Use high-quality simulations to train and test artificial intelligence for business applications

The relationship between simulation and artificial intelligence is increasingly close, particularly for deep reinforcement learning. The AnyLogic Company CEO, Dr. Andrei Borshchev, explores this trend in Using Simulation to Train and Evaluation Artificial Intelligence for Business Applications, a demonstration given in the GE EDGE & Controls Symposium 2019 in Niskayuna, NY.
Below is the movie recording and a brief summary. You can even find the movie on our YouTube station with timestamped topic segments.
The demonstration begins with a quick look in the kind of simulation AnyLogic provides and dispels some misconceptions about electronic twins (4m 46s). A point exemplified with decisionLab’s industrial electronic twin for Siemens – ATOM (10m 37s) (see also, the case study).
Why AI and Simulation?

Evaluation Artificial Intelligence

This query marks the start of the main motif (14m 8s) of this demonstration and shows how simulation fits in with the different types of AI technologies, as well as the Various Ways simulation can be utilized with AI:
To generate synthetic training information
To offer learning environments
As a testbed for trained AI
For all these, it’s possible to observe how simulation helps when data is hard to get, maybe because it wasn’t collected, or set is too dangerous or costly, or just because a true source of data doesn’t exist yet. The wonderful advantages AI professionals are discovering in simulation come from its low-cost and risk-free nature.
The provision of learning environments for training reinforcement policies is emphasized as a rapidly growing area and the demonstration provides both a specialized (21m 21s) (Machine Learning vs Optimization for Traffic Lights) and a commercial example (29m 53s) (Industrial Problem Resolved by AI and Simulation).
To wrap up, the conclusion (32m 26s) highlights three key challenges facing integrated simulation and AI. These centre around the issues with scaling, the abilities required for effective simulation modeling, and issues with the development procedure.
The challenges aren’t insurmountable, however. You will find ways forward — in particular, new tools, such as Pathmind, are beginning to facilitate development processes.
Have a look at the video and find out how the worlds of simulation and artificial intelligence are coming together as well as changing. Leave a comment below: What are your experiences working in these two fields? Do you agree with the challenges above?

Post Author: Tech Review