archive
Every paper Pith has read. Search by title, abstract, or pith.
14903 papers in cs.LG · page 7
-
ARC-STAR cuts PDE rollout error 36x on every cell
ARC-STAR: Auditable Post-Hoc Correction for PDE Foundation Models
-
ARC-STAR cuts PDE model error 36x across all regimes
ARC-STAR: Auditable Post-Hoc Correction for PDE Foundation Models
-
Attention mask forces transformer backtracking to ignore search history
Can Transformers Learn to Verify During Backtracking Search?
-
Strict gate stabilizes self-play RL regardless of reward
Survive or Collapse: The Asymmetric Roles of Data Gating and Reward Grounding in Self-Play RL
-
Kernel embeddings learn safe barriers during deep RL
Kernel-Based Safe Exploration in Deep Reinforcement Learning
-
9B model with skill modules beats 32B LLM
Skill Weaving: Efficient LLM Improvement via Modular Skillpacks
-
Video models top open suturing skill challenge
OSS: Open Suturing Skills Vision-Based Assessment Challenge 2024-2025
-
RL automates adaptive graphs of operations for LLM prompting
Reinforced Graph of Thoughts: RL-Driven Adaptive Prompting for LLMs
-
Two-point feedback lets prediction error set bandit regret
Bandit Convex Optimization with Gradient Prediction Adaptivity
-
GPU batches cut optimal sparse GLM search time by 10-100 times
From Sequential Nodes to GPU Batches: Parallel Branch and Bound for Optimal $k$-Sparse GLMs
-
Telematics and CV fusion boosts MLLM safety event detection
Enhancing Multimodal Large Language Models for Safety-Critical Driving Video Analysis
-
Infinite-order kernels raise neural operator accuracy
IKNO: Infinite-order Kernel Neural Operators
-
4B RL policy beats GPT-5 by picking expert models
Maestro: Reinforcement Learning to Orchestrate Hierarchical Model-Skill Ensembles
-
Metric shows VLM explainers miss text synergy
Measuring Cross-Modal Synergy: A Benchmark for VLM Explainability
-
Pairwise metric on logged pairs lifts latent planning success to 97 percent
Beyond Euclidean Proximity: Repairing Latent World Models with Horizon-Matched Trajectory Reachability Metrics
-
Language models treat star spectra as text to estimate parameters
Spectra as Language: Large Language Models for Scalable Stellar Parameter and Abundance Inference
-
LLMs treat stellar spectra as language sequences to estimate parameters
Spectra as Language: Large Language Models for Scalable Stellar Parameter and Abundance Inference
-
OWPO lets LLMs self-evolve without fixed references
One-Way Policy Optimization for Self-Evolving LLMs
-
Algebraic ML beats cross-validated CNNs on small images
Algebraic Machine Learning for Small-to-Medium Datasets Is Competitive against Strong Standard Baselines
-
Learned transfer keeps relevant facts in long-term KG memory
Short-Term-to-Long-Term Memory Transfer for Knowledge Graphs under Partial Observability
-
30B agents rival 1T models with 25-95% fewer tokens
Efficient Agentic Reasoning Through Self-Regulated Simulative Planning
-
Betting wealth bound yields empirical Bernstein LIL
From Betting to Empirical Bernstein LIL
-
ConvLSTM detects gamma-ray transients after learning from simulated sky
Self-Supervised ConvLSTM for Fermi Large Area Telescope Transient Detection
-
Physics-informed model recovers aerodynamic loads from noisy bridge data
Aerodynamic force reconstruction using physics-informed Gaussian processes
-
Text embeddings boost ImageNet accuracy by up to 2.7 points
TextTeacher: What Can Language Teach About Images?
-
Genetic search designs photonic quantum models reaching 99% accuracy
Q-PhotoNAS: Hybrid Quantum Neural Architecture Search Framework on Photonic Devices
-
Augmentations reduce TTS word error rate from 1.44 to 1.38
RobustSpeechFlow: Learning Robust Text-to-Speech Trajectories via Augmentation-based Contrastive Flow Matching
-
Transformer infers contact states to adapt robots on hardware
CoRMA: Contrastive RMA for Contact-Rich Meta-Adaptation
-
Breath VOCs causally affect blood glucose levels
Can Breath Biomarkers Causally Influence Blood Glucose? Investigating VOC-Mediated Modulation in Diabetes
-
Subproblem curriculum RL improves LLM math reasoning by 4.1 points
From Reasoning Chains to Verifiable Subproblems: Curriculum Reinforcement Learning Enables Credit Assignment for LLM Reasoning
-
Spline-based warp gives accurate start for sparse 3DGS
TWINGS: Thin Plate Splines Warp-aligned Initialization for Sparse-View Gaussian Splatting
-
Prototype stages top time series accuracy on 80 of 128 UCR datasets
Prototype-Guided Classification Sub-Task Decoupling Framework: Enhancing Generalization and Interpretability for Multivariate Time Series
-
Causal attention lifts time-series classification to 98.6 percent
CASE-NET: Deep Spatio-Temporal Representation Learning via Causal Attention and Channel Recalibration for Multivariate Time Series Classification
-
Graph cuts and Bayesian memory defend RAG from dynamic attacks
RADAR: Defending RAG Dynamically against Retrieval Corruption
-
Reasoning paths in training data lift 3D point cloud models
PointLLM-R: Enhancing 3D Point Cloud Reasoning via Chain-of-Thought
-
Finite networks track mean-field limit uniformly in time
Uniform-in-Time Weak Propagation-of-Chaos in Shallow Neural Networks
-
Five lines of code expose an LLM's hidden vocabulary secrets
Check Your LLM's Secret Dictionary! Five Lines of Code Reveal What Your LLM Learned (Including What It Shouldn't Have)
-
RoBERTa reaches 93 percent accuracy on IMDb sentiment task
From TF-IDF to Transformers: A Comparative and Ensemble Approach to Sentiment Classification
-
The paper identifies that in adversarial distillation
Toward Understanding Adversarial Distillation: Why Robust Teachers Fail
-
Auditable encoder reveals semantic nodes are structurally disconnected
Ex-GraphRAG: Interpretable Evidence Routing for Graph-Augmented LLMs
-
Coupled optimization yields verifiable evidence in rankings
ECPO: Evidence-Coupled Policy Optimization for Evidence-Certified Candidate Ranking
-
RL ties LLM reasoning to verifiable stock forecasts for 25.9% gains
Reasoning through Verifiable Forecast Actions: Consistency-Grounded RL for Financial LLMs
-
Sparsity allocation choice changes label-free repair accuracy
How Sparsity Allocation Shapes Label-Free Post-Pruning Recoverability
-
IAdaPID-ADG optimizer fixes Adam convergence and stability
An Improved Adaptive PID Optimizer with Enhanced Convergence and Stability for Deep Learning
-
Medical world model cuts kidney disease forecast error by 7%
ChronoMedicalWorld: A Medical World Model for Learning Patient Trajectories from Longitudinal Care Data
-
Latent memory mixture lifts continual accuracy by 10 percent
Dynamic Mixture of Latent Memories for Self-Evolving Agents
-
Defense blocks semantic attacks on LLM rankings with perfect precision
SCI-Defense: Defending Manipulation Attacks from Generative Engine Optimization
-
Rényi DP audits reach information-theoretic optimality up to logs
Optimal Guarantees for Auditing R\'enyi Differentially Private Machine Learning
-
Irreversibility equates four measures and picks low-entropy paths
Thermodynamic Irreversibility of Training Algorithms
-
CausalGuard weights candidate graphs for covered causal effect estimates
CausalGuard: Conformal Inference under Graph Uncertainty