Quick Buzz Feed

New learning space opening to deepen understanding, problem-solving and reflection in artificial intelligence

Gary Lloyd | Sep 21,26 | 01:32 EST

Technology

Marquette University has announced the opening of a new, state-of-the-art learning and research facility, the AI Discovery Hub. This innovative space, spearheaded by the Opus College of Engineering and located in Haggerty Hall room 268, is designed to provide immersive, hands-on experiences for both students and university employees. Its core mission is to empower individuals to develop essential skills, build confidence, and deepen their understanding of cutting-edge artificial intelligence tools and decision-making processes. The hub will serve as a dynamic environment where human curiosity and problem-solving converge with advanced AI and machine learning technologies, addressing critical topics such as assessing AI bias and reliability, analyzing complex, unstructured MRI data, processing high-dimensional images, and enhancing environmental monitoring. Through a blend of open exploration, integrated course visits, and practical demonstrations, the AI Discovery Hub fosters both guided and self-directed learning across various subfields of artificial intelligence. It aims to prepare ethical leaders who can thoughtfully shape the future of AI for positive societal impact across diverse communities and industries. The hub officially opens its doors for visitors and events on Monday, September 28.

Workstation 1: Assessing computer vision confidence and healthcare applications This workstation provides an interactive platform to delve into the intricacies of computer vision, focusing on object detection and the processing of visual data, with a particular emphasis on its applications within the healthcare sector. Visitors will engage with two distinct sample projects. The first project involves a comprehensive exploration of image detection technology, highlighting the critical ongoing efforts to assess and ensure trust, confidence, and the identification of potential data biases within these complex image detection systems. The second project shifts focus to a practical application, demonstrating how AI-powered image detection and analysis can be effectively utilized in diabetic retinopathy screenings. This showcases AI's potential to not only support but also significantly scale up human-centered clinical reviews, improving efficiency and accessibility in medical diagnostics. Workstation 2: Spectral & hyperspectral imaging to extend beyond human vision At this workstation, visitors will gain an understanding of spectral and hyperspectral imaging, technologies that extend the boundaries of human vision by measuring light wavelengths typically undetectable by standard cameras and the human eye. This advanced capability allows for the revelation of additional, nuanced information and subtle differences between objects that would otherwise remain unseen. Given that hyperspectral images capture hundreds of unique measurements at every single pixel, artificial intelligence becomes an indispensable tool. AI is leveraged not only to efficiently sense and process these immense datasets but also to intelligently recognize intricate patterns within them. This enables the technology to provide profound insights, particularly in distinguishing various materials. The practical applications of this technology are far-reaching, encompassing vital sectors such as food quality assessment, modern agriculture, sophisticated medical diagnostics, precision manufacturing processes, and comprehensive environmental monitoring. Workstation 3: Investigating Large Language Model (LLM) bias and reliability: healthcare and legal domains This workstation offers a critical examination of Large Language Models (LLMs), specifically addressing issues of reliability, accuracy, and inherent biases present in their data collection and generated outputs, particularly within sensitive healthcare and legal contexts. Two compelling projects are presented for exploration. The first project invites participants to compare actual judicial opinions with summaries generated by AI, allowing them to precisely identify discrepancies between LLM outputs and human legal analysis, and to understand instances where errors might occur. This hands-on comparison serves to illuminate both the impressive capabilities and the current limitations of LLMs in nuanced, high-stakes fields. The second project delves into the continuous analysis of LLM systems currently deployed in healthcare. It aims to elucidate the methodologies and rationales behind measuring bias in these systems, highlighting the broader, imperative drive for developing ethical AI. Furthermore, it underscores the extensive research required to establish trust and confidence in LLMs when applied to critical healthcare decision-making and patient care. Workstation 4: AI and vast datasets in healthcare: aphasia care and brain/spinal cord MRIs Focused on the transformative potential of artificial intelligence in addressing complex healthcare challenges that necessitate the analysis of vast and often intricate datasets, this workstation offers deep insights. It features two impactful projects. The first project explores how advanced LLMs, coupled with sophisticated speech information processing and brain signal data analysis, can collectively be utilized to significantly support individuals facing communication challenges related to aphasia. This research aims to develop innovative future assistive tools that can greatly benefit both patients and their caregivers, enhancing communication and quality of life. The second project investigates how AI systems are uniquely positioned to assist in the analysis and detection of subtle patterns within the enormous and often unstructured datasets generated by brain and spinal cord MRIs. This investigation highlights how AI can dramatically accelerate the derivation of critical insights from medical imaging, particularly in scenarios where the sheer volume and lack of structured organization in data render human analysis impractical or exceedingly slow.

Workstation 1: Assessing computer vision confidence and healthcare applications

This workstation provides an interactive platform to delve into the intricacies of computer vision, focusing on object detection and the processing of visual data, with a particular emphasis on its applications within the healthcare sector. Visitors will engage with two distinct sample projects. The first project involves a comprehensive exploration of image detection technology, highlighting the critical ongoing efforts to assess and ensure trust, confidence, and the identification of potential data biases within these complex image detection systems. The second project shifts focus to a practical application, demonstrating how AI-powered image detection and analysis can be effectively utilized in diabetic retinopathy screenings. This showcases AI's potential to not only support but also significantly scale up human-centered clinical reviews, improving efficiency and accessibility in medical diagnostics.

Workstation 2: Spectral & hyperspectral imaging to extend beyond human vision

At this workstation, visitors will gain an understanding of spectral and hyperspectral imaging, technologies that extend the boundaries of human vision by measuring light wavelengths typically undetectable by standard cameras and the human eye. This advanced capability allows for the revelation of additional, nuanced information and subtle differences between objects that would otherwise remain unseen. Given that hyperspectral images capture hundreds of unique measurements at every single pixel, artificial intelligence becomes an indispensable tool. AI is leveraged not only to efficiently sense and process these immense datasets but also to intelligently recognize intricate patterns within them. This enables the technology to provide profound insights, particularly in distinguishing various materials. The practical applications of this technology are far-reaching, encompassing vital sectors such as food quality assessment, modern agriculture, sophisticated medical diagnostics, precision manufacturing processes, and comprehensive environmental monitoring.

Workstation 3: Investigating Large Language Model (LLM) bias and reliability: healthcare and legal domains

This workstation offers a critical examination of Large Language Models (LLMs), specifically addressing issues of reliability, accuracy, and inherent biases present in their data collection and generated outputs, particularly within sensitive healthcare and legal contexts. Two compelling projects are presented for exploration. The first project invites participants to compare actual judicial opinions with summaries generated by AI, allowing them to precisely identify discrepancies between LLM outputs and human legal analysis, and to understand instances where errors might occur. This hands-on comparison serves to illuminate both the impressive capabilities and the current limitations of LLMs in nuanced, high-stakes fields. The second project delves into the continuous analysis of LLM systems currently deployed in healthcare. It aims to elucidate the methodologies and rationales behind measuring bias in these systems, highlighting the broader, imperative drive for developing ethical AI. Furthermore, it underscores the extensive research required to establish trust and confidence in LLMs when applied to critical healthcare decision-making and patient care.

Workstation 4: AI and vast datasets in healthcare: aphasia care and brain/spinal cord MRIs

Focused on the transformative potential of artificial intelligence in addressing complex healthcare challenges that necessitate the analysis of vast and often intricate datasets, this workstation offers deep insights. It features two impactful projects. The first project explores how advanced LLMs, coupled with sophisticated speech information processing and brain signal data analysis, can collectively be utilized to significantly support individuals facing communication challenges related to aphasia. This research aims to develop innovative future assistive tools that can greatly benefit both patients and their caregivers, enhancing communication and quality of life. The second project investigates how AI systems are uniquely positioned to assist in the analysis and detection of subtle patterns within the enormous and often unstructured datasets generated by brain and spinal cord MRIs. This investigation highlights how AI can dramatically accelerate the derivation of critical insights from medical imaging, particularly in scenarios where the sheer volume and lack of structured organization in data render human analysis impractical or exceedingly slow.