INDEX
Every paper Pith has read and judged, in one searchable index.
-
stat.ME arXiv submitted 2026-08-03Six feature sets tie 85% of time-series comparisons
Trent Henderson +1 · “Statistical comparisons of time-series feature sets on classification tasks”
2608.01586 -
cs.CL arXiv submitted 2026-08-03Topological test reads meaning inside language models
Tyler Ashoff +1 · “Semantic Alignment of AI Models: Concept Collapse, Checkpoint Dynamics, and Cross-Lingual Transfer”
2608.01585 -
cond-mat.stat-mech arXiv submitted 2026-08-03Trajectory data alone yields the symmetry algebra of an SDE
Shida Liu +3 · “LieStoNet: Learning Lie Symmetries from Spatiotemporal Data for Stochastic Dynamical Systems”
2608.01582 -
cs.NI arXiv submitted 2026-08-01Delayed task starts lift on-time execution by 70 percent
Keyvan Aghababaiyan +2 · “Execution Timing Control for Deterministic Task Offloading in the IoT-Edge-Cloud Continuum”
2608.00892 -
cs.CR arXiv submitted 2026-08-01Canonical prompts hide up to 13% of unsafe LLM answers
Yongxi Zhou +6 · “Single Canonical Prompts Underestimate LLM Safety's Surface-Form Sensitivity”
2608.02665 -
cs.CV arXiv submitted 2026-08-012,020 math images now come with accessibility descriptions
Rebeka Popek +2 · “MIDAL: A Dataset of Math Image Descriptions for Accessible Learning”
2608.00868 -
cs.LG arXiv submitted 2026-08-01One 13k-parameter model gives calibrated bands for unseen ODEs
Ajitesh Srivastava · “HyperODE: Zero-Shot Surrogate for Simulation and Inference of Dynamical Systems”
2608.00852 -
cs.LG arXiv submitted 2026-08-01Adaptive training lifts quantum PINN accuracy by 60 percent
Fabio Pereira dos Santos +3 · “Adaptive Quantum Physics-Informed Neural Networks for Differential Equations with Applications to Fluid Dynamics”
2608.00850 -
cs.DS arXiv submitted 2026-08-01A Consistency-Robustness Framework for Robust Optimization: Integrating Predictions into Robust Scheduling
Yasser Alghouass +3
2608.00848 -
cs.CV arXiv submitted 2026-08-01Coordinated tools push LaSOT tracking AUC past 80
Wenrui Cai +3 · “Models as Tools: An Agentic Coordination Framework for Unified Multimodal Visual Tracking”
2608.00847 -
cs.CY arXiv submitted 2026-08-01Two-thirds of CS staff approve AI-drawn diagram help
Ethan Dickey +3 · “BoilerSketch: A TA-Supervised, Diagram-First GenAI Practice for Structured Diagrams in CS1/Early CS2”
2608.00844 -
math.NA arXiv submitted 2026-08-01Projection method keeps VMS accuracy at 1.3-2.7x lower cost
Biswajit Khara +3 · “A Helmholtz-Leray projection method with variational multiscale stabilization for the Navier-Stokes equations”
2608.00843 -
physics.class-ph arXiv submitted 2026-08-01Four constants stay fixed on optimal continuous-thrust orbital paths
Aimar Negrete +1 · “Conserved Quantities of Optimal Continuous-Thrust Trajectories in A Central Gravitational Field”
2608.00842 -
cs.DS arXiv submitted 2026-08-01Read a few columns, recover all large eigenvectors
Rajarshi Bhattacharjee +2 · “Sublinear Time Eigenvector Approximation via Column Sampling”
2608.00840 -
cs.CY arXiv submitted 2026-08-01Style feedback tool wins staff interest
Ethan Dickey +3 · “CodeStylist: Supporting Early Undergraduate Programmers with Course-Aware Code Style Feedback”
2608.00839 -
math.PR arXiv submitted 2026-08-01Superpolynomial Dean-Kawasaki accuracy needs strictly positive density
Ana Damnjanovi\'c +2 · “On the role of positivity preservation for high order approximations of the Dean--Kawasaki equation”
2608.00838 -
cs.CR arXiv submitted 2026-08-0132 verified zero-knowledge proofs attest fair lending on 32,768 rows
Mohammad Nasir Uddin +3 · “ZK-SR117: A Chunked Zero-Knowledge Attestation Design for Aggregated Fair-Lending Metrics, with a Control Mapping toward Full SR 11-7 Coverage”
2608.02664 -
cs.CL arXiv submitted 2026-08-01Pruned BPE shrinks token streams without adding vocabulary
Kenny Shao · “Pruned BPE: Post-training Visibility Pruning and Token Reallocation for Byte Pair Encoding”
2608.00837 -
cs.LG arXiv submitted 2026-08-01Beats prior GNNs on signed-directed graphs via a nonlinear Laplacian
Ali Parviz +1 · “Nonlinear Laplacians Improve Signed-Directed Graph Learning”
2608.00836 -
cs.LG arXiv submitted 2026-08-01Eight counter-plan samples cut plan-agent win rate .90 to .82
Alina Kapanova +3 · “AdvPlan-Bench: Adversarial Evaluation of Structured Plan-Generation Agents”
2608.00832 -
physics.med-ph arXiv submitted 2026-08-01Borrow a past patient's change to predict today's anatomy
Yizhou Wu +3 · “Anticipatory Digital Twins for Online Head-and-Neck Adaptive Proton Therapy via Foundation-Model Registration”
2608.00831 -
cs.RO arXiv submitted 2026-08-01A robot draws your portrait in 8.5 seconds
Miguel Altamirano Cabrera +3 · “GeminiPainter's sequence-formed pipeline comprised of perception, cognition, planning, and action stages”
2608.00829 -
cs.AI arXiv submitted 2026-08-01Sharp isotropy cliff predicts LLM accuracy at -0.84
Okan S. Coskun +3 · “Isotropy Cliffs: The Geometric Signature of Decision-Making in Large Language Models”
2608.00828 -
cs.LG arXiv submitted 2026-08-01ICU mortality model hinges on bidirectional graph links
Mohammad Nasir Uddin +6 · “CT-HEG: A Bidirectional, Timestamp-Attributed Event Graph for ICU In-Hospital Mortality Prediction - An Architectural Ablation Study”
2608.02663 -
cs.CR arXiv submitted 2026-08-01Six-factor score ranks XR healthcare security risks
Nafisa Anjum +1 · “XR-PRISM: Data-Driven Privacy and Risk Impact Scoring Metric for Extended Reality in Healthcare”
2608.00826 -
cs.LG arXiv submitted 2026-08-01Vortex-shedding equations recovered without any candidate library
Farbod Faraji +1 · “Verifier-Guided Model Discovery for Physical Dynamical Systems with Pretrained Symbolic Transformers”
2608.02662 -
cs.SE arXiv submitted 2026-08-01Stripping title context collapses annotator agreement from 0.45 to 0.04
Yang Yang · “Structure-Aware Semantic Chunking with Title-Chain Prefixes: A 1600-Query Evaluation and the Measurement Trap in Text-Transform Ablations”
2608.00824 -
cs.GT arXiv submitted 2026-08-01Freemium pays off exactly when data beats compute
Maxim Bichuch +1 · “Freemium Is All You Need”
2608.00823 -
cs.RO arXiv submitted 2026-08-01Riemannian informed sets cut 14-DOF path costs by up to 63.5 percent
Muhayy Ud Din +3 · “RIT*: Riemannian Informed Trees for Cost-Adaptive Optimal Motion Planning”
2608.00822 -
cs.CL arXiv submitted 2026-08-01Pure memorizers can mimic abstraction-first learning
Zachary Nicholas Houghton +1 · “Exemplars in Disguise: Pure Exemplar Models Mimic Abstraction-First Learning”
2608.00821 -
cs.RO arXiv submitted 2026-08-01LooperMuscle cuts body-tracking error 34% in 45 minutes
Boyi Liu +4 · “LooperMuscle: Fast and Stable Learning of Humanoid Whole-Body Tracking via Structured Mixture-of-Experts”
2608.00820 -
cs.AI arXiv submitted 2026-08-01Overestimating AI makes scaling it backfire
Anyan Qi +1 · “The Scaling Paradox in Human-AI Collaboration”
2608.00818 -
cs.AI arXiv submitted 2026-08-01Citations improve MD accuracy but make wrong answers harder to resist
Tirtha Chanda +12 · “Large language models improve physician accuracy but lead to false reliance”
2608.00817 -
cs.IR arXiv submitted 2026-08-01Exponential reward weights beat PPO and DPO for recommender tuning
Keertana Chidambaram +4 · “Exponential Reward Weighting for Fine-Tuning Generative Recommenders under Sparse and Noisy Feedback”
2608.00816 -
math.NA arXiv submitted 2026-08-01Clean Koopman spectra need no invariant dictionary
Siji Chen +2 · “Beyond Invariant Dictionary: Data-Driven Koopman Spectral Recovery with Filtered Extended Dynamic Mode Decomposition”
2608.02661 -
cs.LG arXiv submitted 2026-08-01Paris service density predicts fewer car trips
Andr\'as J. Moln\'ar +3 · “Paris as a 15-Minute City: An Explainable AI Perspective”
2608.00815 -
cs.LG arXiv submitted 2026-08-01Local services track with less car use in central Paris
Andr\'as J. Moln\'ar +3 · “Paris as a 15-Minute City: An Explainable AI Perspective”
2608.00815 -
cs.CL arXiv submitted 2026-08-01Sidecar predicts tool calls out of order for 5.34x speedup
Zhiheng Zhang +3 · “OoO-Spec: Out-of-Order Semantic Speculation for Fast Tool Calling”
2608.00814 -
cs.IT arXiv submitted 2026-08-01Reed-Muller service rates: every two-symbol region is now exact
Chenyang Yu +2 · “New Results Towards the Characterization of Service Rate Region of Reed-Muller Codes”
2608.00810 -
cs.SE arXiv submitted 2026-08-01Coding agents fix 8 more bugs per 100 with a no-model runtime layer
Zehao Wang +6 · “Turning Interaction History into Execution State: A Runtime Layer for Long-Horizon Coding Agents”
2608.00808 -
cs.AI arXiv submitted 2026-08-01Resource-aware benchmark: specialized agents win
Meher Bhaskar Madiraju +1 · “AgentSLABench: Evaluating and Benchmarking Agentic Systems Under Resource Constraints”
2608.00805 -
cs.SD arXiv submitted 2026-08-01Glasses read silent speech at 26.3% word error rate
Ruidong Zhang +3 · “SoniSpeech: A Large-Scale Open-Vocabulary Tri-Modal Dataset for Wearable Silent Speech Interfaces”
2608.00803 -
cs.CR arXiv submitted 2026-08-01Binding AI action logs to hardware attestation catches swapped models
Anton Sokolov · “Hardware-rooted attestation for AI-agent evidence: composing IETF RATS with action evidence packages”
2608.00801 -
cs.CV arXiv submitted 2026-08-01AIMold turns one CAD part into a full injection-mold assembly
Pengyun Qiu +5 · “AIMold: An Autonomous AI-based Pipeline for Complex Mold Design”
2608.00800 -
cs.CV arXiv submitted 2026-08-01Rebuilding CAD one operation at a time beats single-pass generation
Soslan Kabisov +11 · “CADENA: Stepwise CAD Reverse Engineering”
2608.00799 -
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