3/11/2026
Generative Drifting is Secretly Score Matching: a Spectral and Variational Perspective
Theoretical analysis showing generative drifting for one-step image generation is equivalent to score matching under Gaussian kernels.
3/11/2026
SignalMC-MED: A Multimodal Benchmark for Evaluating Biosignal Foundation Models on Single-Lead ECG and PPG
SignalMC-MED benchmark for evaluating biosignal foundation models on synchronized ECG and PPG data from clinical datasets.
3/11/2026
When Learning Rates Go Wrong: Early Structural Signals in PPO Actor-Critic
Study of learning rate sensitivity in PPO actor-critic methods, analyzing early structural signals to predict training stability.
3/11/2026
Towards a Neural Debugger for Python
Neural debugger for Python that trains LLMs on execution traces to predict line-by-line program execution for debugging workflows.
3/11/2026
On the Width Scaling of Neural Optimizers Under Matrix Operator Norms I: Row/Column Normalization and Hyperparameter Transfer
Analysis of neural network optimizers (AdamW, Muon) as steepest descent under matrix norms, addressing width scaling stability.
3/11/2026
From Data Statistics to Feature Geometry: How Correlations Shape Superposition
Mechanistic interpretability study on how feature correlations affect superposition in neural networks beyond idealized uncorrelated settings.
3/11/2026
Task Aware Modulation Using Representation Learning for Upsaling of Terrestrial Carbon Fluxes
TAM-RL framework uses representation learning to improve upscaling of terrestrial carbon flux measurements for climate modeling.
3/11/2026
Skip to the Good Part: Representation Structure & Inference-Time Layer Skipping in Diffusion vs. Autoregressive LLMs
Layer-wise representational analysis comparing diffusion language models and autoregressive LLMs, examining layer-skipping capabilities.
3/11/2026
Unveiling the Potential of Quantization with MXFP4: Strategies for Quantization Error Reduction
Novel techniques (OAS, MBS) improve MXFP4 quantization accuracy for efficient LLM inference, addressing gaps versus NVIDIA's NVFP4.
3/11/2026
KernelCraft: Benchmarking for Agentic Close-to-Metal Kernel Generation on Emerging Hardware
KernelCraft benchmarks agentic LLM systems for generating low-level kernels for novel AI accelerator instruction set architectures.
3/11/2026
ALADIN: Accuracy-Latency-Aware Design-space Inference Analysis for Embedded AI Accelerators
ALADIN framework for design-space analysis of mixed-precision quantized neural networks on resource-constrained embedded AI accelerators.
3/11/2026
Performance Analysis of Edge and In-Sensor AI Processors: A Comparative Review
Review of ultra-low-power edge AI processors including SoCs, neural accelerators, and in-sensor architectures for embedded inference.
3/11/2026
Data-Rate-Aware High-Speed CNN Inference on FPGAs
Research on dataflow-based CNN accelerators on FPGAs addressing data-rate inefficiencies in layers with reduced output dimensions.
3/11/2026
Memory-Augmented Spiking Networks: Synergistic Integration of Complementary Mechanisms for Neuromorphic Vision
Ablation study integrating Leaky Integrate-and-Fire neurons, contrastive learning, and Hopfield networks in spiking neural networks for neuromorphic vision.
3/11/2026
Hebbian-Oscillatory Co-Learning
Hebbian-Oscillatory Co-Learning (HOC-L) combines structural plasticity and phase synchronization in sparse neural architectures using hyperbolic geometry and Hebbian learning.
3/11/2026
Autonomous Edge-Deployed AI Agents for Electric Vehicle Charging Infrastructure Management
Auralink SDC deploys autonomous edge AI agents for electric vehicle charging infrastructure management, achieving autonomous operation with edge computing latency requirements.
3/11/2026
Sensitivity-Guided Framework for Pruned and Quantized Reservoir Computing Accelerators
Sensitivity-guided compression framework for reservoir computing enabling design-space exploration of quantization, pruning, and hardware efficiency trade-offs.
3/11/2026
The AetherFloat Family: Block-Scale-Free Quad-Radix Floating-Point Architectures for AI Accelerators
AetherFloat family proposes block-scale-free quad-radix floating-point architectures reducing silicon area and power overhead in AI accelerators.
3/11/2026
Permutation-Equivariant 2D State Space Models: Theory and Canonical Architecture for Multivariate Time Series
Permutation-equivariant 2D state space models for multivariate time series, formalizing permutation symmetry principle for exchangeable variables.
3/11/2026
On the Formal Limits of Alignment Verification
Formal analysis proving that no verification procedure can simultaneously satisfy soundness, completeness, and decidability for AI alignment certification.
3/11/2026
Micro-Diffusion Compression -- Binary Tree Tweedie Denoising for Online Probability Estimation
Midicoth introduces micro-diffusion denoising for lossless compression to improve probability estimates in adaptive statistical models like PPM.
3/11/2026
MASEval: Extending Multi-Agent Evaluation from Models to Systems
MASEval extends multi-agent evaluation beyond model-centric benchmarks to evaluate LLM-based agentic system components including topology, orchestration, and error handling.
3/11/2026
APPLV: Adaptive Planner Parameter Learning from Vision-Language-Action Model
APPLV automates parameter tuning for autonomous navigation by learning from vision-language-action models, balancing safety assurances with learning flexibility.
3/11/2026
Why Channel-Centric Models are not Enough to Predict End-to-End Performance in Private 5G: A Measurement Campaign and Case Study
Measurement campaign evaluating throughput predictions in private 5G networks, challenging channel-centric models for communication-aware robot planning.
3/11/2026
FedLECC: Cluster- and Loss-Guided Client Selection for Federated Learning under Non-IID Data
FedLECC proposes cluster and loss-guided client selection for federated learning under non-IID data, improving convergence in distributed AI systems.
3/11/2026
Vision-Language Models Encode Clinical Guidelines for Concept-Based Medical Reasoning
Vision-language models encode clinical guidelines for interpretable medical reasoning in Concept Bottleneck Models, enabling transparent AI in medical imaging.
3/11/2026
Optimizing Reinforcement Learning Training over Digital Twin Enabled Multi-fidelity Networks
Deep learning framework using digital network twin for optimizing reinforcement learning training in multi-fidelity 5G networks with antenna tilt adjustment.
3/11/2026
Interpretable Markov-Based Spatiotemporal Risk Surfaces for Missing-Child Search Planning with Reinforcement Learning and LLM-Based Quality Assurance
Guardian system combines reinforcement learning with LLM-based quality assurance to create spatiotemporal risk surfaces for missing-child search planning from unstructured case documents.
3/11/2026
BiCLIP: Domain Canonicalization via Structured Geometric Transformation
BiCLIP adapts vision-language models to specialized domains via structured geometric transformation, extending canonical transformation theory to domain adaptation.
3/11/2026
Kernel Debiased Plug-in Estimation based on the Universal Least Favorable Submodel
Kernel debiased plug-in estimator (ULFS-KDPE) for nonparametric statistical models achieving semiparametric efficiency without explicit influence function derivation.
3/11/2026
Towards Reliable Simulation-based Inference
Research on using machine learning for statistical inference with scientific simulators, focusing on hypothesis testing and model refinement.
3/11/2026
A Consensus-Driven Multi-LLM Pipeline for Missing-Person Investigations
Guardian system uses multi-LLM pipeline for intelligent information extraction in missing-person investigations, coordinating end-to-end execution across tasks.
3/11/2026
A Survey of Reinforcement Learning For Economics
Survey introducing reinforcement learning methods to economists, addressing curse of dimensionality in complex economic models.
3/11/2026
Statistical Inference via Generative Models: Flow Matching and Causal Inference
Reinterprets generative AI through statistical lens using flow matching, connecting generative models to causal inference and interpretability.
3/11/2026
FlexServe: A Fast and Secure LLM Serving System for Mobile Devices with Flexible Resource Isolation
LLM serving system for mobile devices with hardware-based isolation using ARM TrustZone to protect model weights and user data from kernel attacks.
3/11/2026
From Days to Minutes: An Autonomous AI Agent Achieves Reliable Clinical Triage in Remote Patient Monitoring
Autonomous AI agent for clinical triage in remote patient monitoring, using 21 medical tools to process vitals data 24/7 without physician bottleneck.
3/11/2026
Quality over Quantity: Demonstration Curation via Influence Functions for Data-Centric Robot Learning
Data curation method for robot learning using influence functions to select high-quality demonstrations from noisy human teleoperation data.
3/11/2026
Adaptive Active Learning for Online Reliability Prediction of Satellite Electronics
Online reliability prediction for satellite electronics using degradation models and active learning under limited data.
3/11/2026
Verifying Good Regulator Conditions for Hypergraph Observers: Natural Gradient Learning from Causal Invariance via Established Theorems
Theoretical work on observers in hypergraph physics frameworks applying Conant-Ashby Good Regulator Theorem.
3/11/2026
Latent World Models for Automated Driving: A Unified Taxonomy, Evaluation Framework, and Open Challenges
Taxonomy and evaluation framework for latent world models and vision-language-action systems in autonomous driving.
3/11/2026
RubiCap: Rubric-Guided Reinforcement Learning for Dense Image Captioning
Reinforcement learning approach for dense image captioning using rubric-guided optimization to improve diversity and generalization.
3/11/2026
The Costs of Reproducibility in Music Separation Research: a Replication of Band-Split RNN
Replication study of Band-Split RNN for music source separation, documenting reproducibility challenges and training complexity.
3/11/2026
The Reasoning Trap -- Logical Reasoning as a Mechanistic Pathway to Situational Awareness
Examination of logical reasoning as mechanistic pathway to situational awareness in advanced AI systems, exploring emergent capabilities risks.
3/11/2026
Emotion is Not Just a Label: Latent Emotional Factors in LLM Processing
Study of emotion as latent representational factor in LLM reasoning and text processing, beyond sentiment classification tasks.
3/11/2026
MM-Zero: Self-Evolving Multi-Model Vision Language Models From Zero Data
Vision-language models that self-evolve from zero data without seed images, extending self-improvement paradigms from LLMs to multimodal systems.
3/11/2026
A Generative Sampler for distributions with possible discrete parameter based on Reversibility
Generative sampling framework for complex distributions with discrete parameters, avoiding gradient computation issues.
3/11/2026
On Regret Bounds of Thompson Sampling for Bayesian Optimization
Theoretical analysis of Thompson sampling for Bayesian optimization, comparing regret bounds with GP-UCB methods.
3/11/2026
CLoE: Expert Consistency Learning for Missing Modality Segmentation
Medical image segmentation framework handling missing modalities through consistency learning between expert models.
3/11/2026
Flow Field Reconstruction via Voronoi-Enhanced Physics-Informed Neural Networks with End-to-End Sensor Placement Optimization
Physics-informed neural networks for flow field reconstruction with optimized sensor placement, addressing sparse measurement challenges in fluid dynamics.
3/11/2026
Reviving ConvNeXt for Efficient Convolutional Diffusion Models
Research on convolutional neural network backbones for diffusion models, exploring parameter efficiency and hardware friendliness alternatives to Transformers.