archive
Every paper Pith has read. Search by title, abstract, or pith.
14903 papers in cs.LG · page 2
-
Bootstrapped GRTO unifies RL and tool training for segmentation
B-GRTO: Bootstrapped Group Relative Tool Optimization for Referring Segmentation
-
Self-generated tests and code co-evolve to match RLVR results
CoSPlay: Cooperative Self-Play at Test-Time with Self-Generated Code and Unit Test
-
Non-normal operators flag neural network training instabilities
Non-normal spectral signatures of instability in neural network training dynamics
-
DSEBO switches subspace dimension on convergence
Automated Random Embedding for Practical Bayesian Optimization with Unknown Effective Dimension
-
CBANet raises minority recall in aggressive driving detection
CBANet: A Compact Attention-Based CNN-BiLSTM Network for Aggressive Driving Event Detection
-
Static contexts make individual dynamics identifiable from single snapshots
Learning Individual Dynamics from Sparse Cross-Sectional Snapshots
-
S³GNN cuts oversquashing errors up to 10x with 50% fewer parameters
S$^3$GNN: Efficient Global Mixing and Local Message Passing for Long-Range Graph Learning
-
Time-varying transforms block model extraction in sharded training
Unextractable Protocol Models: Collaborative Training and Inference without Weight Materialization
-
Hybrid augmentation raises migraine F1 average to 0.862
Class-Dependent Hybrid Data Augmentation for Multiclass Migraine Classification under Severe Class Imbalance
-
VAE decoder learns to respect non-commutative latent order
Commutator-Induced Uncertainty in VAEs
-
k-WL cannot distinguish all simple-spectrum graphs for any k
Weisfeiler-Leman Is Incomplete on Simple Spectrum Graphs, so Canonicalize Them
-
Onsager-Machlup path transport beats DBVI on large DGP tasks
Onsager-Machlup Posterior Transport for Deep Gaussian Processes
-
Shortest-path tree cuts in-network training traffic by 70%
Sparse In-Network Learning via Shortest-Path Backpropagation and Finite-Rate Gating
-
Oblique trees gain quadratic approximation via Newton least-squares
Hinge Regression Trees and HRT-Boost: Newton-Optimized Oblique Learning for Compact Tabular Models
-
500K trajectories enable BBO foundation models
An Open-Source Training Dataset for Foundation Models for Black-box Optimization
-
Reflection symmetry speeds up state-based RL
Reflex: Reinforcement Learning with Reflection Symmetry Exploitation in State-Based Continuous Control
-
Consistency checks raise LLM multi-agent planning success 9.75%
When Planning Fails Despite Correct Execution: On Epistemic Calibration for LLM-Based Multi-Agent Systems
-
Sample-wise attacks fool TTA while keeping label counts normal
Sample-wise Targeted Adversarial Attacks on Test-time Adaptation
-
Multi-view probes read model weights more accurately
What Linear Probes Miss: Multi-View Probing for Weight-Space Learning
-
Quantum circuits in UNet boost wind downscaling metrics
Hybrid Quantum-Classical Corrective Diffusion Modeling for Meteorological Downscaling
-
PPM maps parametric priors into generative forecasts
Parametric Prior Mapping Framework for Non-stationary Probabilistic Time Series Forecasting
-
Convex factors let energy models scale to larger reasoning tasks
Convex Compositional Reasoning Models
-
One recursion decomposes every component into paths and token credit
Every Component is a Lookup: Token Attribution and Composition from a Single Decomposition
-
Preconditioning caps PINN kernel radius independent of coupling
Coupling-Robust Accuracy in Multiphysics Physics Informed Neural Networks via Kronecker-Preconditioned Optimization
-
Selective dual dispatch improves ambulance response with less fleet use
Selective Ambulance Dispatch Under Contextual Travel-Time Uncertainty
-
VAE turns non-Euclidean tasks into measurable space for RL curricula
Curriculum reinforcement learning with measurable task representation learning
-
One-step MeanFlow policy hits SOTA on locomotion tasks
Score-Based One-step MeanFlow Policy Optimization
-
Adaptive allocation matches oracle rate for multi-judge LLM scoring
Instance-Optimal Estimation with Multiple LLM Judges on a Budget
-
Prudent-Banker keeps bandit safety at constant cost amid delays
Prudent-Banker: No Extra Fees for Baseline Safety in Adversarial Bandits With and Without Delays
-
CDM amortizes Twisted SMC for discrete diffusion with under 5% overhead
Contrastive Distribution Matching for Amortized Sequential Monte Carlo in Discrete Diffusion
-
Spin field model infers traffic phases from trajectories
SpinFlow: A Physics-Informed Spin Field Framework for Traffic Phase Inference and Transition Detection
-
Hybrid search boosts Max-Cut solutions on photonic Ising machines
Accelerating ground state search of spatial photonic Ising machines with genetic-simulated annealing hybrid algorithm
-
Reinforcement learning enforces exact graph assortativity without tuning
Reinforcement Learning for Microcanonical Graph Ensemble with Assortativity Constraints
-
Neural operator deblurs varying blur in pathology slides
Discontinuous Galerkin Neural Operator for Pathology Defocus Deblurring
-
Compact coordinator expands diffusion models to larger domains
Diffusion Domain Expansion: Learning to Coordinate Pre-trained Diffusion Models
-
Structural priors fix bad scores for good equations in symbolic regression
When Good Equations Get Bad Scores: Improving Symbolic Regression Through Better Parameter Optimization
-
Joint training avoids error inheritance from weak privileged data
Coupled Training with Privileged Information and Unlabeled Data
-
Multi-gate residuals stabilize deep nets without extra comms cost
Multi-Gate Residuals
-
Plug-in adds reconstructability routing to KV cache eviction
A Simple Plug-in for Improving Eviction-Based KV Cache Compression
-
Algorithms keep preemptions constant per job using predictions
Learning-Augmented Online Scheduling with Parsimonious Preemption
-
RefCal jointly optimizes calibration and refinement in DNNs
Enhancing Deep Neural Network Reliability with Refinement and Calibration
-
Neural network predicts DLT times 10-100x faster
Accelerating Divisible Load Processing Through Machine Learning: A Practical Framework for Large-Scale Workloads
-
Convex optimization aligns LLMs on one GPU without reference model
Convex Optimization for Alignment and Preference Learning on a Single GPU
-
Multi-perspective clustering pre-trains relational database models
RelPrism: A Multi-Faceted Pre-training Framework with Self-Generated Tasks for Relational Databases
-
Purifier cleans adversarial graphs for any GNN
Self-supervised Adversarial Purification for Graph Neural Networks
-
Generated card games expose jagged strategic skills in top LLMs
GENSTRAT: Toward a Science of Strategic Reasoning in Large Language Models
-
Convex optimization detects accented languages at 97-98% accuracy
Convex Low-resource Accent-Robust Language Detection in Speech Recognition
-
Movement patterns can expose unfairness in predictive models
Assessing Predictive Models for Fairness Based on Movement Patterns
-
Entropy testing needs fewer samples than closeness testing
Entropy Equivalence Testing
-
Automated search yields stronger attacks on world-model agents
WMAttack: Automated Attack Search for Adversarial Evaluation of World-Model Agents