Isolater - Feed

Ax Sihan Zhou, Tiantian He, Yifan Lu, Yaqing Hou, Yew-Soon Ong 4/2/2026

GRASP: Gradient Realignment via Active Shared Perception for Multi-Agent Collaborative Optimization

Multi-agent optimization framework addressing non-stationarity through active shared perception of agent policies.

Ax Bj\"orn Roman Kohlberger (EctoSpace, Dublin, Ireland) 4/2/2026

Spectral Compact Training: Pre-Training Large Language Models via Permanent Truncated SVD and Stiefel QR Retraction

Memory-efficient LLM training via truncated SVD factorization of weight matrices on consumer hardware.

Ax Dong-Jae Lee, Sunghyun Baek, Junmo Kim 4/2/2026

IWP: Token Pruning as Implicit Weight Pruning in Large Vision Language Models

Token pruning framework for vision-language models using attention dual form perspective without retraining.

Ax Swapnil Parekh 4/2/2026

Thinking Wrong in Silence: Backdoor Attacks on Continuous Latent Reasoning

Security analysis of backdoor attacks on language models using continuous latent reasoning without token output.

Ax Dharma Teja Vooturi, Dhiraj Kalamkar, Dipankar Das, Bharat Kaul 4/2/2026

Scalable Pretraining of Large Mixture of Experts Language Models on Aurora Super Computer

LLM pretraining at exascale using Aurora supercomputer with Mula-1B model and Optimus training library.

Ax Hemanth Kotaprolu, Kishan Maharaj, Raey Zhao, Abhijit Mishra, Pushpak Bhattacharyya 4/2/2026

Emotion Entanglement and Bayesian Inference for Multi-Dimensional Emotion Understanding

Bayesian inference framework for multi-dimensional emotion understanding accounting for dependencies among emotions.

Ax Zhanzhi Lou, Hui Chen, Yibo Li, Qian Wang, Bryan Hooi 4/2/2026

Learning to Learn-at-Test-Time: Language Agents with Learnable Adaptation Policies

Language agents with learnable adaptation policies that optimize test-time learning instead of using fixed hand-crafted policies.

Ax Abdullah Al Shafi, Md. Milon Islam, Sk. Imran Hossain, K. M. Azharul Hasan 4/2/2026

KUET at StanceNakba Shared Task: StanceMoE: Mixture-of-Experts Architecture for Stance Detection

Mixture-of-Experts architecture for actor-level stance detection in geopolitical text classification.

Ax Nan Wang, Zhiwei Jin, Chen Chen, Haonan Lu 4/2/2026

PixelPrune: Pixel-Level Adaptive Visual Token Reduction via Predictive Coding

PixelPrune: adaptive visual token reduction for vision-language models using predictive coding for document and GUI tasks.

Ax Razvan Mihai Popescu, David Gros, Andrei Botocan, Rahul Pandita, Prem Devanbu, Maliheh Izadi 4/2/2026

Investigating Autonomous Agent Contributions in the Wild: Activity Patterns and Code Change over Time

Dataset and analysis of autonomous coding agents' contributions in real-world projects, examining code quality and team dynamics over time.

Ax Dylan B. Lewis, Jens Gregor, Hector Santos-Villalobos 4/2/2026

Representation Selection via Cross-Model Agreement using Canonical Correlation Analysis

Training-free canonical correlation analysis method for improving efficiency of pretrained image encoder representations.

Ax Hsin-Ling Hsu, Min-Yu Chen, Nai-Chia Chen, Yan-Ru Chen, Yi-Ling Chang, Fang Yu 4/2/2026

WARP: Guaranteed Inner-Layer Repair of NLP Transformers

WARP: method for repairing adversarial vulnerabilities in transformer NLP models with provable inner-layer repair guarantees.

Ax Ruijie Hao, Longfei Zhang, Yang Dai, Yang Ma, Xingxing Liang, Guangquan Cheng 4/2/2026

Flow-based Policy With Distributional Reinforcement Learning in Trajectory Optimization

Reinforcement learning with flow-based policies and distributional RL for trajectory optimization in multi-solution problems.

Ax Xiangqi Wang, Yue Huang, Haomin Zhuang, Kehan Guo, Xiangliang Zhang 4/2/2026

Dual Optimal: Make Your LLM Peer-like with Dignity

Dignified Peer framework addressing evasive and sycophantic behavior in aligned LLMs through anti-sycophancy and empathy.

Ax Zhengyang Tang, Ke Ji, Xidong Wang, Zihan Ye, Xinyuan Wang, Yiduo Guo, Ziniu Li, Chenxin Li, Jingyuan Hu, Shunian Chen, Tongxu Luo, Jiaxi Bi, Zeyu Qin, Shaobo Wang, Xin Lai, Pengyuan Lyu, Junyi Li, Can Xu, Chengquan Zhang, Han Hu, Ming Yan, Benyou Wang 4/2/2026

Do Phone-Use Agents Respect Your Privacy?

MyPhoneBench: evaluation framework measuring privacy-compliant behavior in mobile phone-use agents during task execution.

Ax Jinkun Hao, Mingda Jia, Ruiyan Wang, Xihui Liu, Ran Yi, Lizhuang Ma, Jiangmiao Pang, Xudong Xu 4/2/2026

EgoSim: Egocentric World Simulator for Embodied Interaction Generation

Egocentric world simulator generating interaction videos with persistent 3D scene state updates for embodied AI.

Ax Yiheng Wang, Lichen Zhu, Yueqian Lin, Yudong Liu, Jingyang Zhang, Hai "Helen" Li, Yiran Chen 4/2/2026

Query-Conditioned Evidential Keyframe Sampling for MLLM-Based Long-Form Video Understanding

Query-conditioned keyframe sampling approach for long-form video understanding with multimodal LLMs using evidential reasoning.

Ax Yiru Wang, Xinyue Shen, Yaohui Han, Michael Backes, Pin-Yu Chen, Tsung-Yi Ho 4/2/2026

OrgAgent: Organize Your Multi-Agent System like a Company

OrgAgent: hierarchical multi-agent framework organizing LLM-based agents into governance, execution, and compliance layers for complex reasoning.

Ax Rafael Sojo, Pedro Larra\~naga, Concha Bielza 4/2/2026

Transfer learning for nonparametric Bayesian networks

Transfer learning algorithms for nonparametric Bayesian network structure learning under limited data.

Ax Zhichen Liu, Tianle Lun, Zhibin Wen, Hao An, Yulin Ou, Jianhui Xu, Hao Zhang, Wenyi Fang, Yang Zheng, Yang Xu 4/2/2026

Fast and Accurate Probing of In-Training LLMs' Downstream Performances

Method for fast probing of LLM downstream performance during training using metrics correlated with performance beyond training loss.

Ax Jingjie Ning, Xueqi Li, Chengyu Yu 4/2/2026

Revision or Re-Solving? Decomposing Second-Pass Gains in Multi-LLM Pipelines

Controlled decomposition of multi-LLM revision pipelines to separate gains into re-solving, scaffold, and content components across benchmarks.

Ax Anubhab Sahu, Diptisha Samanta, Reza Soosahabi 4/2/2026

Automated Framework to Evaluate and Harden LLM System Instructions against Encoding Attacks

Automated framework to evaluate and harden LLM system instructions against encoding-based attacks to prevent credential and policy leakage.

Ax Ying Xie 4/2/2026

VibeGuard: A Security Gate Framework for AI-Generated Code

Security framework to detect and prevent vulnerabilities in AI-generated code through systematic verification of code safety gates.

Ax Awais Khan, Muhammad Umar Farooq, Kutub Uddin, Khalid Malik 4/2/2026

TRACE: Training-Free Partial Audio Deepfake Detection via Embedding Trajectory Analysis of Speech Foundation Models

Training-free detection method for partial audio deepfakes using speech foundation models without frame-level annotations.

Ax Anooshka Bajaj, Deven Mahesh Mistry, Sahaj Singh Maini, Yash Aggarwal, Billy Dickson, Zoran Tiganj 4/2/2026

Temporal Dependencies in In-Context Learning: The Role of Induction Heads

Analysis of how LLMs use induction heads to track and retrieve information from context, revealing serial-recall patterns in in-context learning.

Ax Griffin Pitts, Neha Rani, Weedguet Mildort 4/2/2026

Trust and Reliance on AI in Education: AI Literacy and Need for Cognition as Moderators

Study on how students' trust in AI assistants affects their reliance and critical evaluation of AI-generated output in educational settings.

Ax Atsuyuki Miyai, Mashiro Toyooka, Zaiying Zhao, Kenta Watanabe, Toshihiko Yamasaki, Kiyoharu Aizawa 4/2/2026

Paper Reconstruction Evaluation: Evaluating Presentation and Hallucination in AI-written Papers

PaperRecon evaluation framework for assessing quality and hallucination risks in AI-generated research papers from coding agents.

Ax Mohammad R. Abu Ayyash 4/2/2026

Brainstacks: Cross-Domain Cognitive Capabilities via Frozen MoE-LoRA Stacks for Continual LLM Learning

Brainstacks modular architecture for continual multi-domain LLM fine-tuning using MoE-LoRA stacks composing frozen adapters for domain expertise.

Ax Cai Zhou, Zekai Wang, Menghua Wu, Qianyu Julie Zhu, Flora C. Shi, Chenyu Wang, Ashia Wilson, Tommi Jaakkola, Stephen Bates 4/2/2026

Online Reasoning Calibration: Test-Time Training Enables Generalizable Conformal LLM Reasoning

ORCA framework for test-time calibration of LLM reasoning using conformal prediction, improving efficiency of sampling-based scaling methods.

Ax Ken M. Nakanishi 4/2/2026

Screening Is Enough

Multiscreen architecture introducing explicit query-key relevance rejection mechanism in attention, improving LLM discrimination of irrelevant information.

Ax J. E. Dom\'inguez-Vidal 4/2/2026

A ROS 2 Wrapper for Florence-2: Multi-Mode Local Vision-Language Inference for Robotic Systems

ROS 2 middleware integration for Florence-2 vision-language model in robotics systems, enabling local inference for robotic perception.

Ax Nandan Thakur, Zijian Chen, Xueguang Ma, Jimmy Lin 4/2/2026

ORBIT: Scalable and Verifiable Data Generation for Search Agents on a Tight Budget

ORBIT dataset with 20K reasoning-intensive queries for training search agents combining LMs and web search, using verifiable generation methodology.

Ax Youssef Mroueh, Carlos Fonseca, Brian Belgodere, David Cox 4/2/2026

CliffSearch: Structured Agentic Co-Evolution over Theory and Code for Scientific Algorithm Discovery

Agentic evolutionary framework for scientific algorithm discovery combining LLM-guided search with structured theory and code co-evolution.

Ax Muyu He, Adit Jain, Anand Kumar, Vincent Tu, Soumyadeep Bakshi, Sachin Patro, Nazneen Rajani 4/2/2026

$\texttt{YC-Bench}$: Benchmarking AI Agents for Long-Term Planning and Consistent Execution

Benchmark for evaluating LLM agents on long-term planning over one-year startup simulation with hundreds of turns, testing strategic coherence under uncertainty.

Ax Piyush Garg, Diana R. Gergel, Andrew E. Shao, Galen J. Yacalis 4/2/2026

The Recipe Matters More Than the Kitchen:Mathematical Foundations of the AI Weather Prediction Pipeline

Mathematical framework analyzing AI weather prediction pipelines, emphasizing training methodology and data diversity over architecture choices.

Ax Bhrij Patel, Souradip Chakraborty, Mengdi Wang, Dinesh Manocha, Amrit Singh Bedi 4/2/2026

Code Comprehension then Auditing for Unsupervised LLM Evaluation

Method for unsupervised code correctness evaluation using LLMs through code comprehension before auditing, eliminating need for reference implementations.

Ax Aditi Singh, Abul Ehtesham, Saket Kumar, Tala Talaei Khoei, Athanasios V. Vasilakos 4/2/2026

Agentic Retrieval-Augmented Generation: A Survey on Agentic RAG

Survey of agentic RAG systems combining LLMs with real-time retrieval to address static training data limitations and improve contextual accuracy.

Ax Matthew DosSantos DiSorbo, Harang Ju, Sinan Aral 4/2/2026

Teaching AI to Handle Exceptions: Supervised Fine-Tuning with Human-Aligned Judgment

Research on fine-tuning LLMs as agentic systems to handle exceptions and improve decision-making in complex real-world contexts.

Ax Marco Valentino, Geonhee Kim, Dhairya Dalal, Zhixue Zhao, Andr\'e Freitas 4/2/2026

Mitigating Content Effects on Reasoning in Language Models through Fine-Grained Activation Steering

Study on mitigating reasoning biases in LLMs through activation steering at inference time to improve logical validity discrimination.

Ax Miho Koda, Yu Zheng, Ruixian Ma, Mingyang Sun, Devesh Pansare, Fabio Duarte, Paolo Santi 4/2/2026

LocationReasoner: Evaluating LLMs on Real-World Site Selection Reasoning

Research evaluating LLM reasoning capabilities on real-world site selection tasks, testing if models like o1 and DeepSeek-R1 generalize beyond math/code domains.

Ax Junxing Hu, Ai Han, Haolan Zhan, Pu Wei, Zhiqian Zhang, Yuhang Guo, Jiawei Lu, Zhen Chen, Haoran Li, Zicheng Zhang 4/2/2026

HiMA-Ecom: Enabling Joint Training of Hierarchical Multi-Agent E-commerce Assistants

Benchmark and framework for training hierarchical multi-agent LLM systems with master-coordinator and specialized sub-agents for e-commerce applications.

Ax Chenyu Zhou, Jingyuan Yang, Linwei Xin, Yitian Chen, Ziyan He, Dongdong Ge 4/2/2026

Auto-Formulating Dynamic Programming Problems with Large Language Models

Approach using LLMs to automate formulation of dynamic programming models for operations research, addressing stochastic transitions and data scarcity.

Ax Ammar Ahmed, Azal Ahmad Khan, Ayaan Ahmad, Sheng Di, Zirui Liu, Ali Anwar 4/2/2026

Retrieval-of-Thought: Efficient Reasoning via Reusing Thoughts

Retrieval-of-Thought method that reuses reasoning steps across problems via thought graphs to improve inference efficiency and reduce latency/cost.

Ax Boxuan Zhang, Yi Yu, Jiaxuan Guo, Jing Shao 4/2/2026

Dive into the Agent Matrix: A Realistic Evaluation of Self-Replication Risk in LLM Agents

Research on self-replication risks in LLM agents driven by objective misalignment, moving from theoretical concern to practical reality assessment.

Ax Zheng Zhang, Jiarui He, Yuchen Cai, Deheng Ye, Peilin Zhao, Ruili Feng, Hao Wang 4/2/2026

Genesis: Evolving Attack Strategies for LLM Web Agent Red-Teaming

Genesis: framework evolving attack strategies for red-teaming LLM web agents using behavioral pattern learning.

Ax Xiao Yang, Xuejiao Zhao, Zhiqi Shen 4/2/2026

EHRStruct: A Comprehensive Benchmark Framework for Evaluating Large Language Models on Structured Electronic Health Record Tasks

EHRStruct: benchmark framework evaluating LLM performance on structured electronic health record tasks with standardized metrics.

Ax Xiaohan Zhang, Tian Gao, Mingyue Cheng, Bokai Pan, Ze Guo, Yaguo Liu, Xiaoyu Tao, Qi Liu 4/2/2026