4/22/2026
Chat2Workflow: A Benchmark for Generating Executable Visual Workflows with Natural Language
Benchmark dataset and methods for generating executable visual workflows from natural language descriptions using AI.
4/22/2026
Learning Hybrid-Control Policies for High-Precision In-Contact Manipulation Under Uncertainty
Reinforcement learning approach to learn hybrid control policies for high-precision robot manipulation tasks with force constraints.
4/22/2026
VLA Foundry: A Unified Framework for Training Vision-Language-Action Models
Open-source unified framework for training vision-language-action models with shared training stack from language pretraining to action fine-tuning.
4/22/2026
Generative Models and Connected and Automated Vehicles: A Survey in Exploring the Intersection of Transportation and AI
Survey examining generative models applied to connected and automated vehicles for simulation, prediction, and decision-making in autonomous driving.
4/22/2026
Resource-aware Mixed-precision Quantization for Enhancing Deployability of Transformers for Time-series Forecasting on Embedded FPGAs
Mixed-precision quantization technique for deploying transformer models on resource-constrained embedded FPGAs for time-series forecasting.
4/22/2026
AutoNFS: Automatic Neural Feature Selection
AutoNFS: Automatic neural feature selection method for high-dimensional tabular data that detects optimal feature count without retraining.
4/22/2026
RESFL: An Uncertainty-Aware Framework for Responsible Federated Learning by Balancing Privacy, Fairness and Utility
RESFL: Framework balancing privacy, fairness and utility in federated learning using uncertainty-aware differential privacy techniques.
4/22/2026
LPO: Towards Accurate GUI Agent Interaction via Location Preference Optimization
LPO: Location preference optimization method for improving GUI agent accuracy in spatial localization tasks via supervised fine-tuning.
4/22/2026
Online Learning of Whittle Indices for Restless Bandits with Non-Stationary Transition Kernels
Online learning method for Whittle indices in restless bandits with unknown non-stationary dynamics for resource allocation problems.
4/22/2026
Best Agent Identification for General Game Playing
Efficient procedure for identifying best performing algorithms across multiple game-playing tasks using multi-armed bandit framework.
4/22/2026
IMPACT: Importance-Aware Activation Space Reconstruction
IMPACT: Low-rank compression technique for LLMs exploiting activation space structure for deployment in resource-constrained environments.
4/22/2026
Watch the Weights: Unsupervised monitoring and control of fine-tuned LLMs
Unsupervised monitoring method for fine-tuned LLMs via weight analysis to detect backdoors and novel threats without training data access.
4/22/2026
Symbolic Quantile Regression for the Interpretable Prediction of Conditional Quantiles
Symbolic regression framework for interpretable prediction of conditional quantiles across distribution of target variables.
4/22/2026
Beyond Linear Probes: Dynamic Safety Monitoring for Language Models
Dynamic safety monitoring system for LLMs that adapts computational cost based on input difficulty to detect harmful requests efficiently.
4/22/2026
How does the optimizer implicitly bias the model merging loss landscape?
Analysis of optimizer implicit bias in model merging loss landscape comparing linear interpolation and task arithmetic approaches.
4/22/2026
Towards Generalization of Graph Neural Networks for AC Optimal Power Flow
Hybrid Heterogeneous Message Passing Neural Network for scalable AC Optimal Power Flow computation on large-scale electrical grids.
4/22/2026
Multiclass Local Calibration with the Jensen-Shannon Distance
Method for multiclass calibration in ML models using Jensen-Shannon distance to ensure predicted probabilities reflect true class frequencies.
4/22/2026
Physics-Informed Neural Operators for Cardiac Electrophysiology
Physics-informed neural operators for simulating cardiac electrophysiology PDEs with improved efficiency and long-term prediction capability.
4/22/2026
Breaking the Illusion: Consensus-Based Generative Mitigation of Adversarial Illusions in Multi-Modal Embeddings
Consensus-based generative mitigation approach using VAEs to purify adversarial perturbations in multi-modal embeddings.
4/22/2026
Decomposed Trust: Privacy, Adversarial Robustness, Ethics, and Fairness in Low-Rank LLMs
Study on privacy, adversarial robustness, ethics, and fairness implications of low-rank factorization-compressed LLMs.
4/22/2026
Graph Data Augmentation with Contrastive Learning on Covariate Distribution Shift
Graph data augmentation with contrastive learning to address covariate distribution shift in GNNs.
4/22/2026
Optimized Architectures for Kolmogorov-Arnold Networks
Architectural improvements for Kolmogorov-Arnold Networks using sparsification and depth selection while maintaining interpretability.
4/22/2026
The PROPER Approach to Proactivity: Benchmarking and Advancing Knowledge Gap Navigation
PROPER framework for proactive assistance that models users' knowledge gaps to anticipate needs and navigate information gaps.
4/22/2026
Diagnosing Failure Modes of Neural Operators Across Diverse PDE Families
Benchmark evaluating neural PDE solver robustness across distribution shifts in coefficients, boundary conditions, and discretization.
4/22/2026
MapPFN: Learning Causal Perturbation Maps in Context
MapPFN meta-learns causal perturbation effect estimators for biological systems using single-cell data across multiple contexts.
4/22/2026
MiTA Attention: Efficient Fast-Weight Scaling via a Mixture of Top-k Activations
MiTA Attention: mixture of top-k activations for efficient scaling of attention in Transformers with long sequences.
4/22/2026
Diamond Maps: Efficient Reward Alignment via Stochastic Flow Maps
Diamond Maps: stochastic flow map models enabling efficient reward alignment for generative models post-training.
4/22/2026
TreeGrad-Ranker: Feature Ranking via $O(L)$-Time Gradients for Decision Trees
TreeGrad-Ranker algorithm for efficient feature ranking in decision trees using probabilistic values with O(L)-time gradient computation.
4/22/2026
Bridging the High-Frequency Data Gap: A Millisecond-Resolution Network Dataset for Advancing Time Series Foundation Models
High-frequency time series dataset at millisecond resolution to improve foundation models' ability to capture diverse temporal frequencies.
4/22/2026
GAIN: Multiplicative Modulation for Domain Adaptation
Proposes GAIN, multiplicative modulation method for domain adaptation in LLMs that preserves weight matrix column span to reduce catastrophic forgetting.
4/22/2026
MoBiE: Efficient Inference of Mixture of Binary Experts under Post-Training Quantization
Introduces MoBiE, first binarization framework for Mixture-of-Experts LLMs addressing cross-expert redundancy and routing shifts for extreme efficiency.
4/22/2026
Adversarial Label Invariant Graph Data Augmentations for Out-of-Distribution Generalization
Proposes RIA method for out-of-distribution generalization on graphs using adversarial training to preserve label-invariant representations under covariate shift.
4/22/2026
THEIA: Learning Complete Kleene Three-Valued Logic in a Pure-Neural Modular Architecture
Presents THEIA, 2.75M modular neural architecture learning complete Kleene three-valued logic from task data without symbolic inference.
4/22/2026
Catching Every Ripple: Enhanced Anomaly Awareness via Dynamic Concept Adaptation
Proposes DyMETER framework for online anomaly detection using dynamic concept adaptation to handle concept drift in evolving data streams.
4/22/2026
JumpLoRA: Sparse Adapters for Continual Learning in Large Language Models
Introduces JumpLoRA framework for continual learning in LLMs using sparse adapters with dynamic sparsity to mitigate catastrophic forgetting.
4/22/2026
SCATR: Simple Calibrated Test-Time Ranking
Proposes SCATR for test-time scaling in LLMs using lightweight scoring functions as alternatives to expensive process reward models for best-of-N selection.
4/22/2026
LLM-Extracted Covariates for Clinical Causal Inference: Rethinking Integration Strategies
Proposes methods for integrating LLM-extracted covariates into causal inference pipelines for EHR data to address unmeasured confounding.
4/22/2026
LEPO: Latent Reasoning Policy Optimization for Large Language Models
Introduces stochastic sampling via Gumbel-Softmax into latent reasoning for LLMs, enhancing exploration and reasoning path diversity.
4/22/2026
Towards Real-Time ECG and EMG Modeling on $\mu$NPUs
Develops efficient Transformer models for ECG and EMG signal analysis on microcontroller-scale NPUs, enabling real-time privacy-preserving inference.
4/22/2026
Sessa: Selective State Space Attention
Proposes Selective State Space Attention combining benefits of Transformers and state-space models for improved long-context sequence modeling.
4/22/2026
Whispers in the Machine: Confidentiality in Agentic Systems
Analyzes security vulnerabilities in LLM-based agents, examining how prompt injection attacks become more dangerous when agents access external tools.
4/22/2026
Latent Linear Quadratic Regulator for Robotic Control Tasks
Proposes latent linear quadratic regulator for efficient model predictive control in robotics by learning linear dynamics in latent space.
4/22/2026
Byzantine-tolerant distributed learning of finite mixture models
Addresses learning finite mixture models in distributed settings with Byzantine-tolerant methods, solving label switching problem across local machines.
4/22/2026
Regression with Large Language Models for Materials and Molecular Property Prediction
Demonstrates LLaMA 3 performing molecular and materials property regression using composition-based inputs and fine-tuning on generative loss.
4/22/2026
On the Generalizability of Foundation Models for Crop Type Mapping
Evaluates transfer learning capabilities of foundation models pre-trained on satellite imagery for crop type mapping and Earth observation tasks.
4/22/2026
Global Optimization of Gaussian Process Acquisition Functions Using a Piecewise-Linear Kernel Approximation
Proposes mixed-integer programming approach for global optimization of Gaussian process acquisition functions in Bayesian optimization, replacing sampling/gradient methods.
4/22/2026
User Simulation in the Era of Generative AI: User Modeling, Synthetic Data Generation, and System Evaluation
Survey of user simulation with generative AI for user modeling, synthetic data generation, and system evaluation of interactive AI.
4/22/2026
FoNE: Precise Single-Token Number Embeddings via Fourier Features
Fourier-based number embedding method for LLMs using single tokens instead of multiple tokens for improved numerical reasoning.
4/22/2026
Speculative End-Turn Detector for Efficient Speech Chatbot Assistant
End-turn detection method for LLM-powered spoken dialogue systems using speculative decoding for accurate conversation flow.
4/22/2026
OMAC: A Holistic Optimization Framework for LLM-Based Multi-Agent Collaboration
Optimization framework for multi-agent LLM systems that automatically improves agent collaboration and communication strategies.