Ax Antoni Kowalczuk, Jan Dubi\'nski, Franziska Boenisch, Adam Dziedzic 21d ago

Privacy Attacks on Image AutoRegressive Models

Comprehensive privacy attack analysis on image autoregressive models, identifying membership inference and extraction vulnerabilities.

Ax Mohammad Albinhassan, Pranava Madhyastha, Alessandra Russo 21d ago

$\texttt{SEM-CTRL}$: Semantically Controlled Decoding

Method for enforcing syntactic and semantic constraints in LLM decoding through MCTS-guided token-level control.

Ax Musfiqur Rahman, SayedHassan Khatoonabadi, Emad Shihab 21d ago

OpenClassGen: A Large-Scale Corpus of Real-World Python Classes for LLM Research

Large-scale corpus of 324,843 Python classes from open-source projects for training and evaluating LLMs on code generation.

Ax Dezheng Han, Yibin Jia, Ruxiao Chen, Wenjie Han, Shuaishuai Guo, Jianbo Wang 21d ago

ReCellTy: Domain-Specific Knowledge Graph Retrieval-Augmented LLMs Reasoning Workflow for Single-Cell Annotation

RAG-based LLM workflow using domain-specific knowledge graph for automated single-cell type annotation in biology.

Ax Rui Melo, Claudia Mamede, Andre Catarino, Rui Abreu, Henrique Lopes Cardoso 21d ago

Are Sparse Autoencoders Useful for Java Function Bug Detection?

Study evaluating sparse autoencoders for detecting bugs in Java code, addressing software vulnerability detection.

Ax Ozsel Kilinc, Cem Tarhan 21d ago

RQR3D: Reparametrizing the regression targets for BEV-based 3D object detection

Technique for improving BEV-based 3D object detection in autonomous driving by reparametrizing regression targets.

Ax Charig Yang, Samiul Alam, Shakhrul Iman Siam, Michael J. Proulx, Lambert Mathias, Kiran Somasundaram, Luis Pesqueira, James Fort, Sheroze Sheriffdeen, Omkar Parkhi, Carl Ren, Mi Zhang, Yuning Chai, Richard Newcombe, Hyo Jin Kim 21d ago

Reading Recognition in the Wild

Task and dataset for detecting when users are reading in egocentric smart glasses video using multimodal models.

Ax Thinh Pham, Nguyen Nguyen, Pratibha Zunjare, Weiyuan Chen, Yu-Min Tseng, Tu Vu 21d ago

SealQA: Raising the Bar for Reasoning in Search-Augmented Language Models

Benchmark dataset (SealQA) for evaluating search-augmented LLMs on fact-seeking questions with conflicting or noisy search results.

Ax Adrian-Marius Dumitran, Radu Dita, Angela Liliana Dumitran 21d ago

BacPrep: Lessons from Deploying an LLM-Based Bacalaureat Assessment Platform

Deployment case study of LLM-based platform for automated assessment of Romanian Bacalaureat exam questions using Gemini Flash.

Ax Tianjiao Yu, Vedant Shah, Muntasir Wahed, Ying Shen, Kiet A. Nguyen, Ismini Lourentzou 21d ago

Part$^{2}$GS: Part-aware Modeling of Articulated Objects using 3D Gaussian Splatting

Framework for 3D reconstruction of articulated objects using part-aware Gaussian splatting representation.

Ax Scarlett Raine, Tobias Fischer 21d ago

AI-Driven Marine Robotics: Emerging Trends in Underwater Perception and Ecosystem Monitoring

Survey of AI applications in marine robotics for ecosystem monitoring and conservation using underwater perception.

Ax Alissa A. Valentine, Lauren A. Lepow, Lili Chan, Alexander W. Charney, Isotta Landi 21d ago

Bias Detection in Emergency Psychiatry: Linking Negative Language to Diagnostic Disparities

Analysis of clinician bias in emergency psychiatry using NLP to detect negative language linked to diagnostic disparities.

Ax Himanshu Singh, A. V. Subramanyam, Shivank Rajput, Mohan Kankanhalli 21d ago

Nearest Neighbor Projection Removal Adversarial Training

Adversarial training framework for neural networks that mitigates inter-class feature overlap to improve robustness.

Ax Hyungjin Chung, Hyelin Nam, Jiyeon Kim, Hyojun Go, Byeongjun Park, Junho Kim, Joonseok Lee, Seongsu Ha, Byung-Hoon Kim 21d ago

Video Parallel Scaling: Aggregating Diverse Frame Subsets for VideoLLMs

Inference method for VideoLLMs that processes multiple frame subsets in parallel to improve temporal detail without increasing context window.

Ax Christoph Timmermann, Hyunse Lee, Woojin Lee 21d ago

SeMoBridge: Semantic Modality Bridge for Efficient Few-Shot Adaptation of CLIP

Technique to improve CLIP few-shot classification by addressing modality gap through semantic bridging between image and text embeddings.

Ax Ayan Majumdar, Feihao Chen, Jinghui Li, Xiaozhen Wang 21d ago

Evaluating LLMs for Demographic-Targeted Social Bias Detection: A Comprehensive Benchmark Study

Benchmark for evaluating LLMs on detecting demographic-targeted social biases across diverse content types and demographics.

Ax Hsien-Chin Lin, Benjamin Matthias Ruppik, Carel van Niekerk, Chia-Hao Shen, Michael Heck, Nurul Lubis, Renato Vukovic, Shutong Feng, Milica Ga\v{s}i\'c 21d ago

Prompt reinforcing for long-term planning of large language models

Method to improve LLM performance in multi-turn conversations by reinforcing long-term planning and goal tracking through prompting.

Ax Zhiyu Wang, Bingxin Zhou, Jing Wang, Yang Tan, Weishu Zhao, Pietro Li\`o, Liang Hong 21d ago

Fast and Interpretable Protein Substructure Alignment via Optimal Transport

Protein structure alignment using optimal transport for identifying and comparing local structural motifs.

Ax Gaoxiang Huang, Songning Lai, Yutao Yue 21d ago

Mitigating Spurious Background Bias in Multimedia Recognition with Disentangled Concept Bottlenecks

Lightweight Disentangled Concept Bottleneck Model addressing bias in input-to-concept mapping for interpretable multimedia recognition.

Ax Xi Zhang, Hanwei Zhu, Yan Zhong, Jiamang Wang, Weisi Lin 21d ago

BADiff: Bandwidth Adaptive Diffusion Model

Framework enabling diffusion models to adapt generation quality based on real-time network bandwidth constraints in cloud-to-device scenarios.

Ax Junpei Komiyama, Kyoungseok Jang, Junya Honda 21d ago

Rate-optimal Design for Anytime Best Arm Identification

Minimax optimal algorithm for best arm identification under fixed sampling budget with applications to A/B testing.

Ax Georgios Pantazis, Nicola Mignoni, Raffaele Carli, Mariagrazia Dotoli, Sergio Grammatico 21d ago

Adversarially and Distributionally Robust Virtual Energy Storage Systems via the Scenario Approach

Convex optimization framework for robust scheduling of aggregated EV battery storage under uncertainty.

Ax Bhuvan Sachdeva, Karan Uppal, Abhinav Java, Vineeth N. Balasubramanian 21d ago

Understanding Task Transfer in Vision-Language Models

Study of task transfer in Vision-Language Models examining how finetuning on one perception task affects performance on others.

Ax Austin Spizzirri 21d ago

The Specification Trap: Why Static Value Alignment Alone Cannot Produce Robust Alignment

Philosophical analysis arguing static value alignment approaches cannot ensure robust AI alignment under capability scaling and distribution shift.

Ax Brenda Anague, Bamdad Hosseini, Issa Karambal, Jean Medard Ngnotchouye 21d ago

Physics-Informed Neural Networks for Joint Source and Parameter Estimation in Advection-Diffusion Equations

PINNs applied to source inversion in advection-diffusion equations with sparse measurements for scientific computing.

Ax Jonathan Rystr{\o}m, Zihao Fu, Chris Russell 21d ago

OxEnsemble: Fair Ensembles for Low-Data Classification

OxEnsemble: Fair classification approach for low-data, imbalanced settings with demographic group constraints.

Ax Kohei Nishikawa, Koki Shimizu, Hiroki Hashiguchi 21d ago

Evaluating Singular Value Thresholds for DNN Weight Matrices based on Random Matrix Theory

Method for determining singular value thresholds in DNN weight compression using random matrix theory.

Ax Ayrat Abdullin, Umair Bin Waheed, Leo Eisner, Naveed Iqbal 21d ago

Parameter-Efficient Transfer Learning for Microseismic Phase Picking Using a Neural Operator

Parameter-efficient transfer learning with neural operators for microseismic phase picking across varying signal conditions.

Ax Loris Schoenegger, Benjamin Roth 21d ago

Compact Example-Based Explanations for Language Models

Study on selecting minimal training data subsets for example-based explanations of language model predictions using influence estimation.

Ax Kyriakos Stylianopoulos, Mattia Fabiani, Giulia Torcolacci, Davide Dardari, George C. Alexandropoulos 21d ago

Over-The-Air Extreme Learning Machines with XL Reception via Nonlinear Cascaded Metasurfaces

Wireless ML inference via programmable metasurfaces for over-the-air extreme learning machines in MIMO systems.

Ax Zhicheng Yang, Zhijiang Guo, Yinya Huang, Yongxin Wang, Wenlei Shi, Yiwei Wang, Xiaodan Liang, Jing Tang 21d ago

Accordion-Thinking: Self-Regulated Step Summaries for Efficient and Readable LLM Reasoning

Accordion-Thinking: Framework enabling LLMs to self-regulate reasoning step granularity through dynamic summarization for efficient inference.

Ax Antonin Sulc 21d ago

Differentiable Logical Programming for Quantum Circuit Discovery and Optimization

Neuro-symbolic framework using differentiable logic programming to design and optimize quantum circuits.

Ax Kimon Fountoulakis, David Mart\'inez-Rubio 21d ago

Complexity of Classical Acceleration for $\ell_1$-Regularized PageRank

Complexity analysis of accelerated proximal-gradient methods for ℓ1-regularized PageRank computation.

Ax Shivam Kumar, Yixin Wang, Lizhen Lin 21d ago

Flow Matching is Adaptive to Manifold Structures

Theoretical analysis of flow matching generative models' adaptation to data manifold structures.

Ax Haian Jin, Rundi Wu, Tianyuan Zhang, Ruiqi Gao, Jonathan T. Barron, Noah Snavely, Aleksander Holynski 21d ago

ZipMap: Linear-Time Stateful 3D Reconstruction via Test-Time Training

ZipMap: Stateful 3D reconstruction model achieving linear-time complexity for large image collections via test-time training.

Ax Kevin H. Guo, Chao Yan, Avinash Baidya, Katherine Brown, Xiang Gao, Juming Xiong, Zhijun Yin, Bradley A. Malin 21d ago

Stop Listening to Me! How Multi-turn Conversations Can Degrade LLM Diagnostic Reasoning

Evaluation of 17 LLMs showing diagnostic reasoning degrades across multi-turn conversations compared to single-turn benchmarks.

Ax Xiangyu Zeng, Qi Xu, Yunke Wang, Chang Xu 21d ago

HiCI: Hierarchical Construction-Integration for Long-Context Attention

HiCI: Hierarchical attention module for long-context language modeling, organizing information from local to global levels.

Ax Amuche Ibenegbu, Pierre Lafaye de Micheaux, Rohitash Chandra 21d ago

tBayes-MICE: A Bayesian Approach to Multiple Imputation for Time Series Data

tBayes-MICE: Bayesian approach to multiple imputation for time-series data with missing values via MCMC sampling.

Ax Yulin Zou, Yan Chen, Wenyan Chen, JooYoung Park, Shivaraman Nitin, Luo Tao, Francisco Romero, Dmitrii Ustiugov 21d ago

CodecSight: Leveraging Video Codec Signals for Efficient Streaming VLM Inference

CodecSight optimizes streaming vision-language model inference by leveraging video codec signals for end-to-end efficiency.

Ax Jaehyeok Lee, Xiaoyuan Yi, Jing Yao, Hyunjin Hwang, Roy Ka-Wei Lee, Xing Xie, JinYeong Bak 21d ago

Distributional Open-Ended Evaluation of LLM Cultural Value Alignment Based on Value Codebook

DOVE benchmark for evaluating LLM cultural value alignment using open-ended generation, addressing limitations of multiple-choice formats.

Ax Youssef Abdou 21d ago

The Unreasonable Effectiveness of Data for Recommender Systems

Study investigating saturation points in recommender system performance as training dataset size increases, with reproducible Python implementation.

Ax Asmaa Eldesoukey, Yongxin Chen, Abhishek Halder 21d ago

A Generalized Sinkhorn Algorithm for Mean-Field Schr\"odinger Bridge

Theoretical work on Schrödinger bridge problem with mean-field dynamics for multi-agent systems control.

Ax Hanyang Wang, Mingxuan Zhu 21d ago

The Detection-Extraction Gap: Models Know the Answer Before They Can Say It

Research: Chain-of-thought models generate 52-88% of tokens after answers are already recoverable, revealing inefficiency in reasoning.

Ax Yanan Cao, Ashish Ranjan, Sinduja Subramaniam, Evren Korpeoglu, Kaushiki Nag, Kannan Achan 21d ago

CASE: Cadence-Aware Set Encoding for Large-Scale Next Basket Repurchase Recommendation

ML research on cadence-aware encoding for next-basket recommendation in retail, modeling repurchase timing patterns.

HN chirsz 21d ago

JSON with Commas and Comments

JSON extension format supporting trailing commas and comments. Minor format specification unrelated to AI.

HN kotobuki 21d ago

Some LLM routers are injecting malicious tool calls

Brief mention of LLM routers injecting malicious tool calls as security issue. Insufficient detail.

HN omegaproto 21d ago

.

Bittensor governance controversy and token price impact.

HN stjuan627 21d ago

A few thoughs about AI videos

Personal experience report on using AI video generation models. Notes Gemini Nano, Veo3, Sora2, and Chinese models. Anecdotal observations.

HN obilgic 21d ago

Next SaaS replacement is an agent with a dashboard – kern

Analysis of AI agents as SaaS replacement with integrated database, logic, and UI.

HN vamshidhar199 22d ago

Show HN: Ctxlint – Lint your AGENTS.md for stale refs and token waste

Linter tool for AI agent context files and MCP configs. Detects stale references, token waste, hardcoded secrets. Cites research on context bloat reducing agent performance.