CCB Seminar Series

Speaker: Glenn Ko, SEAS Professor, Co-Founder and CEO of Stochastic AI

When: Thursday, April 11, 11:00 AM ET

Topic: From Harvard Lab to Global Enterprises: The Journey of Stochastic’s Personalized AI

Zoom: Click here

Description: Glenn’s founding story of Stochastic AI as a Harvard academic, followed by some technical details behind Stochastic AI.

Bio: Glenn Ko is an AI systems researcher and entrepreneur focusing on building AI systems that prioritize personalization, privacy, and real-time interaction. He is currently the Co-Founder and CEO of Stochastic, an AI company spun off of his postdoctoral research at Harvard University School of Engineering and Applied Sciences, where he still holds an adjunct research faculty position. Stochastic provides personalized AI agents for automation of decision making and task execution in the realm of enterprise knowledge work such as customer support and internal document processing. He received his PhD on probabilistic computing from UIUC and spent time at IBM Research, Qualcomm and Samsung Electronics where he worked on AI/ML research and contributed to the design of mobile processors for smartphones.

Speaker: Micaela Consense, PhD Candidate, Vector Institute and University of Toronto

When: Thursday, April 11, 11:00 AM ET

Topic: To Transformers and Beyond: Large Language Models for the Genome

Zoom: Click here

Abstract:
In the rapidly evolving landscape of genomics, deep learning has emerged as a useful tool for tackling complex computational challenges. This review focuses on the transformative role of Large Language Models (LLMs), which are mostly based on the transformer architecture, in genomics. Building on the foundation of traditional convolutional neural networks and recurrent neural networks, we explore both the strengths and limitations of transformers and other LLMs for genomics. Additionally, we contemplate the future of genomic modeling beyond the transformer architecture based on current trends in research. The paper aims to serve as a guide for computational biologists and computer scientists interested in LLMs for genomic data. We hope the paper can also serve as an educational introduction and discussion for biologists to a fundamental shift in how we will be analyzing genomic data in the future.

Speaker: Mariano Gabitto, Assistant Investigator, Allen Institute for Brain Science

When: Monday, March 18, 10:00 – 11:00 AM ET

Where: Countway Library, L1-032

Topic: Image processing, visual analytics, and data sharing for whole-slide imaging and spatial profiling of tissues and tumors

Zoom: Click here

Abstract:
Since their inception, single-cell technologies have revolutionized biology, providing exciting opportunities to characterize cellular populations and track their changes with disease. These technologies enable the analysis of the transcriptional (scRNA-seq),chromatin accessibility (scATAC-seq), and surface-protein expression (CITE-seq) landscape of individual cells. How can we effectively discover structure in this complex high-dimensional multimodal data and integrate it into a common representation to better understand disease progression? I will introduce MultiVI, a deep generative model designed for the analysis of single-cell chromatin, transcriptional and protein expression information. MultiVI co-learns informative low-dimensional representations for individual and joint modalities, accurately capturing multimodal properties of individual cells. This approach offers a principled method to analyze samples in which single and multiple genomic data modalities are present. Finally, I will demonstrate how this tool was used within the Seattle Alzheimer’s Disease Cell Atlas (SEA-AD) Consortium to map cellular types in healthy and diseased stated and study their changes associated with Alzheimer’s Disease.

Bio:
Mariano Gabitto is an Assistant Investigator at the Allen Institute for Brain Science. Mariano’s research focuses on applying statistical and machine learning methods to decipher the cellular and molecular programs giving rise to the cellular diversity observed in the brain and how cell types degenerate during neurological diseases. Before joining Allen, he was a research scientist at the Center for Computational Biology at the Flatiron Institute, Simons Foundation and at the Broad Institute of Harvard and MIT. Marianowas a visiting scientist at Mike Jordan’s Group at U.C. Berkeley, developing deep generative models to represent single-cell information. He completed his Ph.D. in Neuroscience at Columbia University in the laboratory of Charles Zuker where he establishedcollaborations with the groups of Liam Paninski, Larry Abbott

Speaker: Jeremy Muhlich, Principal Associate, Director of Software Engineering at the Harvard Medical School Laboratory of Systems Pharmacology

When: Monday, March 11, 10:00 – 11:00 AM ET

Where: Gordon Hall, 106 Waterhouse  Conference Room

Topic: Image processing, visual analytics, and data sharing for whole-slide imaging and spatial profiling of tissues and tumors

Zoom: Click here

Speaker: Stefano M. Iacus, Senior Research Scientist,Director of Data Science and Product Research at the Institute for Quantitative Social Science, Harvard University

When: Monday, Feburary 26, 10:00 – 11:00 AM ET

Where: Countway 518 Lahey Conference Room

Topic: Bringing compute close to large data at Harvard Dataverse

Zoom: Click here