ML+X Forum: Navigating Gravitational Waves with AI Insights
Machine Learning Community Event
We are looking forward to kicking off the spring ML+X forum on Tuesday, Feb. 13, 12-1pm (in-person and online)! We will start February’s forum with announcements on new initiatives in the community and a small group discussion activity to get everyone acquainted. In the second half of the hour, we’ll hear from Bella Finkel, a graduate student who is seeking input on AI explainability methods to better understand convolutional neural networks trained on spectrogram data (representing gravitational waves). Register by Feb. 9th (lunch provided) to guarantee your lunch ticket! - Chris
Welcome and small group discussions, Chris Endemann
For this discussion activity, we will do a round of introductions followed by a focused discussion of the following prompts: (a) Discuss applications of machine learning that can have a positive impact on society, business, or any specific domain; (b) How can we democratize machine learning tools and knowledge to empower a broader audience?; (c) How can ML practitioners effectively convey complex ideas to diverse audiences? (d) What resources do you use to learn machine learning tools and/or concepts
Classifying gravitational wave modes from core-collapse supernovae, Bella Finkel
Core-collapse supernovae are promising gravitational wave sources that play an important role in the evolution of galaxies and the formation of heavy elements. I will describe recently developed software for quickly generating large and diverse sets of synthetic gravitational wave signals and its use in an ongoing project that employs a convolutional neural network to classify gravitational wave signals from core-collapse supernovae. Our approach offers novel opportunities for controlling features of the input waveforms and the noise level of the dataset. At this ML+X forum, I’ll solicit feedback on strategies for interpreting the model to determine salient features in the input spectrograms and optimizing a model that works with inherently stochastic data.
Finding the Orchard View room: The Orchard View room is located on the 3rd floor of Discovery Building — room 3280. To get to the third floor, take the elevator located next to the Aldo’s Cafe kitchen(see photo). If you cannot attend in-person, we invite you to stream the event via Zoom: https://uwmadison.zoom.us/j/92639425571?pwd=Z0tCaWZxK0dDcWs2dm51dXZpcy9mQT09. Meeting ID: 926 3942 5571. Passcode: 111195.