Can AI See Bias in X-ray Images?

Abstract

Article Can AI See Bias in X-ray images? Kwasniewska Alicja *, and Szankin Maciej * 1 SiMa Technologies Inc. 226 Airport Parkway, Suite 550 San Jose, CA 95110, United States 2 Intel Corporation 16409 W Bernardo Dr Suite 100, San Diego, CA 92127, United States * Correspondence: alicja.kwasniewska@sima.ai; maciej.szankin@intel.com Received: 8 November 2022 Accepted: 13 November 2022 Published: 22 December 2022 Abstract: Recent advances in artificial intelligence (AI) have shown promising results in various image-based systems, improving accuracy and throughput, while reducing latency. All these factors are crucial in healthcare and have generated increased interest in this technology. However, there are also multiple challenges integrating AI in existing systems, such as poor explainability, data imbalance and bias. These challenges affect the reliability of the neural networks used in AI applications. The limitations may significantly affect the quality and cost of medical care by introducing false positive diagnosis. The false positives subsequently lead to increased stress in patients and necessitate additional testing and procedures. Lack of rich data representing all socio-economic groups can also undermine reliable decisions for underrepresented groups. Although various studies discussed techniques that may help with bias mitigation, to the best of our knowledge, no practical experiments have been conducted so far that compare different reweighting approaches using convolutional neural networks (CNN). This work focuses on in-depth explanatory analysis of chest X-ray datasets to understand and quantify the problem of class imbalance and bias. After that, various topologies of binary classifications are compared, followed by practical applications of loss reweighting techniques and comparison of their influence of privileged, underprivileged, and overall population. Experiments proved that high classification accuracy can be achieved using an efficient model topology suitable for embedded devices, making it possible to run locally without the need for cloud processing. Preliminary results showed that performance of the model for the underprivileged class can be improved by 15% if proper weighting factors are obtained and applied during the training procedure.

A Unified Brightness Temperature Features Analysis Framework for Mapping Mare Basalt Units Using Chang’e-2 Lunar Microwave Sounder (CELMS) Data

“…Recently, machine learning models have been gradually introduced into lunar and planetary research [52][53][54]. Further studies will explore the potential of more advanced machine learning models with regard to lunar brightness temperature analysis [55][56][57].…”

Differentiable channel pruning guided via attention mechanism: a novel neural network pruning approach

“…Especially, attention scores of individual filters in a layer are recorded in stage one and used to guide the optimization in stage two tions. Moreover, SE block is an effective attention block that can be flexibly embedded into most existing network structures and, consequently, the SE block has been widely used in computer vision applications [45]. An SE block contains two parts in its structure, namely, squeeze and excitation.…”

Is High-Fidelity Important for Human-like Virtual Avatars in Human Computer Interactions?

“…Artificial Intelligent (AI) technology, which makes virtual avatars behave more human-like, especially the ability of reacting like a real human during interactions, has been fast becoming the main focus in HCI in recent years. AI agents, which can act like real humans to achieve a higher level of user engagement, are used not only in games [44] but also in other applications such as education, training [45], and health care [46]. As a result, the intelligence of virtual avatars is also becoming an important part of character design.…”