Erik Amézquita

PFFIE Postdoctoral Fellow. University of Missouri---Columbia (MU), USA.

Date: Friday, July 19, 2024
TIme: 10:00 AM PST / 12:00PM CST.

Shape is foundational to biology. Observing and documenting shape has fueled biological understanding as the shape of biomolecules, cells, tissues, and organisms arise from the effects of genetics, development, and the environment. Less often we do consider that data itself has shape and structure, or that it is possible to measure the shape of data and analyze it. To comprehensively measure shape in a broad scope of different datasets, we will turn to Topological Data Analysis (TDA).
This webinar will focus on the use of TDA characterize and model the spatial distribution of different cytosolic transcripts between individual soybean nodule cells. This modeling can offer exciting insights into how exactly protein production is regulated at the individual cell level.

Valerio Hoyos-Villegas

Assistant Professor in Plant Breeding and Genetics in the Department of Plant Science at McGill University

Date: May 17, 2024
Time: 10:00 AM PST.

Despite progress in multiple areas, measuring phenotypes in plant breeding remains a bottleneck in the crop improvement cycle. This becomes particularly true when dealing with complex trait dissection. In the coming years, a revolution in plant breeding will be realized thanks to advanced in photo optic, ultrasonic, capacitive sensors, etc. This seminar focuses on the work that the Pulse Breeding and Genetics laboratory carries out in the advancement of phenomics and genomics for the acceleration of genetic gain for the development of pulse crop varieties.

Alexander Bucksch, Ph.D.

Associate Professor - School of Plant Sciences.
The University of Arizona

Date: April 23, 2024.
TIme: 10:00 AM PST.

A plant's history can often be inferred from its shape phenotype. This is especially evident in roots, which demonstrate significant plastic responses in their architecture to changing environmental conditions during development. Machine learning and artificial intelligence provide methodologies to organize the complex shape signals detected by sensors, enabling the identification of simple and understandable rules governing root architecture phenotypes. From a shape-oriented perspective, this talk introduces five applications of traditional machine learning and artificial intelligence to measure and understand root architecture both in excavated roots and within the soil.

Jennifer Clarke, Ph.D.

Professor, Departments of Statistics, Food Science and Technology.
University of Nebraska-Lincoln

Date: March 15, 2024
Time: 1:00 PM EST (10:00 AM PST)

In this presentation we will discuss two different philosophies behind scientific modeling that emphasize either simplicity or complexity, respectively. This tension underlies the approach of statisticians to predictive modeling. We will discuss how this tension is handled through a statistical lens with an example from regularization in regression contexts. We will then present how deep learning methods appear to take a different approach, explicitly ignoring this tension with (surprisingly) good results. We follow this with a hypothesis that may explain these results and the associated costs of modeling in high dimensions. We conclude with some challenges to the adoption and use of AI for modeling, with references for further study.