Limitations of public chest radiography datasets for artificial intelligence: label quality, domain shift, bias and evaluation challenges [0.03%]
公共胸片数据集在人工智能应用中的局限性：标签质量、领域差异、偏差和评估挑战
Amy Rafferty,Ajitha Rajan
Amy Rafferty
Artificial intelligence (AI) has shown significant promise in chest radiography, where deep learning models can approach radiologist-level diagnostic performance. Progress has been accelerated by large public datasets, such as MIMIC-CXR, Ch...
Alexander Bastounis,Desmond J Higham,Ivan Tyukin
Alexander Bastounis
Over the past decade, there has been an explosion of activity in the design of algorithms for adversarially attacking artificial intelligence (AI) systems; especially in the context of image classification. For example, a carefully crafted ...
Vulnerability analysis of transformer-based optical character recognition to adversarial attacks [0.03%]
基于变压器的光学字符识别对抗攻击脆弱性分析
Lucas Beerens,Desmond J Higham
Lucas Beerens
We present a novel framework to assess the resilience of state-of-the-art transformer-based optical character recognition (TrOCR) models. In this way, we develop new untargeted and targeted attack algorithms. On a benchmark handwriting data...
Online safety by design: towards user-centric development of safer AI-based systems [0.03%]
以人为本设计的线上安全问题：基于人工智能系统的安全开发措施
Thomas Baldwin-McDonald,Amir Fard,David Fletcher et al.
Thomas Baldwin-McDonald et al.
In this opinion piece, we present the view that the safety and security of end users should be placed at the heart of the design and development of any AI-based system deployed on an online service to ensure that risks of harm are minimized...
Mitigating medical bias in large language models by prompt engineering: an empirical study of effectiveness and trade-offs [0.03%]
通过提示工程减轻大型语言模型中的医学偏见：有效性与权衡的实证研究
Ying Xiao,Zhenpeng Chen,Jie Zhang
Ying Xiao
Large language models (LLMs) demonstrate expert-level performance in various medical scenarios, yet their outputs can exhibit bias against groups or individuals with specific sensitive attributes, posing risks to patient safety and undermin...
Implicit regularization makes overparameterized asymmetric matrix sensing robust to perturbations [0.03%]
隐式正则化使过参数化的非对称矩阵感知具有鲁棒性以应对扰动
Johan Sokrates Wind
Johan Sokrates Wind
Several key questions remain unanswered regarding overparameterized learning models. It is unclear how (stochastic) gradient descent finds solutions that generalize well, and in particular the role of small random initializations. Matrix se...
Tamara G Kolda
Tamara G Kolda
We consider the reality of deploying artificial intelligence (AI) in safety-critical systems, such as autonomous vehicles, medical diagnoses and weather forecasting. Our discussion is grounded in the mathematical nature of AI systems, inclu...
Can model attribution bridge AI's accountability gap in safety-critical domains? [0.03%]
模型溯源能否弥合AI安全关键领域责任缺口？
Marc Juarez
Marc Juarez
The current trend of deploying machine learning models as remote services obscures which specific models users are interacting with, exacerbating an accountability gap that becomes ever more pressing as these technologies make their way int...
Trustworthiness in AI: on SciCompBot-the scientific computing chatbot [0.03%]
人工智能的可靠性：以SciCompBot科学计算聊天机器人为例
Anders Christian Hansen,Fabian Circelli
Anders Christian Hansen
This paper explores the prospect of a scientific computing chatbot (SciCompBot): an artificial intelligence (AI) system designed to address advanced problems in optimization, spectral computations (eigenvalues, eigenvectors, spectra), diffe...
Inverse FIP effect plasma in the solar atmosphere: a synthesis of current understanding and new insights from AR 11967 [0.03%]
太阳大气中的逆FIP效应等离子体：目前认识的总结和新视野活动区11967的新见解
Deborah Baker,David H Brooks,David M Long et al.
Deborah Baker et al.
Wave-plasma interactions and energy transport are fundamental processes in stellar atmospheres, shaping elemental composition through the first ionization potential (FIP) and inverse FIP (IFIP) effects. Although stellar measurements provide...