Artificial Imagination for Reinforcement Learning Now made Possible by deepsense.ai and Google Brain

deepsense.ai, Google Brain, the University of Warsaw and the University of Illinois at Urbana-Champaign collaboratively built a neural network that mimics a simulated environment and effectively enabled artificial intelligence to perform a simulation.

The project was run by data scientists and researchers from deepsense.ai including Henryk Michalewski, Piotr Miłoś, Błażej Osiński and fellow researchers from Google Brain, the University of Warsaw and the University of Illinois at Urbana-Champaign, says the announcement.

What is Reinforcement Learning?

AI aka Artificial Intelligence is an umbrella term for many technologies such as machine learning, deep learning and reinforcement learning etc.

For more curious minds, here is a paper from Google research scientist Kevin Murphy.

The paper states, "Reinforcement learning is the problem of getting an agent to act in the world so as to maximize its rewards."

It quotes an example, "For example, consider teaching a dog a new trick: you cannot tell it what to do, but you can reward/punish it if it does the right/wrong thing. It has to figure out what it did that made it get the reward/punishment, which is known as the credit assignment problem. We can use a similar method to train computers to do many tasks, such as playing backgammon or chess, scheduling jobs, and controlling robot limbs."

Executive Opinion

Chief Executive Officer at deepsense.ai, Tomasz Kułakowski, said, "We consider our research work an essential part of deepsense.ai and a key aspect of the company’s development. This time our researchers have really pushed the boundaries of knowledge with no marketing overstatements. The project brings fresh, new idea into AI research, a thing that is worth contributing into"

Research Findings in Brief

The research found a method to build a neural network to produce or mimic signals for Reinforcement Learning agent. Generally, such signals are consumed from sensors. This method reduces the cost of acquiring data to train reinforcement learning agents.

Atari gaming environment is a popular training ground for reinforcement learning models. The research team also successfully built a neural network that emulates the Atari gaming environment. The announcement noted that the network was able to create a version of the games Pong, Freeway and others that were nearly indistinguishable from Atari's.

The ongoing research is now focusing on the business applications such as video prediction tool and teaching the model to emulate the real world in all of its complexity and unpredictability.