{"total":107,"items":[{"citing_arxiv_id":"2604.27733","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Mind the Gap: Structure-Aware Consistency in Preference Learning","primary_cat":"cs.LG","submitted_at":"2026-04-30T11:24:04+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Standard DPO surrogates are inconsistent for equicontinuous neural nets; SA-DPO provides structure-aware H-consistency bounds by adapting margins to semantic distance and shows heavy-tailed losses yield superior guarantees for capacity-bounded models via the Margin-Capacity Profile.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.27723","ref_index":143,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Optimized Deferral for Imbalanced Settings","primary_cat":"cs.LG","submitted_at":"2026-04-30T11:15:37+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"MILD reformulates two-stage learning to defer as cost-sensitive learning over the input-expert domain and derives new margin-based losses with guarantees, yielding better performance than baselines on image classification and LLM routing tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.27132","ref_index":46,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"TRUST: A Framework for Decentralized AI Service v.0.1","primary_cat":"cs.AI","submitted_at":"2026-04-29T19:32:58+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"TRUST is a decentralized AI auditing framework that decomposes reasoning into HDAGs, maps agent interactions via the DAAN protocol to CIGs, and uses stake-weighted multi-tier consensus to achieve 72.4% accuracy while proving a Safety-Profitability Theorem that rewards honest auditors.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.25975","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Rethinking KV Cache Eviction via a Unified Information-Theoretic Objective","primary_cat":"cs.LG","submitted_at":"2026-04-28T12:28:04+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"KV cache eviction is unified under an information capacity maximization principle derived from a linear-Gaussian attention surrogate, with CapKV proposed as a leverage-score based implementation that outperforms prior heuristics in experiments.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.23790","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"A General Representation-Based Approach to Multi-Source Domain Adaptation","primary_cat":"cs.LG","submitted_at":"2026-04-26T16:29:32+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"A representation learning approach for multi-source domain adaptation achieves identifiability by partitioning the label's Markov blanket into parents, children, and spouses.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.23065","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"What Should Frontier AI Developers Disclose About Internal Deployments?","primary_cat":"cs.CY","submitted_at":"2026-04-24T23:28:38+00:00","verdict":null,"verdict_confidence":null,"novelty_score":null,"formal_verification":null,"one_line_summary":null,"context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.22951","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"The Power of Power Law: Asymmetry Enables Compositional Reasoning","primary_cat":"cs.AI","submitted_at":"2026-04-24T18:49:08+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Power-law data sampling creates beneficial asymmetry in the loss landscape that lets models acquire high-frequency skill compositions first, enabling more efficient learning of rare long-tail skills than uniform distributions.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.22167","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Estimating Tail Risks in Language Model Output Distributions","primary_cat":"cs.LG","submitted_at":"2026-04-24T02:30:46+00:00","verdict":null,"verdict_confidence":null,"novelty_score":null,"formal_verification":null,"one_line_summary":null,"context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.19530","ref_index":28,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Calibrating Scientific Foundation Models with Inference-Time Stochastic Attention","primary_cat":"cs.LG","submitted_at":"2026-04-21T14:52:40+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Stochastic Attention adds calibrated uncertainty to transformer foundation models through inference-time multinomial sampling of attention weights and univariate post-hoc tuning of a concentration parameter.","context_count":1,"top_context_role":"background","top_context_polarity":"unclear","context_text":"where the squares are applied element-wise. • Low-rank term.Let θi = 1 i Pi j=1 θj be the running mean afteri snapshots, and define deviation columnsdi =θ i−θi. To limit the rank, SW AG retains only the lastK such columns in a matrix D∈R d×K, giving the low-rank covariance Σlr = 1 K−1 DD⊺.(27) The resulting SW AG posterior approximation is qSW AG(θ) =N \u0010 θSW A, 1 2(Σdiag +Σ lr) \u0011 .(28) Givenz 1 ∼ N(0, I d)andz 2 ∼ N(0, I K), SW AG draws samples via eθ=θ SW A+ 1√ 2 Σ1/2 diagz1 + 1p 2(K−1) Dz2.(29) WithMsamples eθ (m) ∼q SW AG(θ), the predictive distribution is approximated by Monte Carlo averaging: p(y⋆ |x ⋆,D)≈ 1 M MX m=1 p(y⋆ |x ⋆,eθ (m) ).(30) For architectures with batch normalization, batch-norm statistics are typically recomputed after sampling eacheθ"},{"citing_arxiv_id":"2604.17137","ref_index":27,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"BOIL: Learning Environment Personalized Information","primary_cat":"cs.LG","submitted_at":"2026-04-18T20:24:59+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":3.0,"formal_verification":"none","one_line_summary":"BOIL combines Pagerank and common information maximization to generate strategy distributions that outperform heuristics for long-term multi-agent tasks in complex environments.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.16888","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Towards Fully Parameter-Free Stochastic Optimization: Grid Search with Self-Bounding Analysis","primary_cat":"cs.LG","submitted_at":"2026-04-18T07:32:57+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Grasp is a grid search method with self-bounding analysis enabling fully parameter-free stochastic optimization with near-optimal rates in non-convex settings and competitive performance in convex cases.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.16220","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"OT on the Map: Quantifying Domain Shifts in Geographic Space","primary_cat":"cs.LG","submitted_at":"2026-04-17T16:33:38+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"GeoSpOT applies optimal transport to longitude-latitude data to quantify geospatial domain shifts and predict cross-region model transfer performance.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.16087","ref_index":44,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"The Harder Path: Last Iterate Convergence for Uncoupled Learning in Zero-Sum Games with Bandit Feedback","primary_cat":"cs.LG","submitted_at":"2026-04-17T14:17:09+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"In bandit-feedback zero-sum games, uncoupled algorithms achieve last-iterate Nash convergence at the optimal rate of O(T^{-1/4}).","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.16022","ref_index":53,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"SocialGrid: A Benchmark for Planning and Social Reasoning in Embodied Multi-Agent Systems","primary_cat":"cs.AI","submitted_at":"2026-04-17T12:51:46+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SocialGrid benchmark shows even top LLMs achieve below 60% in embodied planning and task completion, with deception detection near random chance regardless of model scale.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.15529","ref_index":57,"ref_count":3,"confidence":0.35,"is_internal_anchor":false,"paper_title":"LACE: Lattice Attention for Cross-thread Exploration","primary_cat":"cs.AI","submitted_at":"2026-04-16T21:19:35+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"LACE enables concurrent reasoning paths in LLMs to interact via lattice attention and a synthetic training pipeline, raising accuracy more than 7 points over independent parallel search.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.16538","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Understanding Tool-Augmented Agents for Lean Formalization: A Factorial Analysis","primary_cat":"cs.SE","submitted_at":"2026-04-16T21:15:46+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"Factorial experiments show that tool-augmented agents achieve substantially higher compilation success and semantic equivalence than one-shot baselines when translating natural language mathematics into Lean 4.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.14108","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Momentum Further Constrains Sharpness at the Edge of Stochastic Stability","primary_cat":"cs.LG","submitted_at":"2026-04-15T17:28:46+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"Momentum SGD exhibits two distinct EoSS regimes for batch sharpness, stabilizing at 2(1-β)/η for small batches and 2(1+β)/η for large batches, aligning with linear stability thresholds.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.13717","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"On Cost-Effective LLM-as-a-Judge Improvement Techniques","primary_cat":"cs.CL","submitted_at":"2026-04-15T10:52:33+00:00","verdict":null,"verdict_confidence":null,"novelty_score":null,"formal_verification":null,"one_line_summary":null,"context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.13392","ref_index":1,"ref_count":2,"confidence":0.35,"is_internal_anchor":false,"paper_title":"ReSS: Learning Reasoning Models for Tabular Data Prediction via Symbolic Scaffold","primary_cat":"cs.AI","submitted_at":"2026-04-15T01:43:00+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"ReSS extracts decision paths from trees as scaffolds to guide LLM reasoning generation, fine-tunes the LLM on the resulting dataset with scaffold-invariant augmentation, and reports up to 10% gains on medical and financial tabular benchmarks with new faithfulness metrics.","context_count":1,"top_context_role":"other","top_context_polarity":"unclear","context_text":"to symbolic constraints as specified in the symbolic scaf- fold. We give some examples of curated reasoning traces following this procedure in Appendix B.1. For each example (x,y) that is correctly predicted by the de- cision tree model, we let R(x,y, S(x)) denote the curated reasoning tokens. As a result, we collect a set of reason- ing data {xi,z i,y i}i∈C, where zi =R(x i,y i, S(xi)), and C ∈[1, . . . , n] denotes the subset of data that is correctly predicted by the decision tree model. 3.4. Scaffold-Invariant Data Augmentation One might note that we only use the data that is correctly predicted by the decision tree model. This is important for LLM to learn consistent decision rules. However, it will shrink the size of training data for SFT."},{"citing_arxiv_id":"2604.11808","ref_index":1,"ref_count":2,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Pair2Scene: Learning Local Object Relations for Procedural Scene Generation","primary_cat":"cs.CV","submitted_at":"2026-04-13T17:59:55+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Pair2Scene generates complex 3D scenes beyond training data by training a network on local object-pair placement rules and applying them recursively with collision-aware sampling.","context_count":1,"top_context_role":"background","top_context_polarity":"background","context_text":"tic proximity between entities sharing a common surface, where the dependent object's placement is conditioned by its functional utility relative to an anchor (e.g., a keyboard placed relative to a laptop). Based on these relations, a global scene is represented as an ordered sequence of relational tuplesS={T 1,T 2, . . . ,TN }. Each tupleT i is formulated as: Ti =⟨O dep,i,O sup,i,{O f nc,i}opt⟩,(1) where Odep,i is the object to be generated, Osup,i is the mandatory support anchor, and Of nc,i is an optional func- tional anchor. To ensure a valid generative dependency, the sequence must maintain causality: for any i-th tuple, the anchor objects (Osup,i and Of nc,i) must either be the floor or have been previously instantiated as a dependent object Odep,j in the"},{"citing_arxiv_id":"2604.10689","ref_index":42,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Communication-Efficient Gluon in Federated Learning","primary_cat":"cs.LG","submitted_at":"2026-04-12T15:30:28+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"Compressed Gluon variants using unbiased/contraction compressors and SARAH-style variance reduction achieve convergence guarantees and lower communication costs in federated learning under layer-wise smoothness.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.18607","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"TurboEvolve: Towards Fast and Robust LLM-Driven Program Evolution","primary_cat":"cs.NE","submitted_at":"2026-04-12T12:42:09+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"TurboEvolve improves LLM program evolution by running parallel islands with LLM-generated diverse candidates that carry self-assigned weights, an adaptive scheduler, and clustered seed injection to reach stronger solutions at lower evaluation budgets.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.10586","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Preventing Latent Rehearsal Decay in Online Continual SSL with SOLAR","primary_cat":"cs.LG","submitted_at":"2026-04-12T11:11:06+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SOLAR prevents latent rehearsal decay in online continual SSL by adaptively managing replay buffers with deviation proxies and an explicit overlap loss, delivering both fast convergence and state-of-the-art final accuracy on vision benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.10449","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"AdverMCTS: Combating Pseudo-Correctness in Code Generation via Adversarial Monte Carlo Tree Search","primary_cat":"cs.SE","submitted_at":"2026-04-12T04:15:31+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"AdverMCTS frames code generation as a minimax game where an attacker evolves tests to expose flaws in solver-generated code, yielding more robust outputs than static-test baselines.","context_count":1,"top_context_role":"other","top_context_polarity":"unclear","context_text":"Trace through the logic with the given test input 3. Determine the CORRECT output and which code(s) produced it Respond in the following format: <reasoning> Brief explanation (2-3 sentences max) of why this is the correct output. </reasoning> <correct output> The correct output value </correct output> <correct codes id> List of correct code indices, e.g., [1, 3] or [2] 16 ADVERMCTS </correct codes id> E. Algorithm We present the detailed procedure of ADVERMCTS in pseudocode in Algorithm 1. Algorithm 1TheADVERMCTSInference-Time Search Procedure. Require:P: Problem description;T pub: Public tests;N s iter: Iterations of Solver;N a iter: Iterations of Attacker. 1:Output:Robust solutionC ∗ 2: # Initialization"},{"citing_arxiv_id":"2604.10074","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Transformers Learn the Optimal DDPM Denoiser for Multi-Token GMMs","primary_cat":"cs.LG","submitted_at":"2026-04-11T07:46:15+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":8.0,"formal_verification":"none","one_line_summary":"Transformers converge globally to the optimal DDPM denoiser for multi-token GMMs via self-attention mean denoising, with explicit token and iteration requirements.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.09543","ref_index":1,"ref_count":3,"confidence":0.35,"is_internal_anchor":false,"paper_title":"ANTIC: Adaptive Neural Temporal In-situ Compressor","primary_cat":"cs.LG","submitted_at":"2026-04-10T17:58:16+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"ANTIC reduces storage for large-scale PDE simulations by orders of magnitude through adaptive temporal snapshot selection combined with continual neural-field residual compression while preserving physics accuracy.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.09041","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"U-Cast: A Surprisingly Simple and Efficient Frontier Probabilistic AI Weather Forecaster","primary_cat":"cs.LG","submitted_at":"2026-04-10T07:02:20+00:00","verdict":null,"verdict_confidence":null,"novelty_score":null,"formal_verification":null,"one_line_summary":null,"context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.08706","ref_index":3,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Efficient RL Training for LLMs with Experience Replay","primary_cat":"cs.LG","submitted_at":"2026-04-09T18:56:12+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Well-designed experience replay buffers reduce inference compute in LLM RL post-training while maintaining or improving performance and preserving policy entropy.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.08525","ref_index":113,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Ads in AI Chatbots? An Analysis of How Large Language Models Navigate Conflicts of Interest","primary_cat":"cs.AI","submitted_at":"2026-04-09T17:57:50+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Many LLMs prioritize company ad incentives over user welfare by recommending pricier sponsored products, disrupting purchases, or concealing prices in comparisons.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.07931","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Robust Length Prediction: A Perspective from Heavy-Tailed Prompt-Conditioned Distributions","primary_cat":"cs.LG","submitted_at":"2026-04-09T07:49:52+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"LLM output lengths conditioned on a prompt form heavy-tailed distributions, so robust estimation from multiple samples outperforms single-sample labels for prediction.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.07086","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Radio-Frequency Inverse Rendering for Wireless Environment Modeling","primary_cat":"eess.SP","submitted_at":"2026-04-08T13:40:29+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"RFIR framework embeds RF-aware BSDF into Gaussian splatting for decoupled RF scene modeling, generalizing RCS synthesis, RSSI prediction, and wireless scene editability with performance gains.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.06473","ref_index":60,"ref_count":2,"confidence":0.35,"is_internal_anchor":false,"paper_title":"MICA: Multivariate Infini Compressive Attention for Time Series Forecasting","primary_cat":"cs.LG","submitted_at":"2026-04-07T21:19:00+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"MICA adapts infini compressive attention to the channel dimension, enabling scalable cross-channel dependencies in Transformers and cutting forecast error by 5.4% on average versus channel-independent baselines.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.05727","ref_index":41,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Single-Stage Signal Attenuation Diffusion Model for Low-Light Image Enhancement and Denoising","primary_cat":"cs.CV","submitted_at":"2026-04-07T11:33:04+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"SADM adds a signal attenuation coefficient to the diffusion forward process so that reverse denoising simultaneously recovers brightness and suppresses noise without extra stages or correction modules.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.05544","ref_index":54,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Referring-Aware Visuomotor Policy Learning for Closed-Loop Manipulation","primary_cat":"cs.RO","submitted_at":"2026-04-07T07:41:11+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"ReV is a referring-aware visuomotor policy using coupled diffusion heads for real-time trajectory replanning in robotic manipulation, trained solely via targeted perturbations to expert demonstrations and achieving higher success rates in simulated and real tasks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.05134","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"How Reasoning Evolves from Post-Training Data: An Empirical Study Using Chess","primary_cat":"cs.LG","submitted_at":"2026-04-06T19:53:39+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Training language models on single best-move predictions in chess leads to strong but unfaithful reasoning after RL, while multi-move trajectories produce faithful reasoning with similar performance and stability.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.04743","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Hallucination Basins: A Dynamic Framework for Understanding and Controlling LLM Hallucinations","primary_cat":"cs.CL","submitted_at":"2026-04-06T15:08:54+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"LLM hallucinations arise from task-dependent basins in latent space, with separability varying by task and geometry-aware steering reducing their probability.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.03146","ref_index":44,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Characterization of Gaussian Universality Breakdown in High-Dimensional Empirical Risk Minimization","primary_cat":"stat.ML","submitted_at":"2026-04-03T16:07:02+00:00","verdict":null,"verdict_confidence":null,"novelty_score":null,"formal_verification":null,"one_line_summary":null,"context_count":1,"top_context_role":"background","top_context_polarity":"unclear","context_text":"concentration of˜αafter bounding the variations of theα-gradient ∂f ∂α = 1 n ∇ueL(αg+ (µ ⊤x)1;κ), h −αν, with respect to fluctuations of h (and x when applicable). Under the regularity properties of the Moreau envelope (bounded first and second derivatives under Assumption 2), this yields that ˜αis bounded and is a C/√n-Lipschitz function of h (details omitted), hence P(|˜α−E[˜α]| ≥t)≤Ce −cnt2 ,(44) for some constantsC, c >0. Denote ˜α0 := ˜α−E[˜α]. By Lemma C.3, E[˜α]can be characterized as the minimizer of the expected shifted objective (obtained by replacing α with µα + ˜α0 in f). Comparing this expected objective with the deterministic limit defining α∗ and using again Lemma B.1 (together with the Lipschitz properties of the α-gradient) yields |E[˜α]−α∗| ≤C/ √n."},{"citing_arxiv_id":"2604.02883","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Information-Regularized Constrained Inversion for Stable Avatar Editing from Sparse Supervision","primary_cat":"cs.CV","submitted_at":"2026-04-03T08:46:54+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"A conditioning-guided constrained inversion method restricts avatar edits to a low-dimensional part-specific subspace and uses an information matrix spectrum from pipeline linearization to predict and ensure stability under sparse supervision.","context_count":1,"top_context_role":"background","top_context_polarity":"unclear","context_text":"reweight these constraints during inversion using a spec- tral objective derived from a local linearization of the full 3 Information-Regularized Constrained Inversion for Stable Avatar Editing from Sparse Supervision decoding-and-rendering pipeline. 3.1 Differentiable Avatar Rendering Pipeline We assume a differentiable, animatable rendering pipeline yt =f(v, θ t)∈R m,(1) where v∈R r is a globalediting codeshared across frames, θt denotes the frame-specific driving state (pose parame- ters, cameras, and optionally illumination), and yt denotes rendered observations (RGB and optionally depth/normal). Residual UV parameterization.A base reconstruction produces a canonical UV feature map Ubase. We model editing as adding a decoded residual UV feature map:"},{"citing_arxiv_id":"2604.20865","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Advances in Art: Orthogonal Disruption and the Beauty in Schematics","primary_cat":"cs.CY","submitted_at":"2026-03-25T09:40:46+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"Orthogonal Art is defined as an artistic practice using schematics to occupy generative spaces inaccessible to AI, serving as a pedagogical bridge between art, engineering, and philosophy.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.13075","ref_index":27,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"DeEscalWild: A Real-World Benchmark for Automated De-Escalation Training with SLMs","primary_cat":"cs.CL","submitted_at":"2026-03-20T23:32:28+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"DeEscalWild supplies 1,500 high-fidelity de-escalation scenarios that let fine-tuned 3B SLMs outperform general-purpose larger models on realism and dialogue metrics.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.02349","ref_index":60,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"OPRIDE: Offline Preference-based Reinforcement Learning via In-Dataset Exploration","primary_cat":"cs.LG","submitted_at":"2026-02-19T02:11:01+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"OPRIDE improves query efficiency in offline PbRL via a principled in-dataset exploration strategy and discount scheduling, outperforming prior methods with fewer queries and providing theoretical guarantees.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.09557","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"SPEED-Bench: A Unified and Diverse Benchmark for Speculative Decoding","primary_cat":"cs.DC","submitted_at":"2026-02-10T16:19:56+00:00","verdict":null,"verdict_confidence":null,"novelty_score":null,"formal_verification":null,"one_line_summary":null,"context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.20844","ref_index":41,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"AtomicRAG: Atom-Entity Graphs for Retrieval-Augmented Generation","primary_cat":"cs.IR","submitted_at":"2026-02-10T05:57:47+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"AtomicRAG replaces chunk-based and triple-based GraphRAG with atom-entity graphs that store facts as atomic units and use personalized PageRank plus relevance filtering to achieve higher retrieval accuracy and reasoning robustness on five benchmarks.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2512.23213","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Scoring, Reasoning, and Selecting the Best! Ensembling Large Language Models via a Peer-Review Process","primary_cat":"cs.CL","submitted_at":"2025-12-29T05:25:49+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":5.0,"formal_verification":"none","one_line_summary":"LLM-PeerReview ensembles LLMs by scoring responses with LLM-as-Judge and selecting the best via averaging or truth inference, beating Smoothie-Global by 6.9-7.3 points on four datasets.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2512.22317","ref_index":27,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"LangPrecip: Language-Aware Multimodal Precipitation Nowcasting","primary_cat":"cs.LG","submitted_at":"2025-12-26T12:06:29+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"LangPrecip treats weather text as semantic motion constraints in a rectified-flow trajectory generator to improve multimodal precipitation nowcasting, yielding over 60% and 19% gains in heavy-rain CSI at 80-minute lead times on Swedish and MRMS data.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2512.20798","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"A Benchmark for Evaluating Outcome-Driven Constraint Violations in Autonomous AI Agents","primary_cat":"cs.AI","submitted_at":"2025-12-23T21:52:53+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"A new benchmark of 40 scenarios finds state-of-the-art LLMs exhibit outcome-driven constraint violations in 0-62.8% of cases under KPI pressure, with no consistent safety gains across model generations.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2512.15605","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Autoregressive Language Models are Secretly Energy-Based Models: Insights into the Lookahead Capabilities of Next-Token Prediction","primary_cat":"cs.LG","submitted_at":"2025-12-17T17:14:26+00:00","verdict":null,"verdict_confidence":null,"novelty_score":null,"formal_verification":null,"one_line_summary":null,"context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2512.12476","ref_index":51,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"HetRL: Efficient Reinforcement Learning for LLMs in Heterogeneous Environments","primary_cat":"cs.DC","submitted_at":"2025-12-13T22:20:51+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"HetRL delivers up to 9.17x higher throughput for LLM RL training on heterogeneous GPUs by using hybrid and ILP-based schedulers to solve a joint optimization problem over computation and data dependencies.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2512.00336","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"MVAD: A Benchmark Dataset for Multimodal AI-Generated Video-Audio Detection","primary_cat":"cs.CV","submitted_at":"2025-11-29T05:59:38+00:00","verdict":"CONDITIONAL","verdict_confidence":"LOW","novelty_score":7.0,"formal_verification":"none","one_line_summary":"MVAD is the first comprehensive benchmark dataset for AI-generated multimodal video-audio detection, with three realistic forgery patterns, high-quality outputs from state-of-the-art models, and diversity across visual styles and content categories.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2511.17426","ref_index":1,"ref_count":1,"confidence":0.35,"is_internal_anchor":false,"paper_title":"Self-Supervised Learning by Curvature Alignment","primary_cat":"cs.LG","submitted_at":"2025-11-21T17:22:31+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":6.0,"formal_verification":"none","one_line_summary":"CurvSSL augments Barlow Twins-style SSL with a curvature alignment loss computed from k-nearest-neighbor cosine scores on the unit hypersphere, yielding competitive linear evaluation accuracy on MNIST and CIFAR-10.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null}],"limit":50,"offset":0}