Isolater - Feed

Ax Lilan Peng, Yandi Liu, Qingren Yao, Chongshou Li, Tianrui Li 6/15/2026

MP3: Multi-Period Pattern Pre-training for Spatio-Temporal Forecasting

MP3 pre-training method for spatio-temporal forecasting addressing temporal mirage problem in graph neural networks.

Ax Bora Kargi, David Salinas 6/15/2026

From Uncertain Judgments to Calibrated Rankings: Conformal Elo Estimation for LLM Evaluation

Conformal Elo estimation framework for calibrating LLM-as-a-judge rankings while accounting for systematic biases and uncertainty.

Ax Mariya Pavlova, Harrison Bo Hua Zhu, Elizsveta Semenova, Yingzhen Li 6/15/2026

Quantizing Time-Series Models As Dynamical Systems: Trajectory-Based Quantization Sensitivity Score

Trajectory-based Quantization Sensitivity Score metric for post-training quantization using dynamical systems analysis.

Ax Guo Yu, Wenlin Liu, Yulan Hu, Hao-Xuan Ma, Jun-Peng Jiang, Han-Jia Ye 6/15/2026

Dense Supervision, Sparse Updates: On the Sparsity and Geometry of On-Policy Distillation

Analysis of on-policy distillation in language and vision-language models examining sparsity and parameter geometry of updates.

Ax Rajesh Kumar Das, Mohammad Sharif Ul Islam 6/15/2026

Application of Artificial Intelligence and Machine Learning in Libraries: A Systematic Review

Systematic review of AI and machine learning applications in library systems and information management.

Ax Bal\'azs Csan\'ad Cs\'aji, L\'aszl\'o Gy\"orfi, Ambrus Tam\'as, Harro Walk 6/15/2026

On Rate-Optimal Partitioning Classification from Observable and from Privatised Data

Convergence rate analysis for partitioning classification under relaxed conditions with privacy considerations.

Ax Jingbo Liu 6/15/2026

Stability of a Generalized Debiased Lasso with Applications to Resampling-Based Variable Selection

Theoretical analysis of generalized debiased Lasso estimator with stability principles and variable selection applications.

Ax Samar Fares, Klea Ziu, Toluwani Aremu, Nikita Durasov, Martin Tak\'a\v{c}, Pascal Fua, Ivan Laptev, Karthik Nandakumar 6/15/2026

MirrorCheck: Efficient Adversarial Defense for Vision-Language Models

MirrorCheck framework detects adversarial attacks on vision-language models using text-to-image regeneration in multimodal settings.

Ax Vittorio Torri, Elisa Barbieri, Anna Cantarutti, Carlo Giaquinto, Francesca Ieva 6/15/2026

Automatic identification of diagnosis from hospital discharge letters via weakly supervised Natural Language Processing

Weakly supervised NLP pipeline for classifying diagnoses in Italian hospital discharge letters without manual document-level annotation.

Ax Safayat Bin Hakim, Muhammad Adil, Alvaro Velasquez, Houbing Herbert Song 6/15/2026

ANSR-DT: A Neuro-Symbolic Framework for Adaptive and Explainable Digital Twins

Neuro-symbolic framework combining anomaly detection, symbolic reasoning, and RL for interpretable industrial digital twins.

Ax Woongyeong Yeo, Kangsan Kim, Soyeong Jeong, Jinheon Baek, Sung Ju Hwang 6/15/2026

UniversalRAG: Retrieval-Augmented Generation over Corpora of Diverse Modalities and Granularities

UniversalRAG: Multi-modal retrieval-augmented generation over corpora combining text, images, videos. Unifies diverse modality RAG into single framework.

Ax Yuxin Zhang, Meihao Fan, Ju Fan, Mingyang Yi, Yuyu Luo, Guoliang Li, Bin Wu, Wenchao Zhou 6/15/2026

Reward-SQL: Boosting Text-to-SQL via Stepwise Execution-Aware Reasoning and Process-Supervised Rewards

Reward-SQL: Text-to-SQL via RL with execution-aware reasoning and process-supervised rewards. Improves LLM complex query generation with database feedback.

Ax NaHyeon Park, Na Min An, Kunhee Kim, Soyeon Yoon, Jiahao Huo, Hyunjung Shim 6/15/2026

Aligned but Stereotypical? How System Prompts Shape Demographic Bias in LLM-Based Text-to-Image Models

Studies demographic bias in text-to-image models using LLM-based text conditioning. Shows implicit demographic assumptions even with unspecified attributes.

Ax Fabian Kreppel, Reza Salkhordeh, Ferdinand Schmidt-Kaler, Andr\'e Brinkmann 6/15/2026

Shuttling Compiler for Trapped-Ion Quantum Computers Based on Large Language Models

First LLM-based compiler for trapped-ion quantum computers. Fine-tunes LLMs to learn shuttling operations for quantum qubit layout-independent compilation.

Ax Pere Martra 6/15/2026

Fragile Knowledge, Robust Instruction-Following: The Width Pruning Dichotomy in Llama-3.2

Width pruning analysis of Llama-3.2 revealing trade-offs between knowledge retention and instruction-following. Shows GLU-MLP pruning dichotomy in model capabilities.

Ax Zhenyu Gao, Wenxi Jiang, Yutong Yan 6/15/2026

Detecting Lookahead Bias in LLM Forecasts

Method to detect lookahead bias in LLM economic forecasts via Lookahead Propensity statistic. Identifies information leakage in model training data.

Ax Prateek Jain, Shabari S Nair, Ritesh Goru, Prakhar Agarwal, Ajay Yadav, Yoga Sri Varshan Varadharajan, Constantine Caramanis 6/15/2026

Succeeding at Scale: Enterprise Retrieval Benchmark Construction and Index-Preserving Query Adaptation for Multi-Tenant Search

DevRev: Addresses multi-tenant retrieval systems using query adaptation without full re-indexing. Leverages unlabeled query logs for domain adaptation at scale.

Ax Ao Sun, Xiaoyu Wang, Zhe Tan, Yu Li, Jiachen Zhu, Yuheng Jia, Shu Su 6/15/2026

CuMA: Aligning LLMs with Sparse Cultural Values via Demographic-Aware Mixture of Adapters

CuMA: Aligns LLMs with diverse cultural values using demographic-aware mixture of adapters. Addresses mean collapse and cultural sparsity in model alignment.

Ax Nuoya Xiong, Yuhang Zhou, Hanqing Zeng, Zhaorun Chen, Furong Huang, Shuchao Bi, Lizhu Zhang, Zhuokai Zhao 6/15/2026

Token-Level LLM Collaboration via FusionRoute

FusionRoute: LLM collaboration framework routing tokens to specialized models for improved performance across domains. Enables efficient multi-domain LLM inference.

Ax Pietro Barbiero, Mateo Espinosa Zarlenga, Francesco Giannini, Alberto Termine, Filippo Bonchi, Mateja Jamnik, Giuseppe Marra 6/15/2026

Actionable Interpretability Must Be Defined in Terms of Symmetries

Theoretical framework proposing symmetries as basis for defining interpretability in AI models. Argues existing interpretability definitions are untestable.

Ax Yifan Jiang, Cong Zhang, Bofei Zhang, Qiaofeng Zheng, Yifan Yang, Bingzhang Wang, Yew-Soon Ong 6/15/2026

Pix2Fact: When Vision Is Not Enough -- Benchmarking Fine-Grained VQA with Web Verification on High-Resolution Real-World Scenes

Benchmark for vision-language models combining fine-grained visual grounding with knowledge retrieval. Tests VLM capabilities on real-world high-resolution scenes.

Ax Amirreza Kazemi, Seyed Mohammad Azimi-Abarghouyi, Gabor Fodor, Carlo Fischione 6/15/2026

Mitigating Heterogeneity-Induced Drift in Hierarchical Sign-Based Federated Learning

Federated learning method for hierarchical systems using sign-based gradient compression. Machine learning research on distributed training optimization.

Ax Sasha Behrouzi, Lichao Wu, Mohamadreza Rostami, Ahmad-Reza Sadeghi 6/15/2026

NeST: Neuron Selective Tuning for LLM Safety

NeST: neuron selective tuning method for LLM safety that provides parameter-efficient and maintainable alternative to full fine-tuning.

Ax Avni Mittal 6/15/2026

Did You Forget What I Asked? Prospective Memory Failures in Large Language Models

Study of prospective memory failures in LLMs when formatting constraints conflict with concurrent task demands across 8,000 prompts.

Ax Teng Chen, Sheng Xu, Feixiang Guo, Xiaoyu Wang, Qingqing Gu, Hongyan Li, Luo Ji 6/15/2026

Efficient Rationale-based Retrieval: On-policy Distillation from Generative Rerankers based on JEPA

Rabtriever: efficient rationale-based retrieval system using LLM-based generative rerankers with on-policy distillation for cross-encoding.

Ax Koen van der Blom, Diederick Vermetten 6/15/2026

On the Influence of the Feature Computation Budget on Per-Instance Algorithm Selection for Black-Box Optimization

Analysis of feature computation budget's influence on per-instance algorithm selection effectiveness in black-box optimization.

Ax Seunghan Lee, Jun Seo, Jaehoon Lee, Sungdong Yoo, Minjae Kim, Tae Yoon Lim, Dongwan Kang, Hwanil Choi, SoonYoung Lee, Wonbin Ahn 6/15/2026

AdaTKG: Adaptive Memory for Temporal Knowledge Graph Reasoning

AdaTKG: adaptive memory mechanism for temporal knowledge graph reasoning that maintains entity representations based on interaction history.

Ax Diego Casadei 6/15/2026

A Unified Framework for Structured Flow Modeling: From Representation to Verification and Model Discovery

Unified framework for modeling structured flows combining source/sink behavior, cyclic dynamics, and topology-constrained transport.

Ax Jiangyu Chen, Ban Yi 6/15/2026

Evidence-Gated LLM Priors for Multi-Objective Bayesian Optimization

Method for using LLM-guided priors in multi-objective Bayesian optimization with evidence-gating to calibrate LLM confidence to objective values.

Ax Roel Hulsman, Carles Balsells-Rodas, Sara Magliacane 6/15/2026

Identifiable Markov Switching Models with Instantaneous Effects and Exponential Families

Research on identifiable Markov Switching Models with instantaneous effects for modeling non-stationary temporal systems.

Ax Anjun Gao, Yueyang Quan, Yufei Xia, Zhuqing Liu, Minghong Fang 6/15/2026

Patcher: Post-Hoc Patching of Backdoored Large Language Models

Patcher: defense method against backdoor attacks in LLMs that identifies and mitigates poisoned safety alignment without requiring attack details.

Ax Dezhi Yu, Yvonne Qiu, ShuoJia Fu 6/15/2026

Direct Preference Optimization for Chatbot Fine-Tuning: An Empirical Study

Empirical study on Direct Preference Optimization for fine-tuning LLMs, showing improved efficiency and competitive performance with simplified training pipeline.

Ax Dezhi Yu, Kyuil Lee, Yongkang Huang 6/15/2026

Generative Modeling of Bach-Style Symbolic Music: A Comparative Study of Autoregressive, Latent-Variable, and Adversarial Approaches

Comparative study of autoregressive LSTMs, VAEs, and GANs for generating Bach-style symbolic piano music using MIDI data.

HN SaeedZF 6/14/2026