Jan 16 - Ann Arbor AI, ML and Computer Vision Meetup

Register for the event.

Date and Time
Jan 16, 2025 from 5:30 PM to 8:30 PM

Location
The Collaboration Space at UMTRI, 2901 Baxter Rd, Ann Arbor, MI

Mcity Data Engine: From Data Stream to AI Model

The Mcity data engine is designed to train AI models on large volumes of raw sensor data. For our current use case, data is collected from roadside fisheye cameras that capture traffic at intersections. FiftyOne is used to curate the data and select diverse samples for labeling. The labeled dataset is then used to train object detectors/trackers/predictors of vehicles, bicyclists, and pedestrians.

About the Speaker

Daniel Bogdoll is a visiting research scholar at Mcity and PhD student at Karlsruhe Institute of Technology, Germany. He has been working in the field of mobility for 9 years and currently focuses on detecting unusual scenarios.

Building Better Visual AI Datasets for Mobility

Building datasets for training visual AI models is laborious, error-prone, and expensive. Even worse, we lack the ability to measure how good a dataset is for a certain task. And, yet, we spend immense amounts of capital on building datasets, especially the case in the automotive sector. This talk will explore two vectors of inquiry around improving our ability to build measurably high-quality datasets for training models. First, I will explore discuss new methods in pruning large data-lakes down to size for model training guided by the ability to estimate dataset difficulty. Then, I will describe the potential for AI-human teaming when building and annotating datasets, leveraging a human in the loop to guide auto-labeling models when labeling complex scene geometry for mobility datasets.

About the Speaker

Dr. Jason Corso is a professor of electrical engineering and computer science at the University of Michigan and co-founder and CEO of Voxel51, an AI software company creating dev tools for improving the performance of computer vision and machine learning systems.