4/6/2026
ESL-Bench: An Event-Driven Synthetic Longitudinal Benchmark for Health Agents
Event-driven synthetic benchmark for longitudinal health agents reasoning over multi-source trajectories including device streams and clinical data.
4/6/2026
EMS: Multi-Agent Voting via Efficient Majority-then-Stopping
Efficient majority voting method for multi-agent systems that stops early once consensus achieved, reducing computational overhead through agent scheduling.
4/6/2026
Multi-Turn Reinforcement Learning for Tool-Calling Agents with Iterative Reward Calibration
Applies MT-GRPO and GTPO reinforcement learning for training tool-calling agents on multi-turn customer service tasks with sparse reward credit assignment.
4/6/2026
Analysis of Optimality of Large Language Models on Planning Problems
Analyzes frontier LLMs on classic AI planning problems, examining whether models reason optimally or rely on heuristic strategies in Blocksworld domain.
4/6/2026
AgentHazard: A Benchmark for Evaluating Harmful Behavior in Computer-Use Agents
Benchmark for evaluating harmful behavior in computer-use agents, testing safety risks from sequences of individually plausible but collectively harmful actions.
4/6/2026
FoE: Forest of Errors Makes the First Solution the Best in Large Reasoning Models
Analysis of reasoning failures in large reasoning models, showing first solution often optimal despite test-time scaling patterns in DeepSeek-R1.
4/6/2026
InfoSeeker: A Scalable Hierarchical Parallel Agent Framework for Web Information Seeking
Scalable hierarchical parallel agent framework for web information seeking, addressing wide-scale evidence synthesis and context saturation in LLM agents.
4/6/2026
Agentic-MME: What Agentic Capability Really Brings to Multimodal Intelligence?
Benchmark evaluating multimodal LLM agents with tool integration capabilities including visual expansion and web search through agentic reasoning.
4/6/2026
Automatic Textbook Formalization
AI system automatically formalizes 500+ page graduate-level algebraic combinatorics textbook to Lean, achieving 130K lines of formal code.
4/6/2026
Chart-RL: Policy Optimization Reinforcement Learning for Enhanced Visual Reasoning in Chart Question Answering with Vision Language Models
Reinforcement learning approach to improve visual reasoning in chart question answering using vision language models with policy optimization.
4/6/2026
Coupled Control, Structured Memory, and Verifiable Action in Agentic AI (SCRAT -- Stochastic Control with Retrieval and Auditable Trajectories): A Comparative Perspective from Squirrel Locomotion and Scatter-Hoarding
Framework for agentic AI emphasizing control, memory, and verifiable action under partial observability, inspired by squirrel ecology comparisons.
4/6/2026
Linguistic Frameworks Go Toe-to-Toe at Neuro-Symbolic Language Modeling
Evaluates linguistic graph representations combined with pretrained Transformers for language modeling, comparing semantic and syntactic formalisms.
4/6/2026
Reanalyzing L2 Preposition Learning with Bayesian Mixed Effects and a Pretrained Language Model
Bayesian and neural models analyzing Chinese learners' English preposition comprehension, using pretrained language models for linguistic analysis.
4/6/2026
Empirical Sufficiency Lower Bounds for Language Modeling with Locally-Bootstrapped Semantic Structures
Research on language modeling with predicted semantic structure, establishing empirical lower bounds for performance improvements using binary vector representations.
4/6/2026
LLM Reasoning with Process Rewards for Outcome-Guided Steps
Reinforcement learning approach using process rewards to provide intermediate feedback for multi-step mathematical reasoning in LLMs.
4/6/2026
Haiku to Opus in Just 10 bits: LLMs Unlock Massive Compression Gains
Study of LLM-generated text compression using domain-adapted LoRA and arithmetic coding, characterizing lossless and lossy compression frontiers.
4/6/2026
UI-Oceanus: Scaling GUI Agents with Synthetic Environmental Dynamics
Framework for scaling GUI agents using synthetic environmental dynamics and self-supervised learning from ground-truth interaction feedback.
4/6/2026
DrugPlayGround: Benchmarking Large Language Models and Embeddings for Drug Discovery
Benchmark for evaluating LLMs and embeddings on drug discovery tasks including hypothesis generation and candidate prioritization.
4/6/2026
OPRIDE: Offline Preference-based Reinforcement Learning via In-Dataset Exploration
Offline preference-based RL method improving query efficiency by addressing exploration and preference ranking within existing datasets.
4/6/2026
Differentiable Symbolic Planning: A Neural Architecture for Constraint Reasoning with Learned Feasibility
Neural architecture performing discrete symbolic constraint reasoning while maintaining differentiability for planning and feasibility checking.
4/6/2026
An Initial Exploration of Contrastive Prompt Tuning to Generate Energy-Efficient Code
Study using contrastive prompt tuning to optimize LLMs for generating energy-efficient code supporting Green Software Development.
4/6/2026
Prism: Policy Reuse via Interpretable Strategy Mapping in Reinforcement Learning
Framework for zero-shot transfer between RL agents using interpretable discrete concepts validated through causal intervention.
4/6/2026
Using LLM-as-a-Judge/Jury to Advance Scalable, Clinically-Validated Safety Evaluations of Model Responses to Users Demonstrating Psychosis
Framework using LLMs as judges to evaluate safety of model responses for users with psychosis, addressing clinical validation gaps in mental health.
4/6/2026
Beyond Message Passing: Toward Semantically Aligned Agent Communication
Analysis of agent communication protocols for LLM systems organized into communication, syntactic, and semantic layers with systematic evaluation of 18 protocols.
4/6/2026
A Survey on AI for 6G: Challenges and Opportunities
Survey of AI and ML applications in 6G networks covering high data rates, low latency, and emerging applications like autonomous systems.
4/6/2026
Internalized Reasoning for Long-Context Visual Document Understanding
Synthetic data pipeline for reasoning in long-document visual understanding that generates thinking traces for improved LLM performance on enterprise documents.
4/6/2026
Ambig-IaC: Multi-level Disambiguation for Interactive Cloud Infrastructure-as-Code Synthesis
Framework addressing underspecified natural language requests for cloud infrastructure code generation using LLMs with multi-level disambiguation.
4/6/2026
Audio Spatially-Guided Fusion for Audio-Visual Navigation
Audio-visual navigation system for autonomous agents to localize and navigate toward vocalizing targets in 3D environments.
4/6/2026
Improving MPI Error Detection and Repair with Large Language Models and Bug References
Approach using LLMs to detect and repair errors in MPI code for high-performance computing and distributed training frameworks.
4/6/2026
LumiVideo: An Intelligent Agentic System for Video Color Grading
LumiVideo agentic system mimicking professional video colorists' workflows with interpretable iterative control for automated color grading.
4/6/2026
Generative models on phase space
Research on deep generative models (diffusion, flow matching) for high-dimensional distributions on constrained submanifolds in physics data.
4/6/2026
Self-Directed Task Identification
Self-Directed Task Identification framework enabling models to autonomously identify target variables in zero-shot learning without pre-training.
4/6/2026
PlayGen-MoG: Framework for Diverse Multi-Agent Play Generation via Mixture-of-Gaussians Trajectory Prediction
Framework using Mixture-of-Gaussians trajectory prediction for diverse multi-agent play generation in team sports.
4/6/2026
Do We Need Frontier Models to Verify Mathematical Proofs?
Research investigating whether frontier reasoning models are necessary for mathematical proof verification versus smaller LLM judges.
4/6/2026
Skeleton-based Coherence Modeling in Narratives
NLP research on skeleton-based coherence modeling for narrative generation and detection of incoherent story structures.
4/6/2026
When simulations look right but causal effects go wrong: Large language models as behavioral simulators
Empirical evaluation of LLMs as behavioral simulators for predicting intervention effects across 11 climate-psychology interventions using 59,508 participants.
4/6/2026
On the Geometric Structure of Layer Updates in Deep Language Models
Research studying geometric structure of layer-wise updates in deep language models across Transformer and state-space architectures.
4/6/2026
Hierarchical, Interpretable, Label-Free Concept Bottleneck Model
Hierarchical Interpretable Label-Free Concept Bottleneck Model enabling interpretability at multiple abstraction levels unlike single-level existing CBMs.
4/6/2026
Generating Satellite Imagery Data for Wildfire Detection through Mask-Conditioned Generative AI
Diffusion-based foundation model generates synthetic satellite imagery for wildfire detection without task-specific retraining.
4/6/2026
Automated Malware Family Classification using Weighted Hierarchical Ensembles of Large Language Models
Zero-shot malware family classification using weighted hierarchical ensembles of LLMs, avoiding need for labeled datasets and handcrafted features.
4/6/2026
Token-Efficient Multimodal Reasoning via Image Prompt Packaging
Image Prompt Packaging method to reduce token costs in multimodal LLMs by embedding structured text into images, benchmarked across frontier models.
4/6/2026
Social Meaning in Large Language Models: Structure, Magnitude, and Pragmatic Prompting
Study of social meaning in LLMs, introducing calibration metrics and pragmatic prompting strategies to improve quantitative approximation of human reasoning.
4/6/2026
Opal: Private Memory for Personal AI
System for private long-term memory in personal AI using trusted hardware and oblivious RAM to hide data access patterns from providers.
4/6/2026
Jump Start or False Start? A Theoretical and Empirical Evaluation of LLM-initialized Bandits
Theoretical and empirical evaluation of using LLM-generated preferences to warm-start contextual bandits, examining alignment with actual user preferences.
4/6/2026
Feature Attribution Stability Suite: How Stable Are Post-Hoc Attributions?
Analysis of stability in post-hoc feature attribution methods for vision systems under input perturbations, introducing evaluation suite.
4/6/2026
From Theory to Practice: Code Generation Using LLMs for CAPEC and CWE Frameworks
LLM-based code generation for security vulnerabilities using CAPEC and CWE frameworks, addressing gaps in existing vulnerability datasets.
4/6/2026
Pragmatics Meets Culture: Culturally-adapted Artwork Description Generation and Evaluation
Study of cultural bias in LLM text generation, introducing task of culturally-adapted artwork descriptions for different audience groups.
4/6/2026
Generative AI Use in Entrepreneurship: An Integrative Review and an Empowerment-Entrapment Framework
Integrative review of generative AI impact on entrepreneurship across opportunity recognition, evaluation, resource assembly, and venture launch stages.
4/6/2026
Understanding the Effects of Safety Unalignment on Large Language Models
Research on safety alignment vulnerabilities in LLMs, examining jailbreak-tuning and weight orthogonalization methods that can disable safety guardrails.
4/6/2026
High Volatility and Action Bias Distinguish LLMs from Humans in Group Coordination
Comparative study of LLM vs human coordination in group games, revealing volatility and action bias differences in adaptive strategies.