Ax Emmanuel A. Olowe, Danial Chitnis 5/1/2026

Policy-Governed LLM Routing with Intent Matching for Instrument Laboratories

Policy-governed LLM routing system (Routiium) for engineering lab assistance with instructor control over timing and cost.

Ax Hyesun Choung, Soojong Kim 5/1/2026

Can AI be a moral victim? The role of moral patiency and ownership perceptions in ethical judgments of using AI-generated content

Empirical study on moral perception of AI-generated content reuse examining authorship and plagiarism concerns.

Ax Arne Bewersdorff, Nejla Yuruk, Xiaoming Zhai 5/1/2026

Simulating Validity: Modal Decoupling in MLLM Generated Feedback on Science Drawings

Studies validity of multimodal LLM feedback on student hand-drawn scientific diagrams and visual model encoding.

Ax Ji Yeon Kim 5/1/2026

Designing Ethical Learning for Agentic AI: Toegye Yi Hwang's Ethical Emotion Regulation Framework

Proposes ethical emotion regulation framework for agentic AI inspired by 17th-century philosophy for autonomous learning environments.

Ax Elham Nasarian, Abhilash Neog, Kwok-Leung Tsui, Niyousha HosseiniChimeh 5/1/2026

CareGuardAI: Context-Aware Multi-Agent Guardrails for Clinical Safety & Hallucination Mitigation in Patient-Facing LLMs

Multi-agent guardrails system for clinical safety and hallucination mitigation in patient-facing LLM healthcare applications.

Ax Jinhui Han, Ming Hu, Xilin Zhang 5/1/2026

LLM Biases

Study of systematic biases in transformer-based agentic AI systems deployed for user-facing applications like shopping and content navigation.

Ax Pengfei He (Peter), Shaowei Wang (Peter), Tse-Hsun (Peter), Chen, Muhammad Asaduzzaman 5/1/2026

Static Program Slicing Using Language Models With Dataflow-Aware Pretraining and Constrained Decoding

Language models with dataflow-aware pretraining for static program slicing, addressing hallucination and dependency modeling in code analysis.

Ax Bingxi Zhao, Jiahao Zhang, Xubin Ren, Zirui Guo, Tianzhe Chu, Yi Ma, Chao Huang 5/1/2026

DeepTutor: Towards Agentic Personalized Tutoring

DeepTutor: Open-source agentic framework for personalized tutoring using LLMs with adaptive feedback beyond RAG systems.

Ax Mary Nusrat, Sarfuddin Bhuiyan, Gahangir Hossain 5/1/2026

Learning-to-Explain through 20Q Gaming: An Explainable Recommender for Cybersecurity Education

Educational framework using explainable AI and 20 Questions game format for adaptive cybersecurity training with interactive learning.

Ax Jonas Dolezal, Sawood Alam, Mark Graham, Maty Bohacek 5/1/2026

The Impact of AI-Generated Text on the Internet

Analyzes prevalence and impact of AI-generated and AI-edited text on the internet using representative sampling methodology.

Ax Sanjeev Rao Ganjihal 5/1/2026

Predictive Multi-Tier Memory Management for KV Cache in Large-Scale GPU Inference

Addresses KV cache memory bottleneck in large-scale GPU inference serving through predictive multi-tier memory management and unified cache sizing.

Ax Xidong Wu, Yue Zhuan, Ruoqiao Wei, Hangxin Chen, Di Bai, Jintao Liu, Xinyi Wang, Xue Wang, Luoshu Wang, Xinwu Cheng 5/1/2026

AgenticRecTune: Multi-Agent with Self-Evolving Skillhub for Recommendation System Optimization

Multi-agent system with self-evolving skill hub for optimizing large-scale recommendation system pipelines across pre-ranking, ranking, and re-ranking stages.

Ax Mandar Karhade 5/1/2026

Not All Memories Age the Same: Autodiscovery of Adaptive Decay in Knowledge Graphs

Proposes hierarchical framework for knowledge graphs that autodiscovers adaptive decay rates for different knowledge types instead of uniform forgetting curves.

Ax Michael Cardei, Huu Binh Ta, Ferdinando Fioretto 5/1/2026

Simple Self-Conditioning Adaptation for Masked Diffusion Models

Proposes self-conditioning adaptation for masked diffusion models to improve cross-step refinement in discrete sequence generation through iterative denoising.

Ax Akshay Mittal, Elyson De La Cruz 5/1/2026

Agent Name Service (ANS): A Proof-of-Concept Trust Layer for Secure AI Agent Discovery, Identity, and Governance in Kubernetes

Agent Name Service: DNS-inspired trust layer for Kubernetes providing secure discovery, cryptographic authentication, capability attestation, and policy governance for AI agents.

Ax Qisheng Hu, Quanyu Long, Wenya Wang 5/1/2026

When Continual Learning Moves to Memory: A Study of Experience Reuse in LLM Agents

Analysis of continual learning challenges in memory-augmented LLM agents, showing stability-plasticity dilemma resurfaces at memory retrieval level under context limitations.

Ax Gilberto Sussumu Hida, Danilo Monteiro Ribeiro, Erika Yahata 5/1/2026

Beyond Accuracy: LLM Variability in Evidence Screening for Software Engineering SLRs

Empirical study of LLM performance variability and consistency in study screening for systematic literature reviews compared to classical methods.

Ax Alessia Berarducci, Eric Rossetto, Alessandro Antonucci, Marco Zaffalon 5/1/2026

Automatic Causal Fairness Analysis with LLM-Generated Reporting

FairMind: AutoML framework automating fairness analysis at dataset level using causal assumptions and LLM-generated reporting.

Ax Karthik Charan Raghunathan, Christian Metzner, Laura Kriener, Melika Payvand 5/1/2026

NORACL: Neurogenesis for Oracle-free Resource-Adaptive Continual Learning

Neurogenesis-based approach for continual learning that dynamically expands network capacity without oracle task information.

Ax Samuel L Pugh, Eric Yang, Alexander Muir Sutherland, Alessandra Breschi 5/1/2026

Detecting Clinical Discrepancies in Health Coaching Agents: A Dual-Stream Memory and Reconciliation Architecture

Dual-stream memory architecture for persistent LLM health coaching agents reconciling patient self-report with electronic health records.

Ax Boris Shigida, Boris Hanin, Andrey Gromov 5/1/2026

Learning Rate Transfer in Normalized Transformers

Study addressing learning rate transfer in normalized transformers (nGPT) across model dimensions using alignment exponents.

Ax Yibin Luo, Shiwei Gao, Huichuan Zheng, Youyou Lu, Jiwu Shu 5/1/2026

Efficient Training on Multiple Consumer GPUs with RoundPipe

RoundPipe: pipeline parallelism technique with CPU offloading for efficient LLM fine-tuning on consumer-grade GPUs with limited memory.

Ax Mingqian Zheng, Malia Morgan, Liwei Jiang, Carolyn Rose, Maarten Sap 5/1/2026

Useless but Safe? Benchmarking Utility Recovery with User Intent Clarification in Multi-Turn Conversations

CarryOnBench interactive benchmark measuring LLM ability to recover helpfulness through user intent clarification in multi-turn conversations while maintaining safety.

Ax Yassine Ibork, Nhat Ha Nguyen, Myounggyu Won, Lokesh Das 5/1/2026

PALCAS: A Priority-Aware Intelligent Lane Change Advisory System for Autonomous Vehicles using Federated Reinforcement Learning

Multi-agent federated reinforcement learning system for autonomous vehicle lane change advisory with priority-aware decision making.

Ax Haiyu Yang, Miel Hostens 5/1/2026

Lightweight Distillation of SAM 3 and DINOv3 for Edge-Deployable Individual-Level Livestock Monitoring and Longitudinal Visual Analytics

Knowledge distillation approach reducing SAM 3 and DINOv3 foundation models from 446M to 40.66M parameters for edge deployment in livestock monitoring.

Ax Benjamin Probst, Andreas Happe, J\"urgen Cito 5/1/2026

Enhancing Linux Privilege Escalation Attack Capabilities of Local LLM Agents

Research on improving Linux privilege escalation capabilities of locally-hosted open-weight LLM agents for autonomous penetration testing.

Ax Jerry Y. Huang, Justin Lin, Sheel Shah, Kartik Nair, Nicholas M. Boffi 5/1/2026

How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance

Reformulation of guidance in generative modeling as optimal control problem with efficient algorithms for few-step reward-guided sampling.

Ax Fatima Rabia Yapicioglu, Meltem Aksoy, Alberto Rigenti, Tuwe L\"ofstr\"om-Cavallin, Helena L\"ofstr\"om-Cavallin, Seyda Yoncaci, Luca Longo 5/1/2026

ConformaDecompose: Explaining Uncertainty via Calibration Localization

Method for explaining uncertainty in conformal prediction by localizing calibration sources to distinguish aleatoric and epistemic uncertainty at instance level.

Ax Paraskevas V. Lekeas, Giorgos Stamatopoulos 5/1/2026

What Suppresses Nash Equilibrium Play in Large Language Models? Mechanistic Evidence and Causal Control

Mechanistic study of why LLM agents deviate from Nash equilibria in game theory tasks, using interpretability methods on Llama-3 and Qwen2.5 models to identify and reverse deviation causes.

Ax Shouren Wang, Wang Yang, Chuang Ma, Debargha Ganguly, Vikash Singh, Chaoda Song, Xinpeng Li, Xianxuan Long, Vipin Chaudhary, Xiaotian Han 5/1/2026

Path-Lock Expert: Separating Reasoning Mode in Hybrid Thinking via Architecture-Level Separation

Architectural approach separating explicit thinking and no-thinking modes in hybrid reasoning LLMs through Path-Lock Expert mechanism to reduce reasoning leakage.

Ax Halley Young, Nikolaj Bj\"orner 5/1/2026

Theory Under Construction: Orchestrating Language Models for Research Software Where the Specification Evolves

Research on using LLMs for research software development where specifications evolve, identifying hallucination accumulation and unsupported assertion propagation as key failure modes.

Ax Kashif Imteyaz, Isabel Lopez, Nakul Rajpal, Hunjun Shin, Saiph Savage 5/1/2026

Upskilling with Generative AI: Practices and Challenges for Freelance Knowledge Workers

Study on how freelance knowledge workers use generative AI tools for skill acquisition and the challenges they face in rapidly evolving skill demands.

Ax Jason Fournier (Imagine Learning), Kacper {\L}odzikowski (Adam Mickiewicz University, Pozna\'n, Poland) 5/1/2026

Addressing the Reality Gap: A Three-Tension Framework for Agentic AI Adoption

Framework addressing adoption challenges for agentic AI systems in education, balancing implementation feasibility, personalization benefits, and disruption risks.

Ax Jon-Paul Cacioli 5/1/2026

Instruction Complexity Induces Positional Collapse in Adversarial LLM Evaluation

Research on adversarial instruction evaluation showing how LLMs fall back on positional shortcuts under high instruction complexity rather than engaging with question content.

Ax Xingwei Tan, Marco Valentino, Mahmud Elahi Akhter, Yuxiang Zhou, Maria Liakata, Nikolaos Aletras 5/1/2026

Compliance versus Sensibility: On the Reasoning Controllability in Large Language Models

Study on reasoning controllability in LLMs examining whether fundamental reasoning patterns like induction, deduction, and abduction can be decoupled from specific problem instances.

Ax Marco Robol, Paolo Giorgini 5/1/2026

Self-Evolving Software Agents

Paper introducing self-evolving software agents combining BDI reasoning with LLMs to enable autonomous evolution of goals, reasoning, and executable code.

Ax Neha Nagaraja, Hayretdin Bahsi, Carlo R. da Cunha 5/1/2026

From Prompt to Physical Actuation: Holistic Threat Modeling of LLM-Enabled Robotic Systems

Research on threat modeling for LLM-integrated robotic systems, analyzing how compromised inputs and unsafe outputs propagate through planning pipelines to physical consequences.

Ax Chung-Hsiang Lo, Lu Li, Diji Yang, Tianyu Zhang, Yunkai Zhang, Yoshua Bengio, Yi Zhang 5/1/2026

When 2D Tasks Meet 1D Serialization: On Serialization Friction in Structured Tasks

Studies serialization friction when LLMs process 2D structured tasks as 1D token sequences, impacting performance.

Ax Matthew Christian Agustin 5/1/2026

Evaluating Epistemic Guardrails in AI Reading Assistants: A Behavioral Audit of a Minimal Prototype

Examines epistemic guardrails in LLM reading assistants to prevent interpretive displacement and unsafe meaning-making.

Ax Mehmet Iscan 5/1/2026

Learning When to Remember: Risk-Sensitive Contextual Bandits for Abstention-Aware Memory Retrieval in LLM-Based Coding Agents

Risk-sensitive contextual bandit approach for memory retrieval in LLM-based coding agents with safety constraints.

Ax Raviteja Anantha, Nick Levato, Layne C. Price 5/1/2026

BoostLoRA: Growing Effective Rank by Boosting Adapters

BoostLoRA gradient-boosting framework for parameter-efficient fine-tuning that grows adapter effective rank iteratively.

Ax Pedro-Aar\'on Hern\'andez-\'Avalos, Luciano Garc\'ia-Ba\~nuelos 5/1/2026

Pragmos: A Process Agentic Modeling System

Pragmos system uses LLMs and agents for automated business process model derivation from textual descriptions.

Ax Md Hasan Saju, Akramul Azim 5/1/2026

Toward Autonomous SOC Operations: End-to-End LLM Framework for Threat Detection, Query Generation, and Resolution in Security Operations

End-to-end LLM framework for automated security operations combining threat detection, query generation, and resolution.

Ax Vanessa B. Sibug, Emerson Q. Fernando, Almer B. Gamboa, Roque Francis B. Dianelo, Agnes R. Regala, Joseph Alexander Bansil, Jan Henry B. Sunga, Vernon Grace M. Maniago, John Paul P. Miranda 5/1/2026

Exploring the Adoption Intention in Using AI-Enabled Educational Tools Among Preservice Teachers in the Philippines: A Partial-Least Square Modeling

Survey of pre-service teachers' adoption intentions for AI-enabled educational tools using UTAUT2 framework.

Ax Emerson Q. Fernando, Julius Ceazar G. Tolentino, Maria Anna D. Cruz, Jordan L. Salenga, Vernon Grace M. Maniago, Juvy C. Grume, Erika M. Pineda, Aileen P. De Leon, John Paul P. Miranda 5/1/2026

Profiles of AI Dependency: A Latent Class Analysis of Filipino Students' Academic Competencies

Study of AI dependency among Filipino college students and effects on critical thinking and academic skills.

Ax Ta-Yang Wang, Rajgopal Kannan, Viktor Prasanna 5/1/2026

TypeBandit: Type-Level Context Allocation and Reweighting for Effective Attribute Completion in Heterogeneous Graph Neural Networks

TypeBandit method for attribute completion in heterogeneous graph neural networks via type-level context allocation.

Ax Bingli Wang, Huanze Tang, Haijun Lv, Zhishan Lin, Lixin Gu, Lei Feng, Qipeng Guo, Kai Chen 5/1/2026

COHERENCE: Benchmarking Fine-Grained Image-Text Alignment in Interleaved Multimodal Contexts

COHERENCE benchmark for evaluating multimodal LLMs on fine-grained image-text alignment in interleaved document contexts.

Ax Jon-Paul Cacioli 5/1/2026

Beyond the Mean: Within-Model Reliable Change Detection for LLM Evaluation

Applies Reliable Change Index from psychology to evaluate LLM version improvements on MMLU-Pro, finding most items show no reliable change.

Ax Zi Li, Tian Zhou, Wenze Li, Jingyu Hua, Yunlong Mao, Sheng Zhong 5/1/2026

Secret Stealing Attacks on Local LLM Fine-Tuning through Supply-Chain Model Code Backdoors

Security research demonstrating secret extraction from local LLM fine-tuning through compromised model code backdoors.

Ax Zehui Tang, Yuchen Liu, Feihu Huang 5/1/2026

AdaBFL: Multi-Layer Defensive Adaptive Aggregation for Bzantine-Robust Federated Learning

AdaBFL framework for Byzantine-robust federated learning against poisoning attacks through adaptive aggregation.