3/16/2026
Denoising Diffusion Variational Inference: Diffusion Models as Expressive Variational Posteriors
Introduces diffusion models as expressive variational posteriors for black-box inference in latent variable models.
3/16/2026
Sampling and Uniqueness Sets in Graphon Signal Processing
Graph signal processing research extending sampling theory to graphon signals using limits of large graphs.
3/16/2026
Adaptive $Q$-Aid for Conditional Supervised Learning in Offline Reinforcement Learning
Research on offline reinforcement learning combining return-conditioned supervised learning with Q-functions to improve stitching capability and stability.
3/16/2026
What Are Good Positional Encodings for Directed Graphs?
Introduces Walk Profile method and explores positional encodings for directed graphs in graph neural networks and graph transformers.
3/16/2026
Rethinking Attention: Polynomial Alternatives to Softmax in Transformers
Explores polynomial attention alternatives to softmax in transformers, arguing regularization rather than probability distribution drives performance.
3/16/2026
Optimistically Optimistic Exploration for Provably Efficient Infinite-Horizon Reinforcement and Imitation Learning
Proposes first computationally efficient algorithm with optimal regret for infinite-horizon discounted reinforcement and imitation learning.
3/16/2026
Causality Is Key to Understand and Balance Multiple Goals in Trustworthy ML and Foundation Models
Advocates integrating causal methods into ML to balance trustworthiness objectives like fairness, privacy, robustness, and explainability.
3/16/2026
Dual Filter: A Transformer-like Inference Architecture for Hidden Markov Models
Proposes Dual Filter framework connecting Hidden Markov Models to transformer decoder architecture for causal nonlinear prediction.
3/16/2026
Guided Policy Optimization under Partial Observability
Introduces Guided Policy Optimization framework for RL in partially observable environments using privileged information from simulators.
3/16/2026
Backward Oversmoothing: why is it hard to train deep Graph Neural Networks?
Analyzes oversmoothing problem in deep Graph Neural Networks and explores why networks fail to learn non-oversmoothed representations.
3/16/2026
Accelerating Residual Reinforcement Learning with Uncertainty Estimation
Proposes uncertainty estimation improvements to Residual Reinforcement Learning for faster adaptation of pretrained policies with sparse rewards.
3/16/2026
Accelerating Diffusion Model Training under Minimal Budgets: A Condensation-Based Perspective
Data condensation approach for training diffusion models with minimal computational budget by constructing smaller synthetic training datasets.
3/16/2026
Invariant Graph Transformer for Out-of-Distribution Generalization
Graph transformer architecture designed for invariant learning to improve out-of-distribution generalization on graph-structured data.
3/16/2026
Intrinsic training dynamics of deep neural networks
Theoretical study of implicit bias in deep neural network training showing gradient flow induces learning of lower-dimensional parameter structures.
3/16/2026
UniPrompt-CL: Sustainable Continual Learning in Medical AI with Unified Prompt Pools
Continual learning framework with unified prompt pools for medical imaging tasks, addressing domain-specific challenges in adaptive AI.
3/16/2026
Local Mechanisms of Compositional Generalization in Conditional Diffusion
Analysis of compositional generalization mechanisms in conditional diffusion models, studying length generalization on controlled image generation tasks.
3/16/2026
LiLAW: Lightweight Learnable Adaptive Weighting to Meta-Learn Sample Difficulty, Improve Noisy Training, Increase Fairness, and Effectively Use Synthetic Data
Lightweight meta-learning method using three parameters to dynamically adjust sample loss weights for noisy training, fairness, and synthetic data utilization.
3/16/2026
PreLoRA: Hybrid Pre-training of Vision Transformers with Full Training and Low-Rank Adapters
Hybrid pre-training approach using low-rank adapters alongside full training to reduce computational cost for vision transformer training.
3/16/2026
Robust Fine-Tuning from Non-Robust Pretrained Models: Mitigating Suboptimal Transfer With Epsilon-Scheduling
Method for robust fine-tuning non-robust pretrained models using epsilon-scheduling to achieve adversarial robustness and task adaptation simultaneously.
3/16/2026
Disentangling Recall and Reasoning in Transformer Models through Layer-wise Attention and Activation Analysis
Analysis of transformer internals distinguishing recall from reasoning mechanisms through layer-wise attention and activation patterns for interpretability.
3/16/2026
Language Models are Injective and Hence Invertible
Mathematical proof that transformer language models are injective, enabling exact input recovery from representations despite nonlinear components.
3/16/2026
LLM Unlearning with LLM Beliefs
Method for unlearning harmful content from LLMs by analyzing belief redistribution in probability space, avoiding unwanted side effects of gradient ascent.
3/16/2026
Larger Datasets Can Be Repeated More: A Theoretical Analysis of Multi-Epoch Scaling in Linear Regression
Theoretical analysis of data scaling laws in linear regression when training multiple epochs on limited datasets, relevant to LLM training efficiency.
3/16/2026
Information-Consistent Language Model Recommendations through Group Relative Policy Optimization
Method to improve LLM consistency and reliability across semantically equivalent prompts using group relative policy optimization for business-critical applications.
3/16/2026
Epistemic diversity across language models mitigates knowledge collapse
Study demonstrating that ensemble diversity across language models mitigates knowledge collapse from training on model-generated outputs.
3/16/2026
MoHETS: Long-term Time Series Forecasting with Mixture-of-Heterogeneous-Experts
Mixture-of-experts approach with heterogeneous experts for capturing multi-scale temporal dynamics in long-horizon time series forecasting.
3/16/2026
Learnable Koopman-Enhanced Transformer-Based Time Series Forecasting with Spectral Control
Integration of Koopman operator theory with transformer architectures for time series forecasting with learnable spectral parameterizations.
3/16/2026
CORE: Context-Robust Remasking for Diffusion Language Models
Decoding strategy for masked diffusion language models that dynamically adjusts token retention based on context coverage.
3/16/2026
Hierarchical Concept Embedding & Pursuit for Interpretable Image Classification
Interpretable image classification using hierarchical concept embeddings recovered from vision-language model latent spaces.
3/16/2026
RooflineBench: A Benchmarking Framework for On-Device LLMs via Roofline Analysis
Benchmarking framework using roofline analysis to characterize performance of small language models on resource-constrained edge hardware.
3/16/2026
On the Geometric Coherence of Global Aggregation in Federated Graph Neural Networks
Federated learning approach addressing heterogeneous graph structures in distributed GNN training across multiple clients.
3/16/2026
Test-Time Adaptation via Many-Shot Prompting: Benefits, Limits, and Pitfalls
Study of many-shot in-context learning as test-time adaptation for LLMs, analyzing benefits and reliability limits with open-source models.
3/16/2026
Distributional Regression with Tabular Foundation Models: Evaluating Probabilistic Predictions via Proper Scoring Rules
Evaluation framework using proper scoring rules for assessing distributional predictions from tabular foundation models beyond point estimates.
3/16/2026
What do near-optimal learning rate schedules look like?
Search procedure to identify optimal learning rate schedule shapes for neural network training across different workloads.
3/16/2026
H2LooP Spark Preview: Continual Pretraining of Large Language Models for Low-Level Embedded Systems Code
Continual pretraining of LLMs specialized for low-level embedded systems code generation, targeting underrepresented hardware domains.
3/16/2026
Why Softmax Attention Outperforms Linear Attention
Comparative analysis of softmax vs linear attention mechanisms in transformer architectures, examining computational efficiency tradeoffs.
3/16/2026
Latent diffusion models for parameterization and data assimilation of facies-based geomodels
Latent diffusion models for geological parameterization and data assimilation, generating realistic geomodels with reduced variables for history matching.
3/16/2026
Superficial Safety Alignment Hypothesis
Analyzes brittleness of LLM safety alignment mechanisms, proposing superficial safety alignment hypothesis explaining why standard alignment approaches are vulnerable.
3/16/2026
Active Causal Structure Learning with Latent Variables: Towards Learning to Detour in Autonomous Robots
Active causal structure learning framework enabling autonomous robots and AGI agents to dynamically construct causal models of environmental interactions.
3/16/2026
Mask-Enhanced Autoregressive Prediction: Pay Less Attention to Learn More
Training paradigm integrating masked language modeling with next-token prediction to improve in-context retrieval in large language models.
3/16/2026
Understanding Dataset Distillation via Spectral Filtering
Spectral filtering framework unifying dataset distillation methods by interpreting them as filters affecting feature correlation eigenvalues.
3/16/2026
Minimax learning rates for estimating binary classifiers under margin conditions
Theoretical analysis of minimax learning rates for binary classification under geometric margin conditions with horizon function decision boundaries.
3/16/2026
AdaBoN: Adaptive Best-of-N Alignment
Prompt-adaptive Best-of-N alignment strategy using reward models to reduce computational cost of test-time alignment for language models.
3/16/2026
Integration of TinyML and LargeML: A Survey of 6G and Beyond
Survey on integrating TinyML and LargeML for 6G networks, covering deep learning applications in mobile systems, autonomous vehicles, and smart services.
3/16/2026
Token Distillation: Attention-aware Input Embeddings For New Tokens
Attention-aware embedding initialization method for new tokens in LLMs without expensive retraining, addressing vocabulary limitations in specialized domains.
3/16/2026
Towards Reliable Detection of Empty Space: Conditional Marked Point Processes for Object Detection
Conditional marked point processes for reliable object detection uncertainty quantification, addressing miscalibrated confidence scores in neural networks.
3/16/2026
From Video to EEG: Adapting Joint Embedding Predictive Architecture to Uncover Saptiotemporal Dynamics in Brain Signal Analysis
Self-supervised learning approach adapting joint embedding architecture from video to EEG signals for brain activity analysis with limited labeled data.
3/16/2026
Quantum-Informed Machine Learning for Predicting Spatiotemporal Chaos with Practical Quantum Advantage
Quantum-informed ML framework combining quantum generative models with classical predictors for long-term spatiotemporal chaos prediction.
3/16/2026
Aligning Large Language Model Agents with Rational and Moral Preferences: A Supervised Fine-Tuning Approach
Supervised fine-tuning method to align LLM agents with rational and moral preferences in strategic economic games, addressing systematic behavioral deviations.
3/16/2026
SegDAC: Visual Generalization in Reinforcement Learning via Dynamic Object Tokens
Object-centric representations for visual RL policies using dynamic tokens to improve generalization under visual condition changes without fixed-size slots.