Isolater - Feed

Ax Yi Zhao, Yajuan Peng, Cam-Tu Nguyen, Zuchao Li, Xiaoliang Wang, Xiaoming Fu, Hai Zhao 4/17/2026

TrigReason: Trigger-Based Collaboration between Small and Large Reasoning Models

Framework enabling small reasoning models to accelerate large reasoning model inference by detecting capability boundaries and reasoning risks.

Ax Zhonghao Yang, Yu Li, Yanxu Zhu, Tianyi Zhou, Yuejin Xie, Haoyu Luo, Jing Shao, Xia Hu, Dongrui Liu 4/17/2026

Benchmarks for Trajectory Safety Evaluation and Diagnosis in OpenClaw and Codex: ATBench-Claw and ATBench-CodeX

Trajectory-level safety benchmarks for AI agents in OpenClaw and Codex environments with diagnostic evaluation tools.

Ax Rikard Rosenbacke, Carl Rosenbacke, Victor Rosenbacke, Martin McKee 4/17/2026

The Missing Knowledge Layer in AI: A Framework for Stable Human-AI Reasoning

Framework addressing LLM reliability issues in decision-making by adding knowledge layer to detect reasoning drift and uncertainty.

Ax Thanh Linh Nguyen, Nguyen Van Huynh, Quoc-Viet Pham 4/17/2026

Cooperate to Compete: Strategic Data Generation and Incentivization Framework for Coopetitive Cross-Silo Federated Learning

Federated learning framework for cross-silo healthcare collaboration with incentive mechanisms for coopetitive data contribution.

Ax Xinyu Liu, Xin Liu, Bo Jin, Runsong Zhao, Pengcheng Huang, Junhao Ruan, Bei Li, Chunyang Xiao, Tong Xiao, Jingbo Zhu 4/17/2026

MemoSight: Unifying Context Compression and Multi Token Prediction for Reasoning Acceleration

MemoSight: unified framework combining context compression and multi-token prediction to accelerate chain-of-thought reasoning in LLMs.

Ax Marta Sumyk, Oleksandr Kosovan 4/17/2026

Toward Agentic RAG for Ukrainian

Agentic RAG system for Ukrainian combining two-stage retrieval with lightweight agentic query rephrasing and retry loops on small LLM.

Ax Rikard Rosenbacke, Carl Rosenbacke, Victor Rosenbacke, Martin McKee 4/17/2026

Governing Reflective Human-AI Collaboration: A Framework for Epistemic Scaffolding and Traceable Reasoning

Framework for human-AI collaboration treating reasoning as distributed relational process with epistemic scaffolding and traceable reasoning.

Ax Pei-An Chen, Yong-Ching Liang, Jia-Fong Yeh, Hung-Ting Su, Yi-Ting Chen, Min Sun, Winston Hsu 4/17/2026

ADAPT: Benchmarking Commonsense Planning under Unspecified Affordance Constraints

ADAPT/DynAfford benchmark evaluating embodied agents' ability to plan under object affordance constraints and unspecified environmental conditions.

Ax Yifu Chen, Shengpeng Ji, Zhengqing Liu, Qian Chen, Wen Wang, Ziqing Wang, Yangzhuo Li, Tianle Liang, Zhou Zhao 4/17/2026

Dual-Axis Generative Reward Model Toward Semantic and Turn-taking Robustness in Interactive Spoken Dialogue Models

Dual-axis generative reward model for training full-duplex spoken dialogue agents with semantic and turn-taking robustness via reinforcement learning.

Ax Yifu Chen, Shengpeng Ji, Qian Chen, Tianle Liang, Yangzhuo Li, Ziqing Wang, Wen Wang, Jingyu Lu, Haoxiao Wang, Xueyi Pu, Fan Zhuo, Zhou Zhao 4/17/2026

WavAlign: Enhancing Intelligence and Expressiveness in Spoken Dialogue Models via Adaptive Hybrid Post-Training

WavAlign: adaptive hybrid post-training method enhancing open-source spoken dialogue models via online reinforcement learning.

Ax Andrew Dai, Boris Meinardus, Ciaran Regan, Yingtao Tian, Yujin Tang 4/17/2026

Discovering Novel LLM Experts via Task-Capability Coevolution

Task-capability coevolution framework discovering novel LLM experts with emergent diverse skills through open-ended model-task evolution.

Ax Debodeep Banerjee, Burcu Sayin, Stefano Teso, Andrea Passerini 4/17/2026

Hybrid Decision Making via Conformal VLM-generated Guidance

Conformal learning approach for hybrid decision-making where vision-language models provide guidance with uncertainty calibration for human decisions.

Ax Wenji Fang, Yao Lu, Shang Liu, Jing Wang, Ziyan Guo, Junxian He, Fengbin Tu, Zhiyao Xie 4/17/2026

Dr.~RTL: Autonomous Agentic RTL Optimization through Tool-Grounded Self-Improvement

Dr.RTL: autonomous agentic system for realistic RTL circuit optimization using tool-grounded self-improvement with fine-grained modifications.

Ax Heng Ping, Peiyu Zhang, Shixuan Li, Wei Yang, Anzhe Cheng, Shukai Duan, Xiaole Zhang, Paul Bogdan 4/17/2026

COEVO: Co-Evolutionary Framework for Joint Functional Correctness and PPA Optimization in LLM-Based RTL Generation

COEVO: co-evolutionary framework optimizing functional correctness and PPA simultaneously in LLM-based RTL hardware code generation.

Ax Aihua Li 4/17/2026

Towards Faster Language Model Inference Using Mixture-of-Experts Flow Matching

Mixture-of-experts flow matching framework for faster language model inference while maintaining generation quality.

Ax Wentao Zhang 4/17/2026

Autogenesis: A Self-Evolving Agent Protocol

Autogenesis Protocol: self-evolution protocol for LLM-based agent systems addressing lifecycle management, version tracking, and safe updates.

Ax Ke Xu, Yuhao Wang, Yu Wang 4/17/2026

From Reactive to Proactive: Assessing the Proactivity of Voice Agents via ProVoice-Bench

ProVoice-Bench: evaluation framework for proactive voice agents with four novel tasks assessing multimodal interaction beyond reactive responses.

Ax Simeon Allmendinger, Luca Deck, Lucas Mueller 4/17/2026

Where are the Humans? A Scoping Review of Fairness in Multi-agent AI Systems

Scoping review synthesizing research on fairness in multi-agent AI systems across 23 studies, identifying five archetypal fairness approaches.

Ax Kanzhi Cheng, Zehao Li, Zheng Ma, Nuo Chen, Jialin Cao, Qiushi Sun, Zichen Ding, Fangzhi Xu, Hang Yan, Jiajun Chen, Anh Tuan Luu, Jianbing Zhang, Lewei Lu, Dahua Lin 4/17/2026

OpenMobile: Building Open Mobile Agents with Task and Trajectory Synthesis

OpenMobile: open-source framework for mobile agents using vision-language models, with task and trajectory synthesis for Android task automation.

Ax Shay Snyder (George Mason University), Andrew Capodieci (Neya Robotics), David Gorsich (US Army Ground Vehicle Systems Center), Maryam Parsa (George Mason University) 4/17/2026

HyperSpace: A Generalized Framework for Spatial Encoding in Hyperdimensional Representations

Open-source HyperSpace framework decomposing Vector Symbolic Architecture systems into modular operators for hyperdimensional representations.

Ax Shay Snyder (George Mason University), Andrew Capodieci (Neya Robotics), David Gorsich (US Army Ground Vehicle Systems Center), Maryam Parsa (George Mason University) 4/17/2026

SRMU: Relevance-Gated Updates for Streaming Hyperdimensional Memories

Vector Symbolic Architecture method for sequential associative memories in streaming environments with imbalanced, non-stationary data.

Ax Miri Liu, ChengXiang Zhai 4/17/2026

An Axiomatic Benchmark for Evaluation of Scientific Novelty Metrics

arXiv paper on SRMU, hyperdimensional memory architecture for streaming associative memories with relevance-gated updates.

Ax Zihan Liang, Yufei Ma, Ben Chen, Zhipeng Qian, Huangyu Dai, Lingtao Mao, Xuxin Zhang, Chenyi Lei, Wenwu Ou 4/17/2026

IG-Search: Step-Level Information Gain Rewards for Search-Augmented Reasoning

arXiv paper on axiomatic benchmark for evaluating scientific novelty metrics with AI involvement in research generation.

Ax Mitch Adler, Matthew Russo, Michael Cafarella 4/17/2026

Agent-Aided Design for Dynamic CAD Models

arXiv paper on IG-Search reinforcement learning framework for training LLMs with step-level rewards in search-augmented reasoning.

Ax Ziyang Chen, Renbing Chen, Daowei Li, Jinzhi Liao, Jiashen Sun, Ke Zeng, Xiang Zhao 4/17/2026

Meituan Merchant Business Diagnosis via Policy-Guided Dual-Process User Simulation

arXiv paper on agentic systems for iterative CAD model design using code generation and visual feedback loops.

Ax Hatice Merve Vural, Doga Kukul, Ege Erdem Ozlu, Demir Ekin Arikan, Bob Mankoff, Erkut Erdem, Aykut Erdem 4/17/2026

Learning to Think Like a Cartoon Captionist: Incongruity-Resolution Supervision for Multimodal Humor Understanding

arXiv paper on policy-guided dual-process user simulation for merchant behavior analysis and counterfactual evaluation.

Ax Manan Gupta, Inderjeet Nair, Lu Wang, Dhruv Kumar 4/17/2026

Context Over Content: Exposing Evaluation Faking in Automated Judges

arXiv paper on IRS framework for humor understanding with structured reasoning supervision on multimodal datasets.

Ax M\'elanie Roschewitz, Kenneth Styppa, Yitian Tao, Jiwoong Sohn, Jean-Benoit Delbrouck, Benjamin Gundersen, Nicolas Deperrois, Christian Bluethgen, Julia Vogt, Bjoern Menze, Farhad Nooralahzadeh, Michael Krauthammer, Michael Moor 4/17/2026

RadAgent: A tool-using AI agent for stepwise interpretation of chest computed tomography

arXiv paper exposing vulnerability in LLM-as-judge evaluation where contextual framing influences assessment outcomes.

Ax Moin Aminnaseri, Farima Fatahi Bayat, Nikita Bhutani, Jean-Flavien Bussotti, Kevin Chan, Rafael Li Chen, Yanlin Feng, Jackson Hassell, Estevam Hruschka, Eser Kandogan, Hannah Kim, James Levine, Seiji Maekawa, Jalal Mahmud, Kushan Mitra, Naoki Otani, Pouya Pezeshkpour, Nima Shahbazi, Chen Shen, Dan Zhang 4/17/2026

Blue Data Intelligence Layer: Streaming Data and Agents for Multi-source Multi-modal Data-Centric Applications

Blue Data Intelligence Layer: Agent-based system for NL2SQL queries spanning multiple data sources with streaming data and multimodal inputs.

Ax Zhen Yang, Ping Jian, Zhongbin Guo, Zuming Zhang, Chengzhi Li, Yonghong Deng, Xinyue Zhang, Wenpeng Lu 4/17/2026

How Do LLMs and VLMs Understand Viewpoint Rotation Without Vision? An Interpretability Study

Interpretability study examining how LLMs and VLMs understand viewpoint rotation using only text inputs without visual information.

Ax Manan Gupta, Dhruv Kumar 4/17/2026

Diagnosing LLM Judge Reliability: Conformal Prediction Sets and Transitivity Violations

Diagnostic toolkit for LLM-as-judge reliability using conformal prediction and transitivity analysis revealing widespread per-instance inconsistency.

Ax Yao Tong, Jiayuan Ye, Anastasia Borovykh, Reza Shokri 4/17/2026

Generalization in LLM Problem Solving: The Case of the Shortest Path

Controlled study of LLM generalization using shortest-path planning to isolate effects of training data, paradigms, and inference strategies.

Ax Lijie Zhou, Luran Wang 4/17/2026

An Edge-Cloud Collaborative Architecture for Proactive Elderly Care: Real-Time Risk Assessment and Three-Level Emergency Response

Edge-cloud collaborative architecture for elderly care using real-time risk assessment and emergency response systems.

Ax Yihang Ding, Wanke Xia, Yiting Zhao, Jinbo Su, Jialiang Yang, Zhengbo Zhang, Ke Wang, Wenming Yang 4/17/2026

MemGround: Long-Term Memory Evaluation Kit for Large Language Models in Gamified Scenarios

MemGround: Benchmark for evaluating long-term memory capabilities of LLMs in gamified interactive scenarios with dynamic state tracking.

Ax Baocai Shan, Yuzhuang Xu, Wanxiang Che 4/17/2026

HUOZIIME: An On-Device LLM-enhanced Input Method for Deep Personalization

HUOZIIME: On-device LLM-powered input method editor enabling personalized text generation with privacy preservation on mobile devices.

Ax Domonkos Varga 4/17/2026

Can Large Language Models Detect Methodological Flaws? Evidence from Gesture Recognition for UAV-Based Rescue Operation Based on Deep Learning

Study investigating whether LLMs can identify methodological flaws like data leakage in published ML research papers through automated analysis.

Ax Tomer Atia, Yehudit Aperstein, Alexander Apartsin 4/17/2026

SeaAlert: Critical Information Extraction From Maritime Distress Communications with Large Language Models

SeaAlert: information extraction from maritime distress communications using LLMs for emergency response.

Ax Yuxuan Lai, Xiajing Wang, Chen Zheng 4/17/2026

Chinese Essay Rhetoric Recognition Using LoRA, In-context Learning and Model Ensemble

LoRA fine-tuning and in-context learning for Chinese rhetoric recognition in automated essay scoring.

Ax SAGEA Research Team, Basab Jha, Firoj Paudel, Ujjwal Puri, Adrian Liu, Ethan Henkel, Zhang Yuting, Mateusz Kowalczyk, Mei Huang, Choi Donghyuk, Wang Junhao 4/17/2026

SAGE Celer 2.6 Technical Card

SAGE Celer 2.6: general-purpose language models (5B-27B) with inverse reasoning pipeline for validation and hallucination reduction.

Ax Qi Dong, Ziheng Lin, Ning Ding 4/17/2026

Stateful Evidence-Driven Retrieval-Augmented Generation with Iterative Reasoning

Stateful evidence-driven RAG framework with iterative reasoning for grounding LLMs in external knowledge.

Ax Ananda Rimal (Nepal Engineering College), Adarsha Rimal (Tribhuvan University) 4/17/2026

Benchmarking Linguistic Adaptation in Comparable-Sized LLMs: A Study of Llama-3.1-8B, Mistral-7B-v0.1, and Qwen3-8B on Romanized Nepali

Benchmarking linguistic adaptation of Llama-3.1-8B, Mistral-7B, and Qwen3-8B on Romanized Nepali language.

Ax Ziyin Zhou, Jianyi Zhang, Xu ji, Yilong Li, Jiameng Han, Zhangchi Zhao 4/17/2026

Tug-of-War within A Decade: Conflict Resolution in Vulnerability Analysis via Teacher-Guided Retrieval-Augmented Generations

Teacher-guided retrieval-augmented generation framework for resolving conflicting vulnerability information in LLMs.

Ax Mohammad AL-Smadi 4/17/2026

QU-NLP at ArchEHR-QA 2026: Two-Stage QLoRA Fine-Tuning of Qwen3-4B for Patient-Oriented Clinical Question Answering and Evidence Sentence Alignment

Two-stage QLoRA fine-tuning of Qwen3-4B for clinical question answering and evidence sentence alignment task.

Ax Haiyang Zheng, Nan Pu, Yaqi Cai, Teng Long, Wenjing Li, Nicu Sebe, Zhun Zhong 4/17/2026

The Devil Is in Gradient Entanglement: Energy-Aware Gradient Coordinator for Robust Generalized Category Discovery

Energy-aware gradient coordinator addresses gradient entanglement in generalized category discovery optimization.

Ax Junhong Liang, Yifan Lu, Ekaterina Kochmar, Fajri Koto 4/17/2026

Listen, Correct, and Feed Back: Spoken Pedagogical Feedback Generation

SPFG: dataset and task for generating spoken pedagogical feedback with grammatical error correction and explanations.

Ax Zaifu Zhan, Yu Hou, Kai Yu, Min Zeng, Anita Burgun, Xiaoyi Chen, Rui Zhang 4/17/2026

An Underexplored Frontier: Large Language Models for Rare Disease Patient Education and Communication -- A scoping review

Scoping review of LLM applications for rare disease patient education and communication support.

Ax Jiuting Chen, Yuan Lian, Hao Wu, Tianqi Huang, Hiroshi Sasaki, Makoto Kouno, Jongil Choi 4/17/2026

Internal Knowledge Without External Expression: Probing the Generalization Boundary of a Classical Chinese Language Model

318M parameter Transformer trained on Classical Chinese: investigation of internal knowledge vs external expression in language models.

Ax Vrunda N. Sukhadia, Shammur Absar Chowdhury 4/17/2026