Imagry Makes Open Source Motion Planning and Deep Reinforcement Learning Software Available to Developers on GitHub

The Aleph-Star Algorithm Allows Developers to Implement Improvements in Applications Ranging from Robotics to Autonomous Driving
By: Imagry
TEMPE, Ariz. - May 9, 2019 - PRLog -- Imagry, an autonomous vehicle software developer with a pioneering mapless algorithm, today announced that the necessary code needed to reproduce the results of its improvements to the popular A* path-finding algorithm is available on GitHub at Known as Aleph-Star, the general model-based reinforcement learning algorithm can be used by developers seeking to improve the performance of applications such as industrial robots used in manufacturing or, as in the case of Imagry, vision-based autonomous driving.

"What the Monte Carlo Method of artificial intelligence did for the game of Go, the Aleph-Star code is doing for today's machine learning applications without the need for heavy computational resources," said Adham Ghazali, founder and CEO of Imagry. "This approach has allowed us to fundamentally transform autonomous mobility from research to commercialization by pushing the limits of computer vision and artificial intelligence."

The Aleph-star algorithm helps find the minimal cost path from point A to point B. It uses domain knowledge, a so-called "heuristic function", which is an estimated cost to target. By calculating the absolute minimal path while visiting the minimal number of nodes, Aleph-star provides the best domain knowledge in the most efficient way possible.

About Imagry

Imagry is an autonomous vehicle platform developer that's dedicated to pioneering the future of this technology. With headquarters in both the U.S. and Israel, its technology utilizes mapless, real-time software to identify the road, route, vehicles, obstructions, pedestrians and more without a GPS. Its software is trained by a patent-pending simulator that's developed in-house, and it features a proprietary Aleph Star algorithm that uses physics-based planning to compensate for perception errors in real-time. More information is available at

Media Contact
Jillian Carapella
Posted By:*** Email Verified
Tags:Open Source, Autonomous Vehicles, Deep Learning
Location:Tempe - Arizona - United States
Account Email Address Verified     Account Phone Number Verified     Disclaimer     Report Abuse

Like PRLog?
Click to Share