Thursday, December 1, 2016

GoodAI Releases School for AI - A New Update to Brain Simulator

In Brief:
  • We are releasing School for AI, a training and testing environment designed to teach your AI architectures new skills and to measure their performance.
  • Download School for AI, the new update to Brain Simulator.

Today I’m happy to announce a new update to Brain Simulator - School for AI. This is both a training and testing environment that allows you to teach your brain architectures new skills and at the same time measure their performance.

School for AI (or School for short) is a world within Brain Simulator, coupled with an intuitive graphical interface. The training provided by School is separated into learning tasks. A single learning task teaches or tests preferably a single new skill or ability. A collection of learning tasks forms a curriculum. An agent is expected to gradually acquire knowledge as it is progressing through the curriculum. It should acquire knowledge from one learning task and then leverage that knowledge to acquire knowledge from the next learning task faster and with greater ease. We call this accumulation of skills “gradual learning” and we talk about it in more detail in our Framework document.

The training can occur in a range of environments. We've prepared a basic 2D environment (RoguelikeWorld), an advanced 2D environment (ToyWorld), and a basic 3D environment (3D version of ToyWorld). A single curriculum can train the same architecture on multiple different environments.

School for AI does not explicitly separate the training and testing phases. It prefers life-long learning - the agent is learning continuously, without any interruption. The testing is performed during training. This of course poses certain requirements on the datasets - the environments that are used in school are generated rather than fixed. This allows a fast learning agent to progress through the curriculum in shorter time than a slow learning agent. We’re planning to include existing datasets (both fixed and interactive) in School to provide a single interface and make using different datasets a breeze.

The learning tasks that appear in the current version of School are based on our Agent Development Roadmap. They teach the agent basic skills like object detection, color classification, or path finding. We took care to ensure that it is easy to add new learning tasks to the environment. We’ll be happy if you go on, get Brain Simulator code from GitHub and create a learning task of your own!

The school communicates with the agent through a fixed interface. The interface allows the agent to move around the environment, move its focus around the scene and interact with the objects found within the environment (when applicable).

For more information, have a look at the Brain Simulator documentation page for School for AI or just download Brain Simulator and try out School for yourself!

We look forward to receiving your input! please don't hesitate to get in touch at
Read all about GoodAI's latest achievements, technology, and initiatives on our About page.
Marek Rosa
CEO, CTO of GoodAI
CEO of Keen Software House

1 comment:

  1. My name is Harold McBroom, I'm in your friends list as well; I love Space Engineers but it needs a lot of rethinking along the area of what your trying to accomplish as a whole with the game. It's half semi-realistic and half arcade, other games suffer this problem as well, such as 7 Days to Die. We spend a long time collecting resources to fabricate ships and stations, but the arcade aspect of the game destroys all that work in mere minutes; Perhaps making the hull parts stronger, and the inner parts more fragile would help. We need more functionality and things to do in the game. How about a microscope to add to the "Science Lab", so we can unlock a tech tree with research and development. How about software upgrades for the computer terminals, so we can designate one particular terminal as for Engineering, another for Science, another for security, etc. As it is now, we have a huge empty bases with only space, no functional crew. AI is not hard to add to the game, perhaps if you take the Republic Commando approach of pre-designated places where AI can be. Use a "marker" system... Stand over a computer terminal, press a hot key to designate the current block as an "action" point where an AI can perform an action, such as provide "covering fire" in a particular direction, access a computer, or preform a door or wall breach. The marker would stay where they are, much similar to how the projector works, even if the associated block is destroyed, that way the AI can repair those blocks and re-establish it's path finding. AI must be aware of where it belongs, what it's duties are, and when something is not right, have the brains to fix it. Know when something is in its way, I've put a lot of work into AI myself. IF you want to be successful in GoodAI, it starts with Space Engineers; This is your chance to make a huge difference! I'm never motivated by what others CAN accomplish, but more motivated by what they DO accomplish, that others only complain about the impossibilities! For Space Engineers, I want to explore, but there's nothing to explore! I want to do research and development, but there's nothing to research, and nothing to develop! My FS Aurora on the Workshop, is empty waiting for a Captain, it will no longer see, because I'm missing a crew that gives a starship LIFE and purpose. Automated AI ships just don't breath life into this game, or locating ships and stations just as empty as my own, does not breath life into this game. Spending two hours fixing damage to a ship I spent weeks building, repels me away from the game, not closer! This is what can take a good game and destroy it quickly in the gaming market, Mr. Rosa, you MUST be able to go above and beyond what everyone else is doing, and focus on doing what others say CAN NOT be done! God Bless