Waymo Implemented Evolutionary Competition in Its Vehicles
Daniel Johnson — August 5, 2019 — Autos
Waymo uses evolutionary competition in order to improve the AI system implemented in its autonomous vehicles. In order to implement evolutionary competition, Waymo partnered with Deep Mind to create "Population-Based Training," which improves pedestrian alerting by using neural networks. Evolutionary competition works by having neural networks compete in order to find out which one is better. The network that does not perform as a good then gets replaced by the one that performed better.
The initial results of the program indicated success. Pedestrian detection rates performed better after the evolutionary competition was implemented, instances of false positives decreased by 24 percent. This drop occurred in approximately half the time it normally would using previous methods.
Concerns about diversity in neural networks were addressed when Waymo created "niches," which are subgroups.
Image Credit: Waymo
The initial results of the program indicated success. Pedestrian detection rates performed better after the evolutionary competition was implemented, instances of false positives decreased by 24 percent. This drop occurred in approximately half the time it normally would using previous methods.
Concerns about diversity in neural networks were addressed when Waymo created "niches," which are subgroups.
Image Credit: Waymo
Trend Themes
1. Evolutionary Competition in AI - Implementing evolutionary competition in AI systems can improve performance and accuracy by allowing neural networks to compete and replace inferior models.
2. Population-based Training - Population-based training is an effective method of improving alert systems in autonomous vehicles, resulting in increased pedestrian detection rates and reduced false positives.
3. Addressing Diversity in Neural Networks - Creating niches or subgroups within neural networks can help address concerns about diversity and enhance overall system performance.
Industry Implications
1. Autonomous Vehicles - Implementing evolutionary competition and population-based training can revolutionize the performance and safety standards in the autonomous vehicle industry.
2. Artificial Intelligence - Evolutionary competition and population-based training techniques provide disruptive innovation opportunities for improving AI systems across various industries.
3. Technology Research and Development - The development of evolutionary competition and niche-based approaches in neural networks opens up new avenues for research and development in the technology sector.
0.9
Score
Popularity
Activity
Freshness
