Dr. Jasmine Berry: Designing (Artificial) Intelligence
Recent graduate of the University of Southern California’s Computer Science PhD Program with a focus on Artificial Intelligence & Brain Theory.
Data Science Advisor at Intake Health.
CTO at a diabetes-focused startup.
How should you engineer a robot’s brain?
Dr. Jasmine Berry, PhD, is a research scientist trying to answer this question by investigating self-aware computing for autonomous systems. Dr. Berry is designing systems for a future where prosthetic devices, wearable systems, micro machines, and even home appliances have a simulated neural system. She is researching how systems can think with dynamic self-representation using inspiration from self-awareness in humans.
“If I tell you to close your eyes, and touch your nose with your hand,” says Dr. Berry, “You would be able to do that quite feasibly, without any difficulty, even without actually looking at your hands and limbs. You know where your body is in space.” The brain estimates where the arm is in space and predicts the best action to bring your hand to your nose. This ability is made possible by proprioception, the sense of self-movement and body position, which transmits neural signals via muscle and tendons. Dr. Berry published a paper that presents a computational model for sensory “gestalt,” a German word that refers to the totality of an experience, that is inspired by the sensory and motor pathways and signals that allow the human brain to know where the body is in space without even seeing it.
Berry and Valero-Cuevas’ Minimal Self
In Berry and Valero-Cuevas' representation of the minimal self, we see the self as a collection of categorized states in n-dimensional space formed from sensory and motor maps. This data-driven projection method categorizes the set of feasible inputs from muscle spindles for each specific task performed (i.e., arm reach, sit, squat, standing) as a manifold. Transitions from one state to another occur through point-to-point transitions along the manifold. The high-dimensional space of afferent modalities has an underlying structure given by the anatomy of the body and the physical transitions it can undergo such as changing postures via self-movement.
Building Computing Systems for Diabetic Patients
One of the most surprising things that I've encountered throughout my journey of becoming a better scientist and engineer, is that inspiration and motivation can come from anywhere in any place and from anybody, and not just from people who are established professors, or what you would consider as maybe academic elites. I have received a lot of inspiration and motivation for my research, for my direction on what I wanted to do with my career from neighbors, from my friends, from strangers. - Jasmine Berry
In addition to being a research scientist, Dr. Berry is a data science advisor for Intake Health and also the tech lead and software developer of a startup addressing the needs of individuals with diabetes. She is developing a personalized medicine platform driven by AI. The platform provides precision health and wellness services for diabetic patients and patients who are suffering from different co-morbidities.
The CDC has projected that by 2030, one in three people will be diabetic or pre-diabetic. Dr. Berry herself has Type I diabetes and is a member of the Juvenile Diabetes Research Foundation. As she connects with more members of the diabetic community, she continues to see sees how diabetics struggle to discover a diet that controls their blood sugar levels.
Doctors currently try to address the needs of patients by giving generalized dietary recommendations, however, as Dr. Berry says, “…the patients will often leave their appointment, try these recommendations, and play a game of trial and error to figure out how much insulin they should take or what type of foods they should eat. And often times that trial-and-error process can really wreak havoc on their blood sugar levels.”
However, AI may allow Dr. Berry to quickly create quality personalized recommendations for patients. “AI tools coming out, in the form of machine learning and deep learning, might reveal those hidden key insights that cannot necessarily be seen by doctors right away.”
As Artificial Intelligence (AI) increases in popularity, it’s crucial that we have black leaders in the field. Dr. Jasmine Berry is one of these leaders. The field of AI is changing quickly and there is constant innovation. Dr. Jasmine Berry stays focused by knowing her end goal and taking a top-down approach. After she knows her goal, “I start trying to look for the pieces that are necessary in order for me to get to that step, or to that level.” To learn more about her method she recommends reading “The Art of Doing Science and Engineering” by Richard Hamming.
Dr. Berry tells us, “You have the opportunity to make this journey your own, and many people want to see you succeed.” She reminds us that relationships and inspiration can be found anywhere. She provides some words of wisdom for others in or interested in AI, “In this field, it gets tough. No journey that's worthwhile is going to be easy and my advice to you all is to continue persevering and surround yourself with a supportive network. Do that with your colleagues, do that with people who are your family.”
Dr. Berry supports future black leaders in AI and robotics by volunteering with Black In Robotics. To learn more, visit blackinrobotics.org or subscribe and follow Black in Robotics on Twitter and Instagram.
References
Berry, J. A., & Valero-Cuevas, F. J. (2020, July). Sensory-Motor Gestalt: Sensation and Action as the Foundations of Identity, Agency, and Self. In Artificial Life Conference Proceedings (pp. 130-138). MIT Press.
