This inventor holds 1 USPTO granted patent and 1 published patent application. Top assignee: Kinaxis Inc.. Active years: 2026.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile
Company Filing History:
Years Active: 2026
Title: Lauris Petlah: Inventor in Accelerated Tree Learning.
Introduction
Lauris Petlah is an inventor associated with Richmond Hill, California, and Kinaxis Inc. His work includes a patent focused on accelerating the training of gradient-boosted trees.
Latest Patents
Petlah is associated with 1 patent, titled “Systems and Methods for Process Execution.” The patent describes procedures for accelerated tree learning. One approach uses self-adapting learning rates, while another uses multiple learning rates that change during training. In the latter approach, a learning rate increases linearly when a pseudo-residual maintains its direction across successive iterations and decreases exponentially when the residual changes direction. The methods can also be combined with momentum-augmented gradient boosting and Nesterov Accelerated Gradient Boosting. The disclosed techniques aim to reduce the training time and computational resources required for gradient-boosted trees.
Career Highlights
Petlah’s listed patent connects his work at Kinaxis Inc. with machine-learning methods designed to make model training more efficient. The patent’s approaches address how learning rates adapt as training progresses.
Collaborations
Marin Creanga and Dylan Ellicott are identified as Petlah’s coworkers. The available information does not specify their roles in the patent or the work described in it.
Conclusion
Lauris Petlah’s patent describes methods for accelerating gradient-boosted tree training through adaptive learning rates, with the goal of reducing computational time and resources.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile
