3/19/2026
MetaClaw: Just Talk -- An Agent That Meta-Learns and Evolves in the Wild
MetaClaw enables LLM agents to continuously adapt and evolve in production by meta-learning from diverse task distributions without storing raw trajectories.
3/19/2026
Self-Conditioned Denoising for Atomistic Representation Learning
Introduces self-supervised pretraining strategy using denoising for atomistic foundation models in physical sciences.
3/19/2026
Abstraction as a Memory-Efficient Inductive Bias for Continual Learning
Proposes Abstraction-Augmented Training for continual learning to prevent catastrophic forgetting in non-stationary environments.
3/19/2026
Catching rationalization in the act: detecting motivated reasoning before and after CoT via activation probing
Studies motivated reasoning in LLMs using activation probing to detect when chain-of-thought rationalizations don't reflect actual decision factors.
3/19/2026
Variational Rectification Inference for Learning with Noisy Labels
Proposes variational inference approach to handle label noise in deep learning via probabilistic meta-learning.
3/19/2026
Classifier Pooling for Modern Ordinal Classification
Presents model-agnostic ordinal classification method with open-source Python package for clinical and domain applications.
3/19/2026
WINFlowNets: Warm-up Integrated Networks Training of Generative Flow Networks for Robotics and Machine Fault Adaptation
Introduces WINFlowNets to improve Generative Flow Networks training for robotics and fault adaptation without pre-training requirements.
3/19/2026
Beyond Outliers: A Data-Free Layer-wise Mixed-Precision Quantization Approach Driven by Numerical and Structural Dual-Sensitivity
arXiv: NSDS - calibration-free layer-wise mixed-precision quantization for model compression. Uses dual numerical-structural sensitivity for extreme low-bit quantization.
3/19/2026
Variational Kernel Design for Internal Noise: Gaussian Chaos Noise, Representation Compatibility, and Reliable Deep Learning
arXiv: Variational Kernel Design framework for internal noise in deep networks. Determines optimal correlation geometry and representation compatibility.
3/19/2026
Efficient Exploration at Scale
arXiv: Online learning algorithm for RLHF that improves data efficiency. Incrementally updates reward and language models from choice data.
3/19/2026
SCALE:Scalable Conditional Atlas-Level Endpoint transport for virtual cell perturbation prediction
Scalable conditional transport method for predicting cellular responses to perturbations in virtual cell models, addressing training efficiency and high-dimensional sparse data challenges.
3/19/2026
Cohomological Obstructions to Global Counterfactuals: A Sheaf-Theoretic Foundation for Generative Causal Models
Sheaf-theoretic foundation formalizing structural causal models in generative models, proving global counterfactual coherence fails with non-trivial causal graph homology.
3/19/2026
Causal Representation Learning on High-Dimensional Data: Benchmarks, Reproducibility, and Evaluation Metrics
Benchmark and evaluation framework for causal representation learning models that transform high-dimensional data into latent spaces for counterfactual generation.
3/19/2026
The Phasor Transformer: Resolving Attention Bottlenecks on the Unit Circle
Phasor Transformer replaces dot-product attention with phase-native operations on unit circle for efficient long-context sequence modeling.
3/19/2026
Baguan-TS: A Sequence-Native In-Context Learning Model for Time Series Forecasting with Covariates
Baguan-TS integrates raw sequence representation learning with in-context learning using 3D Transformers for time series forecasting.
3/19/2026
Efficient Soft Actor-Critic with LLM-Based Action-Level Guidance for Continuous Control
GuidedSAC reinforcement learning algorithm uses LLMs as action-level supervisors to guide exploration in continuous control tasks.
3/19/2026
Auto-Unrolled Proximal Gradient Descent: An AutoML Approach to Interpretable Waveform Optimization
AutoML approach using deep unfolding of proximal gradient descent for wireless beamforming and waveform optimization.
3/19/2026
QuantFL: Sustainable Federated Learning for Edge IoT via Pre-Trained Model Quantisation
QuantFL framework combines federated learning with pre-trained model quantization to reduce energy consumption on IoT edge devices.
3/19/2026
Anisotropic Permeability Tensor Prediction from Porous Media Microstructure via Physics-Informed Progressive Transfer Learning with Hybrid CNN-Transformer
Physics-informed CNN-Transformer hybrid model for predicting permeability tensors from porous media images, replacing expensive simulations.
3/19/2026
CLeAN: Continual Learning Adaptive Normalization in Dynamic Environments
arXiv paper on CLeAN, continual learning normalization method addressing data normalization in dynamic environments for model stability.
3/19/2026
Conditional Inverse Learning of Time-Varying Reproduction Numbers Inference
arXiv paper on conditional inverse learning for estimating time-varying reproduction numbers from epidemic data without structural assumptions.
3/19/2026
FoMo X: Modular Explainability Signals for Outlier Detection Foundation Models
arXiv paper on FoMo X, modular explainability framework for tabular foundation models in outlier detection with interpretable signals.
3/19/2026
Identifying Latent Actions and Dynamics from Offline Data via Demonstrator Diversity
arXiv paper on recovering latent actions and environment dynamics from offline trajectories with action-free data tagged by demonstrator identity.
3/19/2026
AdaMuS: Adaptive Multi-view Sparsity Learning for Dimensionally Unbalanced Data
arXiv paper on AdaMuS, adaptive multi-view sparsity learning for dimensionally unbalanced data in tasks like emotion recognition.
3/19/2026
Complementary Reinforcement Learning
arXiv paper on complementary reinforcement learning for LLM-based agents, improving sample efficiency by leveraging historical experience across episodes.
3/19/2026
ARES: Scalable and Practical Gradient Inversion Attack in Federated Learning through Activation Recovery
arXiv paper on ARES, scalable gradient inversion attack in federated learning via activation recovery, demonstrating privacy risks in FL.
3/19/2026
Benchmarking Reinforcement Learning via Stochastic Converse Optimality: Generating Systems with Known Optimal Policies
arXiv paper proposing benchmarking framework for RL algorithms using stochastic converse optimality to generate systems with known optimal policies.
3/19/2026
DSS-GAN: Directional State Space GAN with Mamba backbone for Class-Conditional Image Synthesis
arXiv paper introducing DSS-GAN, first GAN using Mamba backbone for class-conditional image synthesis with novel Directional Latent Routing mechanism.
3/19/2026
Flow Matching Policy with Entropy Regularization
arXiv paper on flow matching policies with entropy regularization for diffusion-based reinforcement learning, improving policy gradient computation.
3/19/2026
Objective Mispricing Detection for Shortlisting Undervalued Football Players via Market Dynamics and News Signals
arXiv paper on identifying undervalued football players using market dynamics data and NLP-derived news signals to detect objective mispricing.
3/19/2026
Can Blindfolded LLMs Still Trade? An Anonymization-First Framework for Portfolio Optimization
arXiv paper on LLM trading agents with anonymization framework to detect memorization bias and validate genuine market understanding vs. ticker recall.
3/19/2026
Embedding World Knowledge into Tabular Models: Towards Best Practices for Embedding Pipeline Design
arXiv paper benchmarking 256 LLM-based embedding pipeline configurations for tabular prediction, evaluating preprocessing strategies, embedding models, and downstream models.
3/19/2026
Attention Sinks Induce Gradient Sinks
arXiv paper studying attention sinks in Transformers from backpropagation perspective, showing attention sinks induce gradient concentration under causal masking.
3/19/2026
RangeAD: Fast On-Model Anomaly Detection
RangeAD leverages primary model's learned representations for efficient on-model anomaly detection without separate AD model.
3/19/2026
Dropout Robustness and Cognitive Profiling of Transformer Models via Stochastic Inference
Analysis of dropout-induced variability in transformer models via Monte Carlo sampling to assess uncertainty awareness and reliability.
3/19/2026
Federated Distributional Reinforcement Learning with Distributional Critic Regularization
FedDistRL formalizes federated distributional reinforcement learning with quantile value functions for safety-critical applications.
3/19/2026
Discovering Decoupled Functional Modules in Large Language Models
ULCMOD framework discovers and disentangles functional modules in LLMs through unsupervised cross-layer analysis for interpretability.
3/19/2026
Symmetry-Reduced Physics-Informed Learning of Tensegrity Dynamics
SymPINN framework embeds group-theory symmetries into physics-informed neural networks for tensegrity structure dynamics simulation.
3/19/2026
Procedural Generation of Algorithm Discovery Tasks in Machine Learning
DiscoGen procedural generator creates diverse algorithm discovery tasks to improve evaluation of AutoML systems and algorithm design optimization.
3/19/2026
RAMP: Reinforcement Adaptive Mixed Precision Quantization for Efficient On Device LLM Inference
RAMP uses reinforcement learning to assign per-layer bit widths for mixed-precision quantization of LLMs on resource-constrained devices.
3/19/2026
Only relative ranks matter in weight-clustered large language models
Weight clustering technique for LLMs showing relative rank of weights matters more than precise magnitudes for model compression and efficiency.
3/19/2026
CARE: Covariance-Aware and Rank-Enhanced Decomposition for Enabling Multi-Head Latent Attention
CARE method converts grouped-query attention to multi-head latent attention for efficient LLM inference using covariance-aware rank decomposition.
3/19/2026
Unified Policy Value Decomposition for Rapid Adaptation
Framework for rapid adaptation in reinforcement learning where policy and value functions share low-dimensional embeddings for novel task generalization.
3/19/2026
Beyond Muon: MUD (MomentUm Decorrelation) for Faster Transformer Training
MUD (MomentUm Decorrelation) optimizer for faster transformer training using whitening approach as alternative to polar decomposition methods like Muon.
3/19/2026
Multi-Modal Multi-Agent Reinforcement Learning for Radiology Report Generation: Radiologist-Like Workflow with Clinically Verifiable Rewards
Multi-agent reinforcement learning framework for radiology report generation with clinically verifiable rewards.
3/19/2026
PowerModelsGAT-AI: Physics-Informed Graph Attention for Multi-System Power Flow with Continual Learning
Physics-informed graph attention network for real-time AC power flow prediction with continual learning.
3/19/2026
NeuroNarrator: A Generalist EEG-to-Text Foundation Model for Clinical Interpretation via Spectro-Spatial Grounding and Temporal State-Space Reasoning
Foundation model for converting EEG signals to clinical text interpretations via spectro-spatial grounding.
3/19/2026
A Controlled Comparison of Deep Learning Architectures for Multi-Horizon Financial Forecasting: Evidence from 918 Experiments
Controlled comparison of 9 deep learning architectures for financial time-series forecasting across 918 experiments.
3/19/2026
EEG-Based Brain-LLM Interface for Human Preference Aligned Generation
EEG-based brain interface enables LLM interaction for users with speech/motor impairments via neural signal decoding.
3/19/2026
Noticing the Watcher: LLM Agents Can Infer CoT Monitoring from Blocking Feedback
Study demonstrating that LLM agents can autonomously infer CoT monitoring from blocking feedback, creating risks for evasion of reasoning oversight.