Ax Davide Evangelista, Elena Morotti, Francesco Pivi, Maurizio Gabbrielli 5/14/2026

Improving Diffusion Posterior Samplers with Lagged Temporal Corrections for Image Restoration

Improved diffusion posterior sampling method for image restoration using lagged temporal corrections.

Ax Luca Maria Del Bono, Giulio Biroli, Patrick Charbonneau, Marylou Gabri\'e 5/14/2026

The critical slowing down in diffusion models

Theoretical analysis of computational complexity in diffusion models using statistical field theory.

Ax Ahmed Heakl, Youssef Mohamed, Abdullah Sohail, Rania Elbadry, Ahmed Nassar, Peter W. J. Staar, Fahad Shahbaz Khan, Imran Razzak, Salman Khan 5/14/2026

DocAtlas: Multilingual Document Understanding Across 80+ Languages

DocAtlas: Multilingual OCR and document understanding benchmark covering 82 languages and 9 evaluation tasks.

Ax Eduardo Ochoa Rivera, Ambuj Tewari 5/14/2026

Online Conformal Prediction: Enforcing monotonicity via Online Optimization

Online conformal prediction method enforcing monotonicity across confidence levels. Uncertainty quantification framework.

Ax Tomohiro Hayase, Ryo Karakida 5/14/2026

A Unified Framework for Critical Scaling of Inverse Temperature in Self-Attention

Theoretical analysis of inverse temperature scaling in self-attention for long-context stability. ArXiv paper on attention mechanisms.

Ax Zhiming Yu, Wangtao Lu, Xin Lai 5/14/2026

FePySR: A Neural Feature Extraction Framework for Efficient and Scalable Symbolic Regression

FePySR: Neural feature extraction framework for symbolic regression. Reduces search space by decomposing expressions into reusable components.

Ax Tommaso Giovannelli, Griffin D. Kent 5/14/2026

CHAL: Council of Hierarchical Agentic Language

CHAL proposes hierarchical multi-agent debate framework for LLM reasoning, addressing limitations of majority voting and confidence escalation.

Ax Bastian Lampe, Lutz Eckstein 5/14/2026

A Five-Layer MLOps Architecture for Connected Automated Driving

Five-layer MLOps architecture for continuous assurance of safety and performance in automated driving systems.

Ax Heejin Do, Shashank Sonkar, Mrinmaya Sachan 5/14/2026

Simulating Students or Sycophantic Problem Solving? On Misconception Faithfulness of LLM Simulators

Evaluates whether LLM student simulators faithfully replicate student misconceptions during interaction with AI tutors.

Ax Zhizhen Zhang, Hyemin Gu, Benjamin J. Zhang, Daniel Elenius, Michael Tyrrell, Theo J. Bourdais, Houman Owhadi, Markos A. Katsoulakis, Tuhin Sahai 5/14/2026

ISOMORPH: A Supply Chain Digital Twin for Simulation, Dataset Generation, and Forecasting Benchmarks

ISOMORPH is a public digital twin and benchmark for supply-chain forecasting with configurable parameters and modular topology.

Ax Marcell T. Kurbucz 5/14/2026

When to Trust Confidence Thresholding: Calibration Diagnostics for Pseudo-Labelled Regression

Develops calibration diagnostics for pseudo-labelled regression using confidence thresholding on classifier outputs.

Ax Buyun Liang, Jinqi Luo, Liangzu Peng, Kwan Ho Ryan Chan, Darshan Thaker, Kaleab A. Kinfu, Fengrui Tian, Hamed Hassani, Ren\'e Vidal 5/14/2026

REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations

REALISTA formulates hallucination elicitation in LLMs as constrained optimization finding semantically coherent adversarial prompts.

Ax Kaixiang Zhao, Bolin Shen, Yuyang Dai, Shayok Chakraborty, Yushun Dong 5/14/2026

GraphIP-Bench: How Hard Is It to Steal a Graph Neural Network, and Can We Stop It?

GraphIP-Bench evaluates model-extraction attacks on graph neural networks and tests ownership defenses against surrogate model theft.

Ax Davi Bastos Costa, Renato Vicente 5/14/2026

Persona-Model Collapse in Emergent Misalignment

Studies persona-model collapse in emergent misalignment of LLMs fine-tuned on narrow harmful data, affecting unrelated prompts.

Ax Zhongkai Yu, Yichen Lin, Chenyang Zhou, Yuwei Zhang, Kun Zhou, Junxia Cui, Haotian Ye, Zhengding Hu, Zaifeng Pan, Ruiyi Wang, Yujie Zhao, Hejia Zhang, Jingbo Shang, Jishen Zhao, Yufei Ding 5/14/2026

ChipMATE: Multi-Agent Training via Reinforcement Learning for Enhanced RTL Generation

ChipMATE uses multi-agent reinforcement learning for RTL code generation with self-trained models compatible with vendor security requirements.

Ax Luxu Liang, Xiang Li 5/14/2026

Steer-to-Detect: Probing Hidden Representations for Detection of LLM-Generated Texts

Steer-to-Detect method uses internal LLM representations to detect machine-generated text by steering hidden representations for better discriminative power.

Ax Ryoya Awano, Taiji Suzuki 5/14/2026

The Mechanism of Weak-to-Strong Generalization: Feature Elicitation from Latent Knowledge

Theoretical analysis of weak-to-strong generalization where strong models are fine-tuned on weaker model outputs, studying feature learning under multi-step SGD.

Ax Yuequan Bao, Xing Li, Huabin Sun, Dawei Liu, Yuxuan Tian, Haiyang Hu 5/14/2026

SHM-Agents: A Generalist-Specialist Integrated Agent System for Structural Health Monitoring

SHM-Agents integrates large language models with specialized algorithms for structural health monitoring, combining generalist-specialist reasoning for engineering applications.

Ax One Octadion, Novanto Yudistira, Lailil Muflikhah 5/14/2026

Adaptive Conformal Prediction for Reliable and Explainable Medical Image Classification

Adaptive conformal prediction method for medical image classification with coverage guarantees on difficult inputs.

Ax Young Hyun Cho, Will Wei Sun 5/14/2026

When Should an AI Workflow Release? Always-Valid Inference for Black-Box Generate-Verify Systems

Statistical framework for deciding when LLM-enabled generate-verify workflows should release outputs using always-valid inference.

Ax Xiao Wang, Zihua She, Jianxi Su 5/14/2026

Coreset-Induced Conditional Velocity Flow Matching

Generative model combining coreset-induced source distribution with hierarchical rectified flow for conditional velocity matching.

Ax Jun He, Deying Yu 5/14/2026

Protocol-Driven Development: Governing Generated Software Through Invariants and Evidence

Protocol-driven development framework for governing generated software using invariants and evidence beyond natural language specs.

Ax John R. Minnick, Jinghui Geng, Kamran Hussain, Jesus Gonzalez-Ferrer, Ash Robbins, Mohammed A. Mostajo-Radji, David Haussler, Jason K. Eshraghian, Mircea Teodorescu 5/14/2026

SpikeProphecy: A Large-Scale Benchmark for Autoregressive Neural Population Forecasting

Large-scale benchmark for autoregressive neural population forecasting with structured evaluation metrics beyond correlation.

Ax Toluwani Aremu, Nils Lukas, Jie Zhang 5/14/2026

Watermarking Should Be Treated as a Monitoring Primitive

Proposes watermarking as monitoring primitive for generative models with per-entity attribution keys and detector access.

Ax Ziyang You, Liling Zheng, Xiaoke Yang, Xuxing Lu 5/14/2026

DiffusionHijack: Supply-Chain PRNG Backdoor Attack on Diffusion Models and Quantum Random Number Defense

Supply-chain backdoor attack on diffusion models via PRNG hijacking, with quantum random number defense mechanism.

Ax Nils Loose, Joseph Bienh\"uls, Kristoffer Hempel, Felix M\"achtle, Thomas Eisenbarth 5/14/2026

Code-Centric Detection of Vulnerability-Fixing Commits: A Unified Benchmark and Empirical Study

Benchmark and empirical study using code language models to detect vulnerability-fixing commits across 180k+ commits and 20+ datasets.

Ax Bingying Li, Haiyun He 5/14/2026

On the Generalization of Knowledge Distillation: An Information-Theoretic View

Information-theoretic analysis of knowledge distillation generalization using coupled stochastic processes and KL divergence.

Ax Ishika Agarwal, Sofia Stoica, Emre Can Acikgoz, Pradeep Natarajan, Mahdi Namazifar, Jiaqi Ma, Dilek Hakkani-T\"ur 5/14/2026

AcquisitionSynthesis: Targeted Data Generation using Acquisition Functions

Targeted synthetic data generation method using acquisition functions to improve model training data quality.

Ax Yiyun Zhou, Zhonghua Jiang, Wenkang Han, Kunxi Li, Mingjing Xu, Chang Yao, Jingyuan Chen 5/14/2026

A$_3$B$_2$: Adaptive Asymmetric Adapter for Alleviating Branch Bias in Vision-Language Image Classification with Few-Shot Learning

Adaptive asymmetric adapter (A3B2) for vision-language few-shot learning addressing branch bias between image and text encoders in CLIP-based models.

Ax Beomjin Ahn, Jungmin Kwon, Chanyong Jung, Jaewook Chung 5/14/2026

LoREnc: Low-Rank Encryption for Securing Foundation Models and LoRA Adapters

LoREnc: Training-free encryption framework securing foundation models and LoRA adapters via spectral truncation against IP leakage and model recovery attacks.

Ax Stef van Buuren 5/14/2026

LLMs as Implicit Imputers: Uncertainty Should Scale with Missing Information

Framework for evaluating LLMs as implicit imputers, showing uncertainty should scale with missing information using multiple imputation criteria.

Ax Marte Eggen, Eirik Reiestad, Kristian Gj{\o}steen, Inga Str\"umke 5/14/2026

Backdoor Channels Hidden in Latent Space: Cryptographic Undetectability in Modern Neural Networks

Cryptographically undetectable backdoor attack mechanism for modern neural networks with backdoors hidden in latent space.

Ax Siyuan Zhu, Chao Yu, Rongxin Yang, Zongkai Liu, Jinjun Hu, Qiwen Chen, Yibo Zhang 5/14/2026

GAGPO: Generalized Advantage Grouped Policy Optimization

Generalized Advantage Grouped Policy Optimization for reinforcement learning on LLM agents, improving credit assignment in multi-turn sparse reward settings.

Ax Hanwen Zhang, Dusit Niyato, Wei Zhang, Xin Lou, Malcolm Yoke Hean Low 5/14/2026

An Agentic AI Framework with Large Language Models and Chain-of-Thought for UAV-Assisted Logistics Scheduling with Mobile Edge Computing

Agentic AI framework combining LLMs with chain-of-thought reasoning for UAV logistics scheduling and mobile edge computing task allocation.

Ax Weiqing Luo, Zongye Hu, Xiao Wang, Zhiyuan Yu, Haofeng Zhang, Ziyi Huang 5/14/2026

Utility-Oriented Visual Evidence Selection for Multimodal Retrieval-Augmented Generation

Information-theoretic approach to visual evidence selection in multimodal RAG systems, reformulating selection by information gain rather than semantic similarity.

Ax Jonas Reiter, Jakob Elias Gebler, Hector Geffner 5/14/2026

Differentiable Learning of Lifted Action Schemas for Classical Planning

Differentiable learning approach for discovering lifted action schemas in classical planning for deterministic MDPs with structural generalization.

Ax Zetai Cen, Chenfei Gu, Jin Zhu, Ting Li, Yunxiao Chen, Chengchun Shi 5/14/2026

Learning Perturbations to Extrapolate Your LLM

Framework for enhancing LLM extrapolation performance through learnable continuous perturbations in embedding space rather than discrete fixed designs.

Ax Shubham Kumar Nigam, Suparnojit Sarkar, Piyush Patel 5/14/2026

IndicMedDialog: A Parallel Multi-Turn Medical Dialogue Dataset for Accessible Healthcare in Indic Languages

Parallel multi-turn medical dialogue dataset spanning English and 9 Indic languages with LLM-generated synthetic conversations for healthcare accessibility.

Ax Badr Alboushy, Assef Jafar, Mohamad Aljnidi, Mohamad Bashar Disoki, Aref Shaheed 5/14/2026

Context-Aware Web Attack Detection in Open-Source SIEM Systems via MITRE ATT&CK-Enriched Behavioral Profiling

AI module for open-source Wazuh SIEM using MITRE ATT&CK enrichment for multi-step web attack detection.

Ax Chaehee Song, Minseok Seo, Yeeun Seong, Doyi Kim, Changick Kim 5/14/2026

Query-Conditioned Test-Time Self-Training for Large Language Models

Test-time self-training method enabling LLM parameter updates during inference to correct misconceptions and adapt to queries.

Ax Barathi Ganesh HB, Michal Ptaszynski, Rene Melendez, Juuso Eronen 5/14/2026

Continual Learning with Multilingual Foundation Model

Multi-stage framework using foundation models for detecting reclaimed slurs in multilingual social media.

Ax Yiwen Qiu, Filip Kovacevic, Shimeng Huang, Peter Spirtes, Francesco Locatello 5/14/2026

Towards a holistic understanding of Selection Bias for Causal Effect Identification

Framework for identifying causal effects from observational data with selection bias in biomedical studies.

Ax Huichao Chai, Zhixin Wu, Xuemiao Li, Shiqing Fan, Hengfeng Wang, Maojun Peng, Lu Xu, Yaoyuan Wang, Yibo Jin, Wei Guo, Yongxiang Feng 5/14/2026

TurboGR: An Accelerated Training System for Large-Scale Generative Recommendation

Accelerated training system for large-scale generative recommendation models on Ascend NPUs.

Ax Ruan Visser, Trienko Grobler, Marcel Dunaiski 5/14/2026

Pretraining Language Models with Subword Regularization: An Empirical Study of BPE Dropout in Low-Resource NLP

Empirical investigation of applying BPE dropout during language model pretraining for low-resource NLP tasks.

Ax Yifeng Yu, Lu Yu 5/14/2026

On the Limits of Latent Reuse in Diffusion Models

Study of latent space reuse reliability in diffusion models under distribution shift and subspace misalignment.

Ax H. Moore, S. Qi, D. Milojicic, C. Bash, S. Pasricha 5/14/2026

MARLIN: Multi-Agent Game-Theoretic Reinforcement Learning for Sustainable LLM Inference in Cloud Datacenters

Multi-agent game-theoretic RL system for sustainable LLM inference in cloud datacenters, reducing energy and carbon footprint.

Ax Qingyun Zou, Yingze Li, Tianen Liu, Bingsheng He, Weng-Fai Wong 5/14/2026

Reward-Weighted On-Policy Distillation with an Open Property-Equivalence Verifier for NL-to-SVA Generation

LLM-based generation of SystemVerilog Assertions with reward-weighted distillation and property-equivalence verification.

Ax Edwige Cyffers, Juba Ziani 5/14/2026

Limits of Personalizing Differential Privacy Budgets

Theoretical analysis of personalized privacy budgets in differential privacy, identifying limitations for mean estimation.

Ax Jaeyung Kim, YoungJoon Yoo 5/14/2026

ArcVQ-VAE: A Spherical Vector Quantization Framework with ArcCosine Additive Margin

Novel vector quantization framework using arc-cosine additive margin for improved discrete representation learning in VAEs.

Ax Chengzhi Shen (Cherise), Weixiang Shen (Cherise), Tobias Susetzky (Cherise), Chen (Cherise), Chen (Cherise), Jun Li, Yuyuan Liu, Xuepeng Zhang, Zhenyu Gong, Daniel Rueckert, Jiazhen Pan 5/14/2026

RealICU: Do LLM Agents Understand Long-Context ICU Data? A Benchmark Beyond Behavior Imitation

Benchmark evaluating LLM agents on long-context ICU medical data understanding beyond behavior imitation.