Ax Taijie Chen, Rui Su, Siyuan Feng, Laoming Zhang, Hongyang Zhang, Haijiao Wang, Zhaofeng Ma, Jintao Ke, Li Ma 5/22/2026

D$^3$-Subsidy: Online and Sequential Driver Subsidy Decision-Making for Large-Scale Ride-Hailing Market

Online sequential optimization for driver subsidies in ride-hailing using reinforcement learning.

Ax Alexander Goldberg, Giulia Fanti, Nihar B. Shah 5/22/2026

Smooth Partial Lotteries for Stable Randomized Selection

Lottery-based selection mechanism for stable randomized decision-making in competitive processes.

Ax Xiongbin Wu, Zhihao Luo, Shanzhe Lei, Lechao Zhang, Xuhong Wang, Jie Yang, Zhonglong Zheng, Yuanjie Zheng, Xin Tan, Wei Liu 5/22/2026

GROW: Aligning GRPO with State-Action Modeling for Open-World VLM Agents

Aligns GRPO reinforcement learning with state-action modeling for vision-language model agents in open-world tasks.

Ax Panagiotis Koromilas, Theodoros Giannakopoulos, Mihalis A. Nicolaou, Yannis Panagakis 5/22/2026

Neural Collapse by Design: Learning Class Prototypes on the Hypersphere

Method for achieving neural collapse in supervised classification through prototype learning on hypersphere.

Ax Zhijie Yang, Min Tang, Peng Du, Qiang Zou 5/22/2026

AirfoilGen: A valid-by-construction and performance-aware latent diffusion model for airfoil generation

Latent diffusion model for airfoil generation with geometric validity constraints and physical controllability.

Ax Youngjoon Park 5/22/2026

Decision-Path Patterns as Tree Reliability Signals: Path-based Adaptive Weighting for Random Forest Classification

Adaptive weighting method for random forests using decision-path patterns to improve classification.

Ax Adda Akram Bendoukha, Heber Hwang Arcolezi, Nesrine Kaaniche, Aymen Boudguiga 5/22/2026

Choose Wisely and Privately: Proactive Client Selection for Fair and Efficient Federated Learning

Federated learning approach with proactive client selection to handle non-IID data and improve convergence.

Ax Marcel Hussing, Liv G. d'Aliberti, Claas Voelcker, Benjamin Eysenbach, Eric Eaton 5/22/2026

Behavior-Consistent Deep Reinforcement Learning

Addresses policy divergence in RL by formalizing behavior-consistent training for reliable deployment.

Ax Zeyuan Wang, Da Li, Yulin Chen, Yuehu Gong, Yanming Guo, Ye Shi, Liang Bai, Tianyuan Yu, Yanwei Fu 5/22/2026

Stochastic MeanFlow Policies: One-Step Generative Control with Entropic Mirror Descent

Novel approach combining generative policies with entropic mirror descent for off-policy reinforcement learning.

Ax Nuo Xu, Kaleel Mahmood, Haowen Fang, Ethan Rathbun, Caiwen Ding, Wujie Wen 5/22/2026

Attacking the Spike: On the Transferability and Security of Spiking Neural Networks to Adversarial Examples

Study of adversarial robustness in spiking neural networks with analysis of attack transferability.

Ax Prasad Gabbur 5/22/2026

Improved DDIM Sampling with Moment Matching Gaussian Mixtures

Research on accelerated sampling from diffusion models using Gaussian Mixture Models for moment matching in DDIM.

Ax Hengrui Cai, Shengjie Liu, Rui Song 5/22/2026

Enhancing Causal Reasoning in Large Language Models: A Causal Attribution Model for Precision Fine-Tuning

Causal attribution model enhances LLM interpretability and reasoning through do-operators and precision fine-tuning.

Ax Riyaz Ahuja, Jeremy Avigad, Prasad Tetali, Sean Welleck 5/22/2026

ImProver: Agent-Based Automated Proof Optimization

ImProver agent automatically optimizes formal proofs in Lean by applying style, readability, and modularity criteria.

Ax Jakub Onysk, Quentin J. M. Huys 5/22/2026

Internal narratives parameterise affective states

LLM-based quantification of internal narratives to characterize affective states in psychological assessment.

Ax Yan Jiang, Hao Zhou, Lizhong GU, Tianlong Li, Ruinan Jin, Wanqi Zhou, Ai Han 5/22/2026

NaviAgent: Graph-Driven Bilevel Planning for Scalable Tool Orchestration

NaviAgent uses bilevel graph-based planning for tool orchestration in LLM agents, scaling to hundreds of tools with dependency handling.

Ax Mingkai Deng, Jinyu Hou, Zhiting Hu, Eric Xing 5/22/2026

General Agentic Planning Through Simulative Reasoning with World Models

Agentic planning framework using world models for simulative reasoning, enabling task generalization without re-engineering.

Ax Jon Saad-Falcon, Avanika Narayan, Hakki Orhun Akengin, J. Wes Griffin, Herumb Shandilya, Adrian Gamarra Lafuente, Medhya Goel, Rebecca Joseph, Shlok Natarajan, Etash Kumar Guha, Shang Zhu, Ben Athiwaratkun, John Hennessy, Azalia Mirhoseini, Christopher R\'e 5/22/2026

Intelligence per Watt: Measuring Intelligence Efficiency of Local AI

Study measuring efficiency of small language models on local hardware, comparing energy consumption and performance to cloud-based alternatives.

Ax Chiheng Lou, Sheng Qi, Rui Kang, Yong Zhang, Chen Sun, Pengcheng Wang, Xuanzhe Liu, Xin Jin 5/22/2026

WarmServe: Enabling One-for-Many GPU Prewarming for Multi-LLM Serving

WarmServe improves multi-LLM serving on shared GPUs by predicting workloads and prewarming models to reduce inference latency.

Ax Tobias von Arx, Niels M\"undler, Mark Vero, Maximilian Baader, Martin Vechev 5/22/2026

AutoBaxBuilder: Bootstrapping Code Security Benchmarking

AutoBaxBuilder automates creation of security benchmarks for LLM-generated code using bootstrapping techniques.

Ax Nuoya Xiong, Yuhang Zhou, Hanqing Zeng, Zhaorun Chen, Furong Huang, Shuchao Bi, Lizhu Zhang, Zhuokai Zhao 5/22/2026

Token-Level LLM Collaboration via FusionRoute

FusionRoute enables token-level collaboration between specialized LLMs to achieve broad domain performance without expensive scaling.

Ax Zhanming Shen, Jiaqi Hu, Zeyu Qin, Hao Chen, Wentao Ye, Zenan Huang, Yihong Zhuang, Guoshan Lu, Junlin Zhou, Junbo Zhao 5/22/2026

Training-Trajectory-Aware Token Selection

Training method for efficient LLM distillation addressing performance degradation in student models with strong reasoning ability.

Ax Elias J\"a\"asaari, Ville Hyv\"onen, Teemu Roos 5/22/2026

LEMUR: Learned Multi-Vector Retrieval

Proposes LEMUR, learned multi-vector retrieval system improving on ColBERT's late interaction model for information retrieval efficiency.

Ax Yisi Ke, Tianyu Huang, Yankai Shu, Di He, Jingchu Gai, Liwei Wang 5/22/2026

Towards Solving the Gilbert-Pollak Conjecture via Large Language Models

Explores using large language models to make progress on the Gilbert-Pollak Steiner ratio conjecture in computational geometry.

Ax Jie Jiang, Xinxun Zhang, Enming Zhang, Yuling Xiong, Jun Zhang, Jingwen Wang, Huan Yu, Yuxiang Wang, Hao Wang, Xiao Yan, Jiawei Jiang 5/22/2026

End-to-End Semantic ID Generation for Generative Advertisement Recommendation

Proposes end-to-end semantic ID generation approach for generative advertisement recommendation systems using alternative to residual quantization.

Ax Danlong Yuan, Wei Wu, Zhengren Wang, Xueliang Zhao, Huishuai Zhang, Dongyan Zhao 5/22/2026

SWE-MiniSandbox: Container-Free Reinforcement Learning for Building Software Engineering Agents

Proposes SWE-MiniSandbox, container-free reinforcement learning method for scalable training of software engineering agents without storage overhead.

Ax Simon Rosen, Siddarth Singh, Ebenezer Gelo, Helen Sarah Robertson, Ibrahim Suder, Victoria Williams, Benjamin Rosman, Geraud Nangue Tasse, Steven James 5/22/2026

MoralityGym: A Benchmark for Evaluating Hierarchical Moral Alignment in Sequential Decision-Making Agents

Introduces MoralityGym benchmark with 98 ethical-dilemma problems to evaluate moral alignment in sequential decision-making agents using deontic constraints.

Ax Paul P. Hager, Fabian N. Harang, Luca Pelizzari, Samy Tindel 5/22/2026

The Volterra signature

Proposes Volterra signature as explicit feature representation for history-dependent systems, alternative to implicit memory mechanisms in RNNs and transformers.

Ax Jiawei Chen, Ruoxi Xu, Boxi Cao, Ruotong Pan, Yunfei Zhang, Yifei Hu, Yong Du, Tingting Gao, Yaojie Lu, Yingfei Sun, Xianpei Han, Le Sun, Xiangyu Wu, Hongyu Lin 5/22/2026

Towards Real-world Human Behavior Simulation: Benchmarking Large Language Models on Long-horizon, Cross-scenario, Heterogeneous Behavior Traces

OmniBehavior benchmark evaluates LLMs as user simulators on real-world long-horizon heterogeneous behavior traces.

Ax Ben Wan, Yan Feng, Zihan Tang, Weizhe Huang, Yuting Zeng, Jia Wang, Tongxuan Liu 5/22/2026

RTPrune: Reading-Twice Inspired Token Pruning for Efficient DeepSeek-OCR Inference

Token pruning optimization for DeepSeek-OCR visual-language model reducing inference cost while preserving text fidelity.

Ax Yiqiao Jin, Yiyang Wang, Lucheng Fu, Yijia Xiao, Yinyi Luo, Haoxin Liu, B. Aditya Prakash, Josiah Hester, Jindong Wang, Srijan Kumar 5/22/2026

UniSD: Towards a Unified Self-Distillation Framework for Large Language Models

Unified self-distillation framework for LLMs without external teachers, addressing free-form trajectory supervision challenges.

Ax Alexandre Cristov\~ao Maiorano 5/22/2026

Evaluating Prompt Injection Defenses for Educational LLM Tutors: Security-Usability-Latency Trade-offs

Evaluation methodology for prompt injection defenses in LLM tutors, analyzing security-usability-latency trade-offs.

Ax Tom Sander, Hongyan Chang, Tom\'a\v{s} Sou\v{c}ek, Tuan Tran, Valeriu Lacatusu, Sylvestre-Alvise Rebuffi, Alexandre Mourachko, Surya Parimi, Christophe Ropers, Rashel Moritz, Vanessa Stark, Hady Elsahar, Pierre Fernandez 5/22/2026

TextSeal: A Localized LLM Watermark for Provenance & Distillation Protection

TextSeal watermarking technique for LLMs based on Gumbel-max sampling with zero inference overhead and speculative decoding support.

Ax Guruprasad Raghavan, George Nychis, Rohan Narayana Murthy 5/22/2026

X-SYNTH: Beyond Retrieval -- Enterprise Context Synthesis from Observed Digital Human Attention

Framework for AI agents to synthesize enterprise context from multiple sources beyond simple retrieval-based approaches.

Ax Fengfei Yu, Ruijia Niu, Dongxia Wu, Yian Ma, Rose Yu 5/22/2026

Calibrating LLMs with Semantic-level Reward

Method for calibrating LLMs with semantic-level rewards to improve uncertainty estimation in high-stakes applications.

Ax Kunyang Li, Mubarak Shah, Yuzhang Shang 5/22/2026

Attend Locally, Remember Linearly: Linear Attention as Cross-Frame Memory for Autoregressive Video Diffusion

Linear attention mechanism for autoregressive video diffusion reducing quadratic complexity via cross-frame memory.

Ax Carol Xuan Long, David Simchi-Levi, Feng Zhu, Huangyuan Su, Andre P. Calmon, Flavio P. Calmon 5/22/2026

Reliability and Effectiveness of Autonomous AI Agents in Supply Chain Management

Study of autonomous AI agents in supply chains using Beer Game, shows reasoning models outperform humans with 67% cost reduction.

Ax Kuei-Chun Kao, Daixuan Huo, Yuanhao Ban, Cho-Jui Hsieh 5/22/2026

AutoRubric-T2I: Robust Rule-Based Reward Model for Text-to-Image Alignment

Rule-based reward model for evaluating text-to-image generation alignment, offering interpretability over preference-trained models.

Ax Lautaro Estienne, Erik Ernst, Mat\'ias Vera, Pablo Piantanida, Luciana Ferrer 5/22/2026

$ECUAS_n$: A family of metrics for principled evaluation of uncertainty-augmented systems

Proposes ECUAS_n metrics for evaluating uncertainty-augmented ML systems with principled evaluation framework.

Ax Akshay Manglik (Emily), Apaar Shanker (Emily), Kaustubh Deshpande (Emily), Jason Qin (Emily), Yash Maurya (Emily), Veronica Chatrath (Emily), Vijay S. Kalmath (Emily), Levi Lentz (Emily), Yuan (Emily), Xue 5/22/2026

Insights Generator: Systematic Corpus-Level Trace Diagnostics for LLM Agents

Framework for systematic corpus-level diagnostics of LLM agent execution traces and failure patterns.

HN def- 5/22/2026

Finding Bugs Using LLMs

Materialize uses LLM-based coding agents with Claude to find bugs in code and pull requests, sharing implementation considerations and lessons learned.

HN hamid914 5/22/2026

FreeLLMAPI – One key. One billion free LLM tokens. Every month

Self-hosted OpenAI-compatible proxy aggregating free-tier LLM providers into single API with automatic budget routing.

HN Austin_Conlon 5/22/2026

Best Practices to Produce Maintainable Code with AI [video]

SF Swift talk on production AI coding practices including guardrails, scripts, specialized agents for PRs, and agent reference documentation.

HN essekar 5/22/2026

NVSX: Adopt NVSentinel the Easy Way

NVSentinel GPU fault detection tool with LLM-powered runbook conversion support via NVSX control plane.

HN mooreds 5/22/2026

The AI-Native Interview

Coding agents like Codex and Claude Code are transforming software engineering. Engineers must shift focus from writing precise code to designing agents that produce correct outcomes.

HN bridgettegraham 5/22/2026

Why does it look like LLMs consistently overestimate implementation time?

Discussion: LLMs consistently overestimate implementation time, possibly because training data reflects human timelines rather than AI agent capabilities with Claude/GPT.

HN diebillionaires 5/22/2026

The AI Bubble – No One's Happy

OpenAI's CFO discusses financing chip and data center commitments through banks, private equity, and federal backing at WSJ Tech Live event.

HN d0able 5/22/2026

Making sure what your code does you think it does, with Vet

Vet tool uses LLM to verify git diffs match intended code changes, with walkthrough of debugging p5.js sketches.

HN c4pt0r 5/22/2026

Yet Another AI Teammate

Curated directory of AI teammate products including Perplexity Computer, a multi-model AI agent workspace for research, coding, data analysis, and workflow automation.

BL 5/22/2026

OpenAI named a Leader in enterprise coding agents by Gartner

Gartner names OpenAI a leader in Enterprise AI Coding Agents quadrant, recognizing Codex for innovation and deployment. Industry recognition announcement.

YT DeepLearningAI 5/21/2026

AI Dev 26 x SF | Atai Barkai: Fullstack Agents & Generative UI with AG UI

Talk on fullstack agents and generative UI using AG UI framework.