Isolater - Feed

Ax Danial Monachan, Samira Nazari, Mahdi Taheri, Ali Azarpeyvand, Milos Krstic, Michael Huebner, Christian Herglotz 3/24/2026

Mix-and-Match Pruning: Globally Guided Layer-Wise Sparsification of DNNs

Mix-and-Match Pruning globally-guided layer-wise sparsification framework for compressing DNNs with minimal accuracy loss on edge devices.

Ax Songfeng Zhu 3/24/2026

HSI Image Enhancement Classification Based on Knowledge Distillation: A Study on Forgetting

Knowledge distillation approach mitigating catastrophic forgetting in incremental hyperspectral image classification without storing old samples.

Ax Govinda Kolli, Adinath Madhavrao Dukre, Behzad Bozorgtabar, Dwarikanath Mahapatra, Imran Razzak 3/24/2026

VGS-Decoding: Visual Grounding Score Guided Decoding for Hallucination Mitigation in Medical VLMs

VGS-Decoding training-free method to reduce hallucinations in medical vision-language models using visual grounding scores during inference.

Ax Shasha Yu, Fiona Carroll, Barry L. Bentley 3/24/2026

The Causal Impact of Tool Affordance on Safety Alignment in LLM Agents

Empirical study of how executable tool access impacts safety alignment in LLM agents. Shows tool affordance increases capability-safety misalignment.

Ax Ernest Fokou\'e, Gregory Babbitt, Yuval Leventhal 3/24/2026

Decorrelation, Diversity, and Emergent Intelligence: The Isomorphism Between Social Insect Colonies and Ensemble Machine Learning

Demonstrates mathematical isomorphism between ant colony decision-making and random forest ensemble learning under stochastic ensemble intelligence framework.

Ax Edoardo Calvello, Elizabeth Carlson, Nikola Kovachki, Michael N. Manta, Andrew M. Stuart 3/24/2026

Operator Learning for Smoothing and Forecasting

Theory for data-driven operator learning methods in smoothing and forecasting of dynamical systems and data assimilation.

Ax Erez Badash, Dan Boneh, Ilan Komargodski, Megha Srivastava 3/24/2026

Hawkeye: Reproducing GPU-Level Non-Determinism

Hawkeye system reproduces GPU matrix operations on CPU for verifiable ML without precision loss. Enables reproducible ML inference analysis.

Ax Joanna Zou, Han Cheng Lie, Youssef Marzouk 3/24/2026

Goal-oriented learning of stochastic dynamical systems using error bounds on path-space observables

Goal-oriented learning of surrogate models for stochastic dynamical systems with error bounds on path-dependent observables.

Ax Jerry M. Huang, Lukas Schumacher, Niek Stevenson, Stefan T. Radev 3/24/2026

CogFormer: Learn All Your Models Once

Simulation-based inference framework enabling rapid neural network fitting across varying cognitive modeling assumptions and parameterizations.

Ax Tony Mason 3/24/2026

Epistemic Observability in Language Models

Study finding language models report highest confidence when fabricating, with formal proof this is observational not capability limitation.

Ax Hang-Cheng Dong, Pengcheng Cheng, Shuhuan Li 3/24/2026

Interpretable Operator Learning for Inverse Problems via Adaptive Spectral Filtering: Convergence and Discretization Invariance

SC-Net operator learning framework for regularized inverse problems using spectral filtering with improved interpretability and generalization.

Ax Zhilong Zhang, Haoxiang Ren, Yihao Sun, Yifei Sheng, Haonan Wang, Haoxin Lin, Zhichao Wu, Pierre-Luc Bacon, Yang Yu 3/24/2026

Towards Practical World Model-based Reinforcement Learning for Vision-Language-Action Models

World model-based reinforcement learning approach for training Vision-Language-Action robotic models without costly real-world interaction.

Ax Kry Yik Chau Lui, Cheng Chi, Kishore Basu, Yanshuai Cao 3/24/2026

LassoFlexNet: Flexible Neural Architecture for Tabular Data

Neural architecture incorporating five inductive biases for improved performance on tabular data compared to tree-based models.

Ax Shixiang Liu, Zhifan Li, Hanming Yang, Jianxin Yin 3/24/2026

High-dimensional online learning via asynchronous decomposition: Non-divergent results, dynamic regularization, and beyond

Asynchronous decomposition framework for high-dimensional online learning with dynamic regularization avoiding error bound divergence.

Ax Soudeep Ghoshal, Himanshu Buckchash, Sarita Paudel, Rub\'en Ruiz-Torrubiano 3/24/2026

NDT: Non-Differential Transformer and Its Application to Sentiment Analysis

Non-Differential Transformer architecture for improved sentiment analysis in text, inspired by Differential Transformer variants.

Ax Zihao Zheng, Hangyu Cao, Jiayu Chen, Sicheng Tian, Chenyue Li, Maoliang Li, Xinhao Sun, Guojie Luo, Xiang Chen 3/24/2026

RoboECC: Multi-Factor-Aware Edge-Cloud Collaborative Deployment for VLA Models

RoboECC edge-cloud collaborative deployment framework for Vision-Language-Action models enabling real-time embodied AI inference.

Ax Zacharie Bugaud 3/24/2026

Multi-RF Fusion with Multi-GNN Blending for Molecular Property Prediction

Multi-RF Fusion ensemble method achieving top OGB leaderboard rank for molecular property prediction via Random Forests and GNNs.

Ax Kaiyuan Li, Jing-Cheng Pang, Yang Yu 3/24/2026

RLVR Training of LLMs Does Not Improve Thinking Ability for General QA: Evaluation Method and a Simple Solution

Evaluation and solution for improving general QA performance of LLMs trained with reinforcement learning from verifiable rewards.

Ax Enguang Wang, Qiang Wang, Yuanchen Wu, Ke Yan, Xinbin Yuan, Shouhong Ding, Xialei Liu, Ming-Ming Cheng 3/24/2026

Predictive Regularization Against Visual Representation Degradation in Multimodal Large Language Models

Study of visual representation degradation in MLLMs with predictive regularization technique to preserve visual competence during language training.

Ax Milos Gravara, Juan Luis Herrera, Stefan Nastic 3/24/2026

Compass: Optimizing Compound AI Workflows for Dynamic Adaptation

Compass framework for optimizing compound AI workflows across multiple specialized models with dynamic adaptation under varying loads.

Ax Yujin Park, Haejun Chung, Ikbeom Jang 3/24/2026

Dodgersort: Uncertainty-Aware VLM-Guided Human-in-the-Loop Pairwise Ranking

Dodgersort system for efficient pairwise ranking via CLIP-based pre-ordering, neural ranking, and uncertainty-aware pair selection.

Ax Xiangyu Zeng, Qi Xu, Yunke Wang, Chang Xu 3/24/2026

HiCI: Hierarchical Construction-Integration for Long-Context Attention

HiCI hierarchical attention module for long-context language modeling with segment-level and global context integration.

Ax Siddharth Srivastava, Adam Smith, Scott Brooks, Jack Bacon, Till Bretschneider 3/24/2026

Ensemble of Small Classifiers For Imbalanced White Blood Cell Classification

Lightweight ensemble approach for white blood cell classification addressing class imbalance in medical imaging.

Ax Kaustubh D. Dhole, Eugene Agichtein 3/24/2026

RubricRAG: Towards Interpretable and Reliable LLM Evaluation via Domain Knowledge Retrieval for Rubric Generation

RubricRAG system for interpretable LLM evaluation via domain knowledge retrieval to generate detailed rubrics instead of scalar scores.

Ax Melissa Adrian, Daniel Sanz-Alonso, Rebecca Willett 3/24/2026

Auto-differentiable data assimilation: Co-learning of states, dynamics, and filtering algorithms

Framework for jointly learning state, dynamics, and filtering algorithm parameters in data assimilation via auto-differentiable filtering.

Ax Tanay Varshney, Annie Surla, Michelle Xu, Gomathy Venkata Krishnan, Maximilian Jeblick, David Austin, Neal Vaidya, Davide Onofrio 3/24/2026

LLM Router: Prefill is All You Need

Router mechanism for LLMs using internal prefill activations and Encoder-Target Decoupling to route queries to best-performing models.

Ax Lisan Al Amin, Md Ismail Hossain, Rupak Kumar Das, Mahbubul Islam, Saddam Mukta, Abdulaziz Tabbakh 3/24/2026

Democratizing AI: A Comparative Study in Deep Learning Efficiency and Future Trends in Computational Processing

Benchmark study of deep learning model efficiency comparing Conv6, VGG16, and other architectures under computational constraints.

Ax Bert de Vries 3/24/2026

Active Inference for Physical AI Agents -- An Engineering Perspective

Framework applying Active Inference and Free Energy Principle to physical AI agents and robots operating under resource constraints in real-world environments.

Ax Debdeep Pati 3/24/2026

Stability of Sequential and Parallel Coordinate Ascent Variational Inference

Theoretical analysis of coordinate ascent variational inference stability differences in sequential vs parallel variants for high-dimensional linear regression.

Ax Leandro Vignolo, Matias Gerard 3/24/2026

MOELIGA: a multi-objective evolutionary approach for feature selection with local improvement

MOELIGA multi-objective evolutionary algorithm for feature selection balancing subset size and classification accuracy.

Ax Bo Jiang 3/24/2026

DiscoUQ: Structured Disagreement Analysis for Uncertainty Quantification in LLM Agent Ensembles

DiscoUQ framework extracts semantic structure from disagreement patterns in multi-agent LLM ensembles for improved uncertainty quantification.

Ax Rushikesh Deotale, Adithya Srinivasan, Yuan Tian, Tianyi Zhang, Pavlos Vlachos, Hector Gomez 3/24/2026

ALL-FEM: Agentic Large Language models Fine-tuned for Finite Element Methods

ALL-FEM framework fine-tunes LLMs to automatically generate and analyze finite element method code for engineering simulations.

Ax Erich Studerus, Vivienne Jia Zhong, Stephan Vonschallen 3/24/2026

A Framework for Low-Latency, LLM-driven Multimodal Interaction on the Pepper Robot

Framework integrating LLMs into Pepper robots with low-latency multimodal interaction, enabling speech processing and agentic control capabilities.

Ax Jinquan Zheng, Jia Yuan, Jiacheng Yao, Chenyang Gu, Pujun Zheng, Guoxiu He 3/24/2026

Mitigating Selection Bias in Large Language Models via Permutation-Aware GRPO

Permutation-Aware GRPO method reducing selection bias in LLMs during multiple-choice evaluation by training models to produce consistent answers across option permutations.

Ax Yunhai Hu, Junwei Zhou, Yumo Cao, Yitao Long, Yiwei Xu, Qiyi Jiang, Weiyao Wang, Xiaoyu Cao, Zhen Sun, Yiran Zou, Nan Du 3/24/2026

DSL-R1: From SQL to DSL for Training Retrieval Agents across Structured and Unstructured Data with Reinforcement Learning

DSL-R1 framework training retrieval agents via reinforcement learning to bridge structured metadata and unstructured content using domain-specific language.

Ax Ziquan Wang, Zhongqi Lu 3/24/2026

Knowledge Boundary Discovery for Large Language Models

Knowledge Boundary Discovery framework using reinforcement learning to systematically map what LLMs can and cannot answer reliably.

Ax Taiga Saito, Yu Otake, Daijiro Mizutani, Stephen Wu 3/24/2026

TabPFN Extensions for Interpretable Geotechnical Modelling

TabPFN transformer model applied to geotechnical site characterization using sparse borehole data for uncertainty quantification and interpretability.

Ax Shuoxun Xu, Zhanhao Yan, Lexin Li 3/24/2026

Statistical Learning for Latent Embedding Alignment with Application to Brain Encoding and Decoding

Statistical learning framework for latent embedding alignment in brain encoding/decoding with limited fMRI data.

Ax Yingzhen Yang, Ping Li 3/24/2026

Gradient Descent with Projection Finds Over-Parameterized Neural Networks for Learning Low-Degree Polynomials with Nearly Minimax Optimal Rate

Improved sample complexity bounds for training over-parameterized neural networks to learn low-degree spherical polynomials.

Ax Tin Van Huynh, Kiet Van Nguyen, Ngan Luu-Thuy Nguyen 3/24/2026

ViCLSR: A Supervised Contrastive Learning Framework with Natural Language Inference for Natural Language Understanding Tasks

ViCLSR: contrastive learning framework improving Vietnamese NLU with limited annotated data through supervised representation learning.

Ax Shih-Wen Liu, Yen-Chang Chen, Wei-Ta Chu, Fu-En Yang, Yu-Chiang Frank Wang 3/24/2026