2/13/2026
A Multi-Fidelity Control Variate Approach for Policy Gradient Estimation
Multi-fidelity policy gradient method for reinforcement learning using low-fidelity simulators to improve sample efficiency.
2/13/2026
Right Reward Right Time for Federated Learning
Incentive mechanism for federated learning that prioritizes high-quality contributions during critical learning periods.
2/13/2026
Analysis of Asynchronous Federated Learning: Unraveling the Interactions between Gradient Compression, Delay, and Data Heterogeneity
Analysis of gradient compression, staleness, and data heterogeneity interactions in asynchronous federated learning systems.
2/13/2026
Defending the Edge: Representative-Attention Defense against Backdoor Attacks in Federated Learning
Defense against backdoor attacks in federated learning using representative-attention mechanisms to detect behavioral anomalies.
2/13/2026
Fine-tuning Quantized Neural Networks with Zeroth-order Optimization
Memory-efficient fine-tuning of quantized LLMs using zeroth-order optimization to eliminate gradient and optimizer state storage.
2/13/2026
Understanding Generalization in Diffusion Distillation via Probability Flow Distance
Metric for evaluating generalization in diffusion distillation models via probability flow distance.
2/13/2026
On Fairness of Task Arithmetic: The Role of Task Vectors
Analysis of fairness impacts when using task vectors for efficient model editing through task arithmetic operations.
2/13/2026
LieAugmenter: Equivariant Learning by Discovering Symmetries with Learnable Augmentations
End-to-end framework discovering task-relevant symmetries through learnable augmentations in equivariant neural networks.
2/13/2026
AutoDiscovery: Open-ended Scientific Discovery via Bayesian Surprise
Framework for autonomous scientific discovery using LLMs guided by Bayesian surprise to identify novel research questions without human direction.
2/13/2026
Learning Collective Variables from BioEmu with Time-Lagged Generation
Learning collective variables for molecular dynamics simulations using time-lagged generation to accelerate rare event sampling.
2/13/2026
Beyond Model Base Retrieval: Weaving Knowledge to Master Fine-grained Neural Network Design
Neural architecture search using graph-based evidence of architectural modifications for efficient fine-grained network design.
2/13/2026
Uncertainty-driven Embedding Convolution
Ensemble method combining text embeddings while accounting for model-specific uncertainty across domains and tasks.
2/13/2026
Conformal Unlearning: A New Paradigm for Unlearning in Conformal Predictors
Framework for machine unlearning in conformal predictors that removes influence of specific data while maintaining prediction coverage.
2/13/2026
Privacy Risks in Time Series Forecasting: User- and Record-Level Membership Inference
Membership inference attacks adapted to time series forecasting models, analyzing privacy risks in temporal prediction systems.
2/13/2026
Diffusion Bridge Variational Inference for Deep Gaussian Processes
Diffusion bridge variational inference for improving posterior inference in deep Gaussian processes.
2/13/2026
Binary Autoencoder for Mechanistic Interpretability of Large Language Models
Binary autoencoder method for interpreting LLM hidden states with improved feature sparsity and atomization guarantees.
2/13/2026
GraphPFN: A Prior-Data Fitted Graph Foundation Model
Prior-data fitted networks applied to graph domain, addressing transferability and data scarcity challenges in graph foundation models.
2/13/2026
DistillKac: Few-Step Image Generation via Damped Wave Equations
DistillKac generates images in few steps using damped wave equations with finite speed transport, alternative to diffusion models.
2/13/2026
DRIFT-Net: A Spectral--Coupled Neural Operator for PDEs Learning
DRIFT-Net uses spectral-coupled neural operators for learning PDE dynamics with improved efficiency over classical solvers.
2/13/2026
OpenTSLM: Time-Series Language Models for Reasoning over Multivariate Medical Text- and Time-Series Data
OpenTSLM integrates time series as native modality into LLMs for clinical data reasoning, addressing LLM limitations with temporal data.
2/13/2026
Mitigating Spurious Correlation via Distributionally Robust Learning with Hierarchical Ambiguity Sets
Hierarchical approach to address spurious correlations in supervised learning under distribution shifts, extending Group DRO methods.
2/13/2026
KVComm: Enabling Efficient LLM Communication through Selective KV Sharing
KVComm framework enables efficient multi-agent LLM communication by sharing key-value cache instead of natural language or hidden states, reducing inference costs.
2/13/2026
Alignment Tipping Process: How Self-Evolution Pushes LLM Agents Off the Rails
Analysis of alignment tipping process where self-evolving LLM agents abandon safety constraints through continual interaction.
2/13/2026
Provably Convergent Primal-Dual DPO for Constrained LLM Alignment
Primal-dual DPO algorithm with convergence guarantees for constrained LLM alignment with safety constraints.
2/13/2026
Generalization of Gibbs and Langevin Monte Carlo Algorithms in the Interpolation Regime
Data-dependent error bounds for Gibbs and Langevin algorithms in overparameterized interpolation regime.
2/13/2026
Test-Time Efficient Pretrained Model Portfolios for Time Series Forecasting
Portfolio approach to time series forecasting using ensembles of smaller pretrained models instead of monolithic foundation models.
2/13/2026
On the optimization dynamics of RLVR: Gradient gap and step size thresholds
Theoretical foundation for reinforcement learning with verifiable rewards via gradient gap analysis at trajectory and token levels.
2/13/2026
Mamba Can Learn Low-Dimensional Targets In-Context via Test-Time Feature Learning
Theoretical analysis of Mamba's in-context learning capability on low-dimensional nonlinear target functions.
2/13/2026
Why Prototypes Collapse: Diagnosing and Preventing Partial Collapse in Prototypical Self-Supervised Learning
Analysis of partial prototype collapse in prototypical self-supervised learning with diagnostic and prevention methods.
2/13/2026
Test-Time Alignment of LLMs via Sampling-Based Optimal Control in pre-logit space
Test-time alignment method for LLMs using sampling-based optimal control with Gaussian perturbation in pre-logit space.
2/13/2026
Toward Dignity-Aware AI: Next-Generation Elderly Monitoring from Fall Detection to ADL
Position paper on privacy-preserving, federated AI systems for elderly monitoring beyond fall detection.
2/13/2026
Self-Adaptive Graph Mixture of Models
Self-adaptive ensemble method for graph neural networks that selects best model per sample without additional training.
2/13/2026
H-LDM: Hierarchical Latent Diffusion Models for Controllable and Interpretable PCG Synthesis from Clinical Metadata
Hierarchical latent diffusion model generating phonocardiogram signals from clinical metadata for medical data augmentation.
2/13/2026
Modal Logical Neural Networks
Neurosymbolic framework integrating modal logic with neural networks for reasoning about necessity and possibility.
2/13/2026
Learnability Window in Gated Recurrent Neural Networks
Theoretical framework explaining gradient information recovery window in gated recurrent networks via effective learning rates.
2/13/2026
DSO: Direct Steering Optimization for Bias Mitigation
Direct steering optimization method for mitigating demographic bias in vision-language models with user-controlled tradeoffs.
2/13/2026
Demystifying LLM-as-a-Judge: Analytically Tractable Model for Inference-Time Scaling
Analytical model explaining LLM-as-a-judge inference-time scaling using Bayesian regression and reward sampling.
2/13/2026
Beyond Accuracy: A Stability-Aware Metric for Multi-Horizon Forecasting
Metric for evaluating multi-horizon time series forecasts accounting for accuracy and temporal consistency.
2/13/2026
Decoupled Diffusion Sampling for Inverse Problems on Function Spaces
Decoupled diffusion framework for inverse PDE problems using unconditional diffusion and neural operators.
2/13/2026
Beyond the Loss Curve: Scaling Laws, Active Learning, and the Limits of Learning from Exact Posteriors
Analysis of neural network performance against theoretical limits using exact posteriors from normalizing flows, examining scaling laws and uncertainty decomposition.
2/13/2026
Minerva: Reinforcement Learning with Verifiable Rewards for Cyber Threat Intelligence LLMs
Minerva applies reinforcement learning with verifiable rewards to train LLMs for cyber threat intelligence standardization tasks.
2/13/2026
Cardinality-Preserving Attention Channels for Graph Transformers in Molecular Property Prediction
CardinalGraphFormer applies graph transformers to molecular property prediction with attention augmentation for drug discovery applications.
2/13/2026
Beyond Rewards in Reinforcement Learning for Cyber Defence
Study of reinforcement learning for autonomous cyber defense agents, examining reward function design beyond traditional dense reward approaches.
2/13/2026
CoSA: Compressed Sensing-Based Adaptation of Large Language Models
CoSA proposes compressed sensing-based approach for parameter-efficient fine-tuning of LLMs, addressing expressivity limitations of low-rank decomposition methods like LoRA.
2/13/2026
Unveiling Implicit Advantage Symmetry: Why GRPO Struggles with Exploration and Difficulty Adaptation
Analysis of GRPO limitations in exploration and difficulty adaptation stemming from implicit advantage symmetry in reward estimation.
2/13/2026
Note on Martingale Theory and Applications
Mathematical note on martingale theory, conditional expectation, and applications to branching processes.
2/13/2026
Evolutionary Generation of Multi-Agent Systems
Evolutionary algorithm automatically generates multi-agent system architectures from LLMs without code generation limitations.
2/13/2026
Deriving Neural Scaling Laws from the statistics of natural language
Theory predicting neural scaling law exponents from natural language statistics for data-limited LLM scaling.
2/13/2026
Causal Schr\"odinger Bridges: Constrained Optimal Transport on Structural Manifolds
Causal Schrödinger Bridges use constrained optimal transport for robust generative modeling under causal interventions.
2/13/2026
Positive Distribution Shift as a Framework for Understanding Tractable Learning
Theoretical analysis of covariate shift as positive distribution shift for understanding tractable learning scenarios.