NTR Webinar: TAO: a New Method to Train Powerful and Efficient Tree-based Models

NTR organizes and hosts technical webinars about neural networks and invites speakers from all over the world to lead them. Webinars are held on Tuesdays and most are in Russian. 

On July 13 Arman Zharmagambetov, University of California, Merced, USA, led a technical Zoom webinar on TAO: a New Method to Train Powerful and Efficient Tree-based Models.

About the webinar: 

In this talk, I presented a novel algorithm, Tree Alternating Optimization (TAO), to train decision trees and tree-based models; and to show its effectiveness in a number of tasks. 

Specifically, I showed that we can build a forest (ensemble of trees), where each tree is trained by the TAO algorithm, which can outperform state-of-the-art forests (like XGBoost, Random Forest, etc.). 

Furthermore, I demonstrated that the TAO algorithm is so generic that it can be applied to train hybrids of decision trees and neural nets. The resulting model strikes a good balance between accuracy, model size and interpretability.

Webinar recording (in Russian)

Webinar presentation (in English) 

Moderator and contact: NTR CEO Nick Mikhailovsky: nickm@ntrlab.com.

Leave a Reply

Your email address will not be published. Required fields are marked *