+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Replaces determinant growth with generalized Rayleigh quotient for rare switching in private linear bandits to control worst-direction volume despite non-monotonic design matrices from noise.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.15622","ref_index":19,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"Position: Zeroth-Order Optimization in Deep Learning Is Underexplored, Not Underpowered","primary_cat":"cs.LG","submitted_at":"2026-05-15T05:11:43+00:00","verdict":"UNVERDICTED","verdict_confidence":"UNKNOWN","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Zeroth-order optimization is underexplored rather than underpowered in deep learning, with limitations stemming from full-space designs that can be addressed via subspace, spectral, and systems-aware approaches.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.10774","ref_index":14,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"When Are Trade-Off Functions Testable from Finite Samples?","primary_cat":"math.ST","submitted_at":"2026-05-11T16:07:48+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Trade-off functions between two distributions are finitely testable if and only if their Neyman-Pearson rejection regions are attainable by a VC-class of sets.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"Our eO(n−2/5)rate is slower, as expected, because it comes from a generic trade-off testing procedure and applies to arbitrary benchmark curves, not only to total-variation closeness. 14 7 Efficient computation The statistical results above define the test as a scan over a class of witness setsS: ψn,δ(S) =1 n ∃S∈ S:h + Qn(S∁) \u0001 < f0 h+(Pn(S)) \u0001o .(14) At first sight, this may appear computationally difficult, becauseSmay be infinite and the rejection condition is nonlinear in the empirical probabilities. In this section, we show that the scan has a simple algorithmic reduction: it can be implemented using a finite number of cost-sensitive empirical risk minimization problems overS. This reduction separates the statistical and computational aspects of the method."},{"citing_arxiv_id":"2605.08651","ref_index":61,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"Privacy-Aware Video Anomaly Detection through Orthogonal Subspace Projection","primary_cat":"cs.CV","submitted_at":"2026-05-09T03:46:39+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"A new orthogonal projection module for video anomaly detection suppresses facial attributes via weak face-presence signals and cosine alignment while preserving anomaly-relevant features like pose and motion.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.06433","ref_index":39,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"Federated Cross-Client Subgraph Pattern Detection","primary_cat":"cs.LG","submitted_at":"2026-05-07T15:35:37+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"A per-step layer-wise embedding exchange in federated GNNs recovers centralized node representations for cross-client subgraph patterns under an extended-subgraph assumption.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.04901","ref_index":122,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"On the (In-)Security of the Shuffling Defense in the Transformer Secure Inference","primary_cat":"cs.CR","submitted_at":"2026-05-06T13:31:15+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"An attack aligns differently shuffled intermediate activations from secure Transformer inference queries to recover model weights with low error using roughly one dollar of queries.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.01425","ref_index":8,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"Barriers to Counterfactual Credit Attribution for Autoregressive Models","primary_cat":"cs.LG","submitted_at":"2026-05-02T12:53:18+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"CCA does not compose autoregressively and retrofitting requires exponential query complexity under weak optimality.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.20985","ref_index":14,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"Differentially Private Model Merging","primary_cat":"cs.LG","submitted_at":"2026-04-22T18:13:37+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Post-processing via random selection or linear combination of differentially private models allows meeting arbitrary target privacy parameters without additional training.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.19015","ref_index":119,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"FedProxy: Federated Fine-Tuning of LLMs via Proxy SLMs and Heterogeneity-Aware Fusion","primary_cat":"cs.LG","submitted_at":"2026-04-21T03:06:24+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"FedProxy replaces weak adapters with a proxy SLM for federated LLM fine-tuning, outperforming prior methods and approaching centralized performance via compression, heterogeneity-aware aggregation, and training-free fusion.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2605.18762","ref_index":28,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"ALDEN: Boosting Private Data Extraction from Retrieval-Augmented Generation Systems via Active Learning and Distribution Estimation","primary_cat":"cs.IR","submitted_at":"2026-04-10T08:26:40+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"ALDEN boosts private data extraction rates from RAG systems by combining active learning for query diversification with dynamic estimation of the underlying knowledge-base topic distribution.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2304.01373","ref_index":221,"ref_count":1,"confidence":0.55,"is_internal_anchor":false,"paper_title":"Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling","primary_cat":"cs.CL","submitted_at":"2023-04-03T20:58:15+00:00","verdict":"ACCEPT","verdict_confidence":"LOW","novelty_score":8.0,"formal_verification":"none","one_line_summary":"Pythia releases 16 identically trained LLMs with full checkpoints and data tools to study training dynamics, scaling, memorization, and bias in language models.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null}],"limit":50,"offset":0}