INDEX
Every paper Pith has read and judged, in one searchable index.
-
cs.RO arXiv submitted 2026-07-30One policy grasps, moves, spins, and slides held objects
Hui Zhang +3 · “UniCross: Unified Cross-Skill Dexterous Manipulation Synthesis”
2607.28198 -
cs.CL arXiv submitted 2026-07-30Compressed LLMs clear every guard then invent procedure steps
I. Kennedy +1 · “Fidelity Is Not Safety: Gently-Compressed LLMs Pass Every Data-Free Quality Guard Yet Invent Procedure Steps in Agentic Execution”
2607.28196 -
cs.DC arXiv submitted 2026-07-30One continuum substrate runs control and monitoring CPS
Fabio Orazio Mirto +9 · “A Cloud Continuum Research Infrastructure for Distributed CPS Experimentation”
2607.28193 -
cs.CR arXiv submitted 2026-07-30Secure EEG federated learning hides each hospital’s update
Pouya Rajabi +1 · “Secure Aggregation for Privacy-Preserving Federated Learning on Clinical EEG Data”
2607.28191 -
cs.CL arXiv submitted 2026-07-30Trial interviews scored to MADRS totals at 0.867 Spearman
Mila Fodor +4 · “The MADRS Pipeline: Supporting Depression Assessment in Clinical Trials”
2607.28190 -
cs.AI arXiv submitted 2026-07-30Simple image edits bypass commercial AI moderators
Marco Alecci +4 · “Old Tricks, New Models: How Simple Image Transformations Break Modern AI-based Content Moderation”
2607.28187 -
cs.CV arXiv submitted 2026-07-30Farm maps improve when the model asks for missing time and space
Haiyang Wu +6 · “Think with Extra-Image: A Farmland Segmentation Agent Driven by Spatio-Temporal Information Gain”
2607.28186 -
cs.LG arXiv submitted 2026-07-30Gaussian noise stops graph nets from oversmoothing
Mostafa Haghir Chehreghani · “Persistent Gaussian Perturbations Prevent Oversmoothing in Recurrent Graph Neural Networks”
2607.28185 -
cs.LG arXiv submitted 2026-07-30Local causal reallocations beat global budget optimizers
Changjian Liu +9 · “Multi-channel Uplift Policy Learning”
2607.28182 -
cs.CY arXiv submitted 2026-07-30Digital tabletops scale cyber team training in class
Jan Vykopal +3 · “Technology-Enhanced Tabletop Exercises for Cybersecurity Education: Lessons Learned”
2607.28179 -
cs.SE arXiv submitted 2026-07-30AI in RE class works when students revise first
Hansika Ekanayake Mudiyanselage +3 · “Integrating AI into Requirements Quality Learning in Software Engineering Education: A TPACK-Guided Empirical Study”
2607.28176 -
cs.AI arXiv submitted 2026-07-30Clean speech text that rewrites itself mid-utterance
Zixuan Jiang +5 · “AgenticASR: Refining Speech Recognition in Real-World Scenarios via an Agentic Approach”
2607.28175 -
cs.LG arXiv submitted 2026-07-30Two search rules beat prior optimal-tree solvers
Jacobus G. M. van der Linden +2 · “Search Strategies for Optimal Classification and Regression Trees”
2607.28170 -
math.OC arXiv submitted 2026-07-30Two graph steps pick leaders for full matrix-network control
Lanhao Zhao · “On Leader Selection for Strong Structural Controllability in Matrix-Weighted Networks”
2607.28168 -
cs.CL arXiv submitted 2026-07-30Verify the answer span, not a confidence proxy, to exit DLMs early
Chia-Ming Lee +4 · “Where and When to Commit: Candidate-Aware Decoding for Diffusion Language Models”
2607.28166 -
cs.CV arXiv submitted 2026-07-30One photo yields a 3D head avatar that holds up from the back
Hail Song +4 · “S-Avatar: Diffusion-Guided Gaussian Head Avatars from a Single Image”
2607.28164 -
cs.IT arXiv submitted 2026-07-30Level-k bits best tell two correlations apart
Jun Chen +2 · “On The Most Discriminative Boolean Functions for Correlated Sources”
2607.28162 -
cs.DS arXiv submitted 2026-07-30Three-block graph matching is near-linear
Sebastian Angrick +4 · “String Matching in (Block) Graphs: A Full Classification by Walk Length”
2607.28159 -
eess.SY arXiv submitted 2026-07-30Local sensors track global Boolean state with a proven error bound
Rong Yang +1 · “Stochastic Average Consensus Filtering and Distributed State Estimation for Boolean Control Networks”
2607.28158 -
cs.CL arXiv submitted 2026-07-30Agents learn how to search long videos
Jingxiang Fan +2 · “RRM: Experience-Driven Reflective Retrieval Memory for Long-Horizon Multimodal Reasoning”
2607.28156 -
cs.HC arXiv submitted 2026-07-30Math makes torn animated scores still playable
Patricio F. Calatayud +2 · “A Mathematical Framework for Reading the Autopsias' Meta - Compositional System”
2607.28155 -
cs.CV arXiv submitted 2026-07-30Latent distillation absorbs multimodal CoT without help images
Shoutai Zhu +5 · “OPLD: On-Policy Latent Distillation for Multimodal Reasoning”
2607.28154 -
cs.DC arXiv submitted 2026-07-30Selective KV transfer cuts second-token wait by 4.3×
Xuchuan Luo +3 · “SmartGen: Seamless Disaggregated LLM Inference with Selective KV Cache Transfer”
2607.28150 -
cs.CV arXiv submitted 2026-07-30Tongue AI gains 20% from data
Longxia Gao +5 · “What Makes Deep Learning Work for Traditional Chinese Medicine Tongue Diagnosis? A Comprehensive Ablation Study”
2607.28148 -
cs.CR arXiv submitted 2026-07-30Black-box queries steal multi-agent orchestration harnesses
Yu Cui +6 · “Agent Harness Distillation: Inference-Time Harness Extraction and Exploitation in Autonomous Multi-Agent Systems”
2607.28147 -
cs.CL arXiv submitted 2026-07-30Most AI agents can't keep a lie for a whole game
Niklas Bauer +5 · “Can Agents Deceive? Evaluating Reasoning and Deception in ParliamentBench using a Social Deduction Game”
2607.28146 -
cs.SE arXiv submitted 2026-07-30GPU ragged-array cuts run 5–12× faster
Ianna Osborne +1 · “Advancing Awkward Arrays for High-Performance CPU and GPU Processing”
2607.28145 -
eess.IV arXiv submitted 2026-07-30Talking-head video at 1/10 the bitrate
Zheyuan Zhang +1 · “ReGenVC: End-to-End Real-Time Generative Video Coding at Ultra-Low Bitrate”
2607.28144 -
eess.SY arXiv submitted 2026-07-30One projection cancels actuator dead-zones from raw data
Mahmood Mazare +1 · “Data-Driven Dead-Zone Compensation via Projection in Predictive Control Setting”
2607.28142 -
cs.CY arXiv submitted 2026-07-30LLMs don’t hold commitments—only humans do
Enzo Fenoglio · “Asymmetric Communication: Large Language Models and Language Games”
2607.28137 -
cs.DS arXiv submitted 2026-07-30Depth-first k²-trees cut space and memory on matrix ops
Gabriel Carmona +3 · “Extended Depth-First Representations of $k^2$-trees”
2607.28136 -
cs.LG arXiv submitted 2026-07-30Online language model beats GPU-GRASP by 28 makespan units
Mohand Mezmaz +1 · “LM-GRASP: Instance-Specific Language Models for Combinatorial Construction via Online Imitation Learning”
2607.28135 -
cs.CV arXiv submitted 2026-07-30Local context in the shader sharpens multi-view 3D meshes
Juheon Hwang +3 · “Convolutional Neural Shading for High-Quality 3D Reconstruction from Multi-View Images”
2607.28132 -
cs.CV arXiv submitted 2026-07-30Multiple blurry faces fuse into one sharp identity prior
Juheon Hwang +2 · “Collaborative Feature Aggregation for Face Super-Resolution and Robust Re-Identification”
2607.28130 -
cs.CV arXiv submitted 2026-07-30Convex hulls pull face and voice of one person together
Taewan Kim +1 · “Face and Voice Cross-modal Association with Learning Convex Feature Embedding”
2607.28129 -
cs.CL arXiv submitted 2026-07-30Helpfulness scores miss when tutors just give the answer
Shuyi Fan +6 · “Rethinking LLM-Judged Helpfulness as a Pedagogy Signal: A Pre-Registered Audit Across Tutor Models”
2607.28128 -
cs.CL arXiv submitted 2026-07-30Train sentiment on market returns
Giorgos Iacovides +2 · “FinSMART: Financial Sentiment Analysis for Algorithmic Trading through Market-Aligned Reinforcement Learning”
2607.28127 -
cs.AI arXiv submitted 2026-07-30Contribution-ranked memory cuts inspection tokens 88%
Bingchen Liu +8 · “ConMem: Contribution-Aware Memory for Long-Horizon Manufacturing Inspection Logs”
2607.28126 -
cs.CV arXiv submitted 2026-07-30No-label vote picks the best UDA model for clinical scans
Yiheng Xiong +4 · “Towards Practical Algorithm Selection for Unsupervised Domain Adaptation in Medical Imaging”
2607.28125 -
cs.LG arXiv submitted 2026-07-30Only 14–20% of history yields accurate
Xu Zheng +5 · “Information Bottleneck Learning for Faithful Time Series Forecasting Explanations”
2607.28124 -
eess.SY arXiv submitted 2026-07-30Regularized projection cuts error under process noise
Mahmood Mazare +1 · “Projection-Regularized Indirect Data-Driven Predictive Control”
2607.28123 -
cs.CL arXiv submitted 2026-07-30LLMs beat student annotators on hard Appraisal labels
Mirela Imamovic +3 · “Challenges in annotations by humans and LLMs: A case study of evaluative language”
2607.28119 -
eess.SP arXiv submitted 2026-07-30Linear RUL labels miss how bearings actually vibrate toward failure
Behrad Mousaei Shir-Mohammad +5 · “When Linear RUL Labels Disagree with Vibration Degradation: A Stage-Aware Target and Dual-Scale Predictor Evaluated on XJTU-SY and IMS”
2607.28115 -
cs.NI arXiv submitted 2026-07-306G sites cut energy 11% by treating power as packets
Adnan Aijaz +1 · “Powering Net-Zero 6G: Packetized Energy Management for Grid-Interactive Telecom Infrastructure”
2607.28111 -
cs.AI arXiv submitted 2026-07-30Typed edits fix failed Lean blueprints nearly as often
Ruslan Khrulev · “BlueprintRepair: Typed Local Edits for Failed Lean Proof Blueprints”
2607.28110 -
cs.AI arXiv submitted 2026-07-30Book-shaped synthetic textbooks beat rewritten snippets
Jiawen Tao +9 · “Beyond Rephrasing: Book-Level Organization Improves Synthetic Textbook Data for Mid-Training”
2607.28109 -
cs.CV arXiv submitted 2026-07-30Indoor radar twin recalls 71% of real response regions
Jianyi Zhou +3 · “mmRadarTwin: A Measurement-Calibrated Signal-Level Digital Twin Platform for Indoor mmWave Radar”
2607.28108 -
math.NA arXiv submitted 2026-07-30Adaptive level-set transport matches fine grids at interface cost
Eugenio Aulisa +4 · “An Adaptive Finite Element Method for Marker-Driven Level-Set Transport on Hierarchical Meshes”
2607.28104 -
cs.AI arXiv submitted 2026-07-30Memory defense halves agent attack success without extra lag
Dongyi Liu +3 · “MIND: Lightweight and Effective Memory Injection Defense for LLM Agents via Intent-Aware Information Bottleneck”
2607.28103 -
cs.DC arXiv submitted 2026-07-30Bounded Stern–Brocot IDs cut CRDT lag and memory
Moulay Driss Mechaoui +1 · “ESBT: A Scalable and Deterministic Sequence CRDT for Distributed Collaborative Editing”
2607.28101