INDEX
Every paper Pith has read and judged, in one searchable index.
-
eess.SP arXiv submitted 2026-08-01Uncertainty score lifts RF modulation accuracy to 92.6%
Nurettin Safak +5 · “An Uncertainty-Driven Hybrid Deep Learning Approach for Broad-Coverage RF Modulation Recognition”
2608.00796 -
cs.CG arXiv submitted 2026-08-01Exact Steiner-point tree algorithm ignores the explosion of k
Eungyu Woo +1 · “Exact Algorithms for Minimum Steiner Point Trees”
2608.00795 -
cs.AI arXiv submitted 2026-08-01Agentic benchmark scores retain at most 36% of intended signal
William Caban · “Measurement Without Validity: The Compounding Reliability Problem in Agentic AI Evaluation”
2608.00794 -
cs.RO arXiv submitted 2026-08-01Two motion cues beat static-frame robot policies by 22.9 points
Yunfan Lou +8 · “DynamicWAM: Dual-Path Motion Conditioning for World-Action Models in Dynamic Manipulation”
2608.00793 -
cs.RO arXiv submitted 2026-08-01StochSIPP gives provably safe routes that cut arrival time
Ajith Kemisetti +2 · “StochSIPP: Safe Interval Path Planning in Stochastic Dynamic Environments”
2608.00792 -
cs.CL arXiv submitted 2026-08-01Distill only failed prompts to beat naive RL-plus-distillation
Zhuowen Han +11 · “Distill Where You Fail: Recovering Learning Signals of Negative RL-Groups from Adaptive Teacher Guidance”
2608.00782 -
cs.RO arXiv submitted 2026-08-01Map-free predictor beats HD-map models by reading BEV per agent
Feifei Liu +8 · “SIPTraj: Map-Free End-to-End Trajectory Prediction via Physics-Guided Scene Interaction”
2608.00779 -
cs.LG arXiv submitted 2026-08-01A frozen image model beats quantum descriptors for yield prediction
Qiwei Han +1 · “Generic Vision and Cross-Attention for Reaction Yield Prediction”
2608.00776 -
cs.RO arXiv submitted 2026-08-01One mixed-reality backend runs four robots from teach-in or voice
Ivan Snegirev +5 · “ORCESTRA: VLM-driven Visual Robot programming in Mixed Reality”
2608.00775 -
cs.CV arXiv submitted 2026-08-01One-step video editing cuts warping error 78% and flicker 49%
Zhiqiang Lao · “ChordVideo: One-Step, Training-Free, Temporally Consistent Video Editing via Low-Energy Transport”
2608.00769 -
cs.LO arXiv submitted 2026-08-01SAT solves all Binairo
Lukas Zandomeneghi +2 · “Taming the Search Space: Solving and Generating Hitori and Binairo Puzzles”
2608.00768 -
cs.GR arXiv submitted 2026-08-01Training-free 3D part splitter outdoes learned models on PartNet
Ruoyu Wu +7 · “Hi-TOPS: Hierarchical Topology-aware Scoring Prior for 3D Part Decomposition”
2608.00767 -
cs.CL arXiv submitted 2026-08-01Images make RAG 8x cheaper, 16 points more accurate
Jiayang Yu +2 · “RAGOCR: Optical Compression of Retrieval-Augmented Text via Visual Representation”
2608.00765 -
cs.AI arXiv submitted 2026-08-01Best deep-research agents hit ~40% on financial indicators
Chaoqun Yang +7 · “FinDeepIndicator: Benchmarking Deep Research Agents in End-to-End Financial Indicator Construction”
2608.00764 -
cs.HC arXiv submitted 2026-08-01Device type barely moves face-login accuracy
Miriam Palmetshofer +2 · “Continuous Face Authentication on Mobile and Desktop Platforms: A Comparative Study”
2608.00763 -
math.NA arXiv submitted 2026-08-01Reversing multigrid steps builds exact adjoints for Bayesian inference
Andr\'es Ram\'irez +11 · “Variational Inference Using a Differentiable Multigrid Linear Solver”
2608.00760 -
cs.SE arXiv submitted 2026-08-01Cheaper test runs beat real-system tuning in 5 of 6 cases
Yulong Ye +2 · “Less Is More: Tuning Configurable Systems with Imperfect Fidelity”
2608.00759 -
cs.CV arXiv submitted 2026-08-01One neural field fuses all heart MRI views and fixes slice motion
Nil Stolt-Ans\'o +4 · “NISF++: Geometrically-grounded implicit representations of 3D+time cardiac function from 2D short- and long-axis MR views”
2608.00752 -
cs.IR arXiv submitted 2026-08-01Dense token rewards lift session utility and ad target metrics
Kaifeng Guo +8 · “Hierarchical Residual Policy Optimization for Generative Recommendations”
2608.00750 -
cs.CY arXiv submitted 2026-08-01Configured prompt lifts AI teaching accessibility from 24% to 97%
Hector R. Amado-Salvatierra · “A Protocol for Evaluating the Accessibility of AI-Generated Educational Materials: Prompt Configuration, WCAG-Derived Criteria, and Content Overload”
2608.00749 -
cs.HC arXiv submitted 2026-08-01Only 45% of AI-companion Reddit posts are about users' own bots
Richard A. Fabes · “Me and My Bot: What Users Talk About in AI Companion Communities on Reddit”
2608.00748 -
cs.RO arXiv submitted 2026-08-01Up to 100% of runs: one injected robot hijacks its teammates
Neha Nagaraja +2 · “When Prompts Control Robots: Prompt Injection Attacks in Multi-Agent Robotic Systems”
2608.00747 -
cs.RO arXiv submitted 2026-08-01Injected text makes clean robots act in 100% of runs
Neha Nagaraja +2 · “When Prompts Control Robots: Prompt Injection Attacks in Multi-Agent Robotic Systems”
2608.00747 -
cs.AI arXiv submitted 2026-08-01Timing regularity exposes adaptive malware at 93% detection
Zihan Luo · “Behavioral Grammar: Detecting Adaptive Malware via Tiny Language Model Priors and Second-Order Temporal Analysis”
2608.00745 -
cs.CV arXiv submitted 2026-08-01Answer labels alone can train latent visual reasoning
Jiaxuan Kang +8 · “LUT: Latent Utility Training for Visual Reasoning”
2608.00743 -
cs.DC arXiv submitted 2026-08-01Slingshot HPC runs multi-tenant Kubernetes at full fabric speed
Jake Watson +8 · “Multi-tenant Kubernetes Use Cases for AI, Secure Computing and Data Services, and More”
2608.00742 -
cs.CV arXiv submitted 2026-08-01Image editors beat restoration specialists on art damage
Mridula Vijendran +2 · “MDTD-ArtIR: Benchmarking Image Editing and Restoration Models for Art Image Restoration under Texture-Overlay Degradations”
2608.00736 -
cs.LG arXiv submitted 2026-08-01Trap backdoors in a decoy branch
Zixuan Zhu +3 · “Mitigating Backdoors via Decoy Shortcuts and Knowledge Decoupling”
2608.00732 -
cs.CV arXiv submitted 2026-08-01Event-camera panoramas jump 13 dB with learned spherical pipeline
Zhenyang Li +4 · “E2Pano: Learning Event-to-Panorama Image Reconstruction”
2608.00694 -
cs.CL arXiv submitted 2026-08-01One attention pass ranks every SQL schema candidate
Jinwang Song +5 · “AttnLink: Turning Attention into Schema Links for Text-to-SQL”
2608.00693 -
cs.SE arXiv submitted 2026-08-01Agents fix 41 of 100 Python vulnerabilities
Tan Bui +6 · “Vul4Py: Benchmarking Automated Vulnerability Repair in Python with Paired Exploit and Functional Oracles”
2608.00692 -
cs.NI arXiv submitted 2026-08-01Selecting helpers before LLM reasoning improves cooperative driving
Ahmad Sarlak +3 · “LLM-Assisted Coalition Formation for Cooperative Perception in Autonomous Driving”
2608.00690 -
cs.CV arXiv submitted 2026-08-01LiDAR codec survives 70% bitstream truncation
Yihan Qiu +4 · “Proteus: A Truncation-Robust Entropy Model for Progressive LiDAR Compression”
2608.00687 -
cs.CY arXiv submitted 2026-08-01Developers owe AI assistant users four fiduciary duties
Benjamin Lange · “AI Alignment and Fiduciary Obligation”
2608.02660 -
cs.AI arXiv submitted 2026-08-01More LLM calls buy ~4 accuracy points at 2–4x tokens
Nicolas Leins +3 · “When Does LLM Orchestration Pay Off? A Controlled Evaluation of Accuracy, Cost, and Task Difficulty”
2608.00685 -
cs.CV arXiv submitted 2026-08-01Split light and color to erase shadows in satellite images
Wei Lu +2 · “BRIC-Net: Boundary-Reliable Illumination-Color Interaction for Remote Sensing Image Deshadowing”
2608.00682 -
cs.AI arXiv submitted 2026-08-01AI evaluation gets a nine-dimension process check
Abdalla Doleh +1 · “Multi-Dimensional Assessment for AI Cognition (MAAC): A Theoretical Framework for Process-Oriented Cognitive Evaluation of Text-Based AI Systems”
2608.00680 -
cs.AI arXiv submitted 2026-08-01Graph AI safety filter cuts EV-grid voltage violations
Xiangwei Wang +4 · “HetGPS: Scalable Graph Multi-Agent Reinforcement Learning with Physics-Anchored Adaptive Safety for EV Charging”
2608.00679 -
cs.CV arXiv submitted 2026-08-01Rotation-invariant training constraint cuts tilt-induced depth error
Kaihua Tang +6 · “Breaking the Horizontal Prior: From Long-Tailed Orientation Bias to Roll-Robust Monocular Depth Estimation”
2608.00678 -
cs.CL arXiv submitted 2026-08-01Evolving environments breach AI agents 85% of the time
Yunhao Chen +8 · “OpenART: Scaling Agent Red Teaming via Open-Ended Environment Evolution”
2608.00677 -
stat.ML arXiv submitted 2026-08-01One round trip predicts rollout error without ground truth
Alexander Scheinker · “Round-Trip Consistency: Bidirectional Diffusion Models Can Predict Their Own Rollout Errors”
2608.00675 -
cs.CV arXiv submitted 2026-08-01One proxy image sharpens fine-grained subject identity in seconds
Peng Zheng +3 · “CopyCat: Improving Fine-Grained Subject Consistency in Subject-to-Image Models within Seconds”
2608.00674 -
cs.CE arXiv submitted 2026-08-01Deep learning reconstructs all vibration modes from one eigenvector
Sebastian Rodriguez +5 · “A plausible Parametrization of Modal Basis for Dynamical Systems Analysis”
2608.00673 -
cs.CR arXiv submitted 2026-08-01SOC practitioners rate LLMs unready for incident analysis
Jonas Thurner +6 · “From Chasing Ghosts to Missed Attacks: Perspectives and Perceptions of SOC Practitioners on LLM Integration, Risks, and Readiness”
2608.00672 -
cs.IR arXiv submitted 2026-08-01Graph grounding beats prompt-only KG in 24 of 27 tests
Yong Wang +6 · “GARDRec: Decision-Level Graph Grounding for Large Language Model Recommendation”
2608.00669 -
math.NT arXiv submitted 2026-08-01Deterministic integer-root finding drops to Õ(n^{3/2} b)
Itamar Nir · “Faster Deterministic Integer Root Finding for Integer Polynomials”
2608.00668 -
eess.AS arXiv submitted 2026-08-01Cut modal error 66% with count-first resonator fitting
Minhui Lu +1 · “Band-Count Dense Modal Estimation with Fixed-Frequency Differentiable Resonator Refinement”
2608.00667 -
cs.CE arXiv submitted 2026-08-01Converter asymmetry sends 97% of Polymarket arbitrage through one path
Jonas Gebele +2 · “Executable Arbitrage and Market Efficiency in Prediction Markets”
2608.00666 -
cs.CV arXiv submitted 2026-08-01Blendshape geometry hands talking faces a continuous emotion dial
Chenggong Hu +5 · “Geometry-guided Emotion Modulation for Controllable and Photorealistic Emotional Talking Face Generation”
2608.00663 -
cs.SE arXiv submitted 2026-08-0186,726 code errors: even top AI models skip basic checks
Rodrigo Pato Nogueira +2 · “Unreliable in Practice? A Comprehensive Study of Errors in LLM-Generated Code”
2608.00661