This is a past event.
Dr. Neerav Kaushal, a deep learning scientist at Flagship Pioneering in Boston, MA, will present at this week's Physics Colloquium.
Kaushal's presentation is titled "Machine Learning in Drug Discovery".
The seminar will be presented at 4:00 p.m. on Thursday (Oct. 31) in Fisher 139. The coffee hour will be held at 3:30 in the Fisher Hall Lobby.
Abstract: The design space of drug-like small molecules is of the order of 10^63. Optimizing this parameter space is a computationally intensive and near impossible challenge because it is large, discrete, and unstructured. However, recent advancements in machine learning, particularly deep learning, have enabled the development of continuous, data-driven molecular representations, transforming complex molecular structures into machine-readable formats. In this talk, I will discuss how I conceptualize, prototype, and build machine and deep learning frameworks capable of understanding, navigating, and optimizing this immense design space to discover and design novel drug candidates. I will also discuss how I leverage generative machine learning to generate novel RNAs optimized for specific functions. The foundational models grasp the chemical language of drugs and the biological language of RNAs, leading to the development of fine-tuned models that accelerate drug discovery and delivery. The recent application of machine learning to drug discovery benefits pharmaceutical researchers, chemists, and computational biologists by speeding up and optimizing the process, ultimately benefiting patients through the development of innovative and targeted therapies. Lastly, I will discuss my transition from academia to industry, highlighting how this journey shaped and refined my professional skills along the way.
Bio: Dr. Kaushal is currently a deep learning scientist at Flagship Pioneering in Boston, where he applies machine and deep learning algorithms to discover and design novel small molecule and lipid-based drugs for targeted delivery. He earned his Master in Physics from Panjab University in 2017 and later his PhD in Engineering Physics from Michigan Technological University in 2022, where he worked at the intersection of machine learning and cosmology to study and simulate the evolution of the large-scale structure of the Universe. In addition to drug discovery and cosmology, his interdisciplinary research extends across multiple fields, including astroparticle physics, where he worked on relativistic image doubling; geoinformation sciences, where he developed software pipelines to study the relationship between aerosol optical depth and PM2.5; and structural biology, where he optimized novel RNA components for functional enhancement.
0 people added
No recent activity