UMass Boston

Artificial intelligence (AI) is ushering in a new era, revolutionizing fields such as protein structure prediction and medical imaging with groundbreaking advancements.

The broader scientific community is recognizing AI's potential to revolutionize science, opening new pathways, sparking innovative inquiries, and driving profound leaps forward with significant societal impact. 

AI and Scientific Research 

In my research, I focus on AI-enabled scientific discoveries. AI can be leveraged to drive breakthroughs in science, and scientific principles can be applied to advance AI methodologies. The convergence of these fields accelerates innovation. 

I am one of the principal investigators for a National Institutes of Health-funded Research Project Grant (R01) study with Exercise Physiology Professor Scott Crouter from the University of Tennessee Knoxville. Professor Crouter, our student researchers, and I are focusing on how AI and machine learning (ML) can be used to measure physical behavior in adults as they go about their day in their natural environments. Through our collaboration, we are creating highly accurate methods for measuring physical activity based on physics-driven AI algorithms guided by physical knowledge. 

It has become common to measure steps with our phones or to use wearable devices to track how many flights of stairs we climb, miles we run, and the number of hours we sleep. Have you ever wondered how devices detect and differentiate activities such as running versus jumping? The key to this starts with data from two types of sensors, accelerometers and gyroscopes. An accelerometer tracks how fast people's movements are speeding up or slowing down as they move side to side, up and down, and forward and backward. A gyroscope measures how quickly and in which direction an object is rotating. 

Our research explores ways to develop and validate AI and ML algorithms using raw accelerometer and gyroscope sensor data collected at the hip and wrist. The aim of this work is to find highly effective ways to translate unprocessed data detected by the sensors and to turn it into information that accurately describes the physical behaviors that took place. 

The development of AI/ML methodology to assess physical activity is a high priority for public health because it enables health care providers to customize exercise prescriptions to closely align with what an individual is doing. These AI/ML algorithms can be used to inform innovations in the development of wearable devices. They can also benefit those examining dose response relationships between sedentary behaviors, physical activity, and health outcomes in adults. 

Ethical Considerations in AI Applications 

As we advance AI technologies in health monitoring and other sensitive areas, it's crucial to address ethical concerns such as privacy, consent, data security, and bias. Collecting and analyzing personal health data raises questions about how this information is stored, who has access to it, and how it may be used beyond its original intent. Ensuring that AI systems are designed with ethical principles in mind helps protect individual rights and fosters trust between users and technology. 

In our work, we prioritize 

• Data Privacy and Security: Implementing robust measures to safeguard personal information against unauthorized access and breaches 
• Informed Consent: Ensuring participants are fully aware of how their data will be used and have the option to opt out 
• Bias Mitigation: Developing algorithms that are fair and do not disproportionately affect any group based on race, gender, age, or other characteristics 
• Transparency and Explainability: Creating AI models whose decision-making processes can be understood and scrutinized by stakeholders 

AI Fluency: Preparing for Today’s Workforce 

Understanding both the opportunities and hazards of AI enables everyone to contribute to its informed, equitable, and ethical integration into society, maximizing benefits while managing risks. Recognizing AI's growing importance, UMass Boston established the Paul English Applied AI Institute (PEAAII) to unify the efforts around AI taking place throughout our campus.  

The PEAAII is driving efforts to incorporate AI into the curricula at UMass Boston; the initial focus is on supporting students as they develop their AI literacy, the PEAAII offers programs designed to move students beyond thinking of generative AI (GenAI) as a shortcut for doing assignments. Instead, our students are becoming skilled at leveraging AI as a tool for increasing their creative, analytical, and problem-solving skills. For example, in an “AI for All: Demystifying AI in Creative Problem Solving” session, Communications Professor Gamze Yilmaz presents ways to use AI for creative problem solving. In the session, students learn about GenAI, ethical considerations, effective prompting techniques, and how to critically evaluate Al-generated content.  

The PEAAII is instrumental in preparing our students, and other members of the UMass Boston community, for a world in which AI will inevitably continue to grow and shape industries and daily life.

Soccer play while wearing the COSMED K5 to measure energy expenditure. Image By: Steven Bridges