💻 Category

cs

2608 papers

💻 computer science

Rank-Gated and Sparse Low-Rank Adaptation for Efficient Fine-Tuning of Vision--Language Models

This paper introduces Rank-Gated LoRA (RG-LoRA), a method that assigns learnable importance weights to LoRA rank components to enable post-training pruning for memory efficiency, but initial experiments on Qwen3-VL-8B show that without explicit sparsity regularization, the gates fail to learn meaningful sparsity, resulting in negligible latency or VRAM gains despite validating the proposed train-prune-evaluate mechanism.

Qinwu Xu2026-09-17
💻 computer science

Ideology by Alphabet: Option Order and the Measurement of Machine Political Preferences

This paper demonstrates that the apparent political ideologies of large language models are largely artifacts of fixed answer-order biases rather than genuine preferences, as randomized option ordering dissolves previously identified ideological clusters and reveals that specific slot arrangements can arbitrarily label models with conflicting political stances.

Tamas Olah, Laszlo Erdey, Tibor Tokes, Levente Nadasi2026-09-17
💻 computer science

Explainable Machine Learning for Extreme Weather Event Prediction Using Long-Term Environmental Measurements

This study presents an explainable multi-tower machine learning framework using 15-minute meteorological data from six Oak Ridge Reservation towers to forecast five types of extreme weather events, demonstrating that gradient boosting models outperform deep learning and CNN architectures while identifying key tower-based and physics-based features that drive prediction accuracy across varying time horizons.

Jose Tupayachi, Ziwei Liu, Kevin Birdwell, Xueping Li, Xiao-Ying Yu2026-09-17
💻 computer science

Context-Aware Fine-Grained Ranking of Art Exam Works Under Shared Evaluation Conditions

This paper addresses the challenge of fine-grained ranking in art examinations by introducing a new dataset organized by exam groups and proposing a group-conditioned framework that leverages dual-path reference encoding and adaptive routing to capture context-aware relative quality differences among visually similar submissions.

Jiahao Li, Yingjie Zhang, Zhongwei Huang, Haoze Chen, Zongming Tan, Baohua Tan, Chao Chen2026-09-16
💻 computer science

LLMs for Survey Text Analysis – A Performance Comparison Between Humans and GPT-5 on Inductive Content Analysis

This study demonstrates that GPT-5.4 can approximate human performance in inductive content analysis of survey data, achieving coding and theme generation agreement levels comparable to human internal consistency, thereby validating its potential as a scalable support tool for qualitative research.

Leonardo Bergmann, Renata Gheorghiu, Ana Gvritishvili, Alex Mican, Chris Stewart, Topias Tolonen-Weckström2026-09-16
💻 computer science

A Risk-Based Framework for Hardening Active Directory Infrastructure Against Credential Theft and Privilege Escalation

This paper proposes a Risk-Based Active Directory Hardening Framework (RB-ADHF) that integrates five security domains and aligns with NIST and CIS standards to provide a structured, measurable approach for assessing and improving enterprise identity infrastructure against credential theft and privilege escalation.

NAWAF ABDULJALIL ABDULAZIZ ABDULNOOR, RANI ABDULJALIL ABDULAZIZ ABDULNOOR, AHMED RAZZAZ AHMED QAID2026-09-16