AI News, BOOK REVIEW: Minecraft to run artificial intelligence experiments
- On Sunday, June 3, 2018
- By Read More
Minecraft to run artificial intelligence experiments
Microsoft, owner of the popular video game, revealed that computer scientists and amateurs will be able to evaluate and develop AI software using its virtual landscapes from July.
To take advantage of the offer, users will need to install AIX - a software platform that hooks into Minecraft and allows the artificial intelligence code to control a character and get feedback about the consequences of its actions.
'People build amazing structures that do amazing things in Minecraft, and this allows experimenters to put in tasks that will stretch AI technology beyond its current capacity,' explained Katja Hofmann, who leads the project at Microsoft Research's Cambridge lab in the UK.
'But eventually, we will be able to scale this up further to include tasks that allow AI agents to learn to collaborate with humans and support them in a creative manner.
Microsoft expects one of the most popular types of research will be reinforcement learning, in which an AI agent learns how best to carry out a task via a mix of experimentation and use of prior knowledge, rather than being told what to do.
Last week, it triumphed against one of the best ever players of the board game Go after having played thousands of games against itself to discover new strategies.
Minecraft Is a Testing Ground for Human-AI Collaboration
The blockish and slightly dorky computer game Minecraft may turn out to be a great place for humans and AI to learn how to work together.
An experimental new version of the game, released by Microsoft researchers this month, can be used to train an AI to perform all sorts of tasks, from crossing bridges to building complex objects.
Hofmann, who gave a demo of the software to AI researchers at an academic conference in New York last week, says that human-AI collaboration is a key goal for the project: “We’ve built in all the capabilities that a researcher would need in order to work toward collaborative AI.” Malmo is geared toward testing reinforcement-learning algorithms, a way of training a computer to perform a task by providing simulated rewards.
It is possible, for instance, to use reinforcement learning to train an AI controlling a Minecraft character to navigate through a room filled with obstacles by providing a reward when it reaches the other side.
She adds that the platform will offer an efficient way to gather large amounts of data, for example concerning human-AI interaction, that is needed for modern machine learning.
While Malmo is primarily aimed at those working on artificial intelligence, machine learning, and robotics, anyone with sufficient technical savvy can download it and experiment with an in-game AI.
Microsoft is using Minecraft to train AI and wants you to help out
Computer scientists at Microsoft have developed a new artificial intelligence platform atop the hugely popular video game Minecraft.
The AI not only needs to understand its surroundings, but it also needs to figure out the difference between day and night, why walking on lava is probably a bad idea, and when exactly it has achieved its goal via a system of rewards.
While Microsoft says it's entirely possible to program an AI, stick it inside a robot, and have that try to climb a real-world hill, it's an extremely costly experiment—particularly if (when) the AI fails and the robot tumbles down the hill.
AI research has come on leaps and bounds in recent years, but it typically focuses on having a computer learn a specific task, whether that's beating humans in Jeopardy, or, uhh, beating humans at the game of Go.
Microsoft says that by using Minecraft researchers will be able to create AI that's more adept at 'general intelligence,' learning in a similar way to humans by parsing information from light, smell, sound, and touch.
AI Agents Learn to Work Together by Wrangling Virtual Swine
AI researchers often develop software capable of performing a specific human task, such as playing chess or Go, and then measure it according to its ability to defeat a human player.
However, a great deal of human intelligence involves communication, social intelligence, and theory of mind, or the ability to anticipate and interpret another intelligent agent’s intentions.
Much further progress will be needed before AI agents can team up in useful ways or assist humans, but the contest offers a way to test some early ideas.
The winners of the contest, a team from the University of Oxford in the U.K., used reinforcement learning, a kind of machine learning inspired by the way animals learn through experimentation (see “10 Breakthrough Technologies: Reinforcement Learning”).
“There was no single type of approach that emerged as a clear winner,” she adds, saying it’s likely that hybrid approaches “will prove particularly promising directions for future research.” The pig-wrestling challenge takes inspiration from a thought experiment known as the Stag Hunt, which explores concepts within game theory, a branch of mathematics concerned with cooperation and negotiation strategies.
Software can become overoptimized for that particular environment and therefore less useful in the real world, he says, although more sophisticated simulated worlds are starting to change this.
- On Sunday, September 22, 2019
Project Malmo – a platform for fundamental AI research
I present Project Malmo – an AI experimentation platform built on top of the popular computer game Minecraft, and designed to support fundamental research in ...
Prof_Watson, Autonomous AI in Minecraft
Here is a little project I'm working on. A virtual character, based on my eXistenZ AI engine. Enjoy!
Moral Math of Robots: Can Life and Death Decisions Be Coded?
A self-driving car has a split second to decide whether to turn into oncoming traffic or hit a child who has lost control of her bicycle. An autonomous drone needs ...
Research in Focus: AI experimentation platform—Project Malmo
This session talks about Project Malmo, a unique AI experimentation platform based on the game Minecraft. Microsoft Research's Katja Hofmann and Evelyne ...
DeepMind Learns Parkour
Google's Deepmind AI learns to navigate obstacles using reinforcement learning. Paper: Click Here To Subscribe
Unreal Engine 4 Tutorial - Basic AI Navigation
In this tutorial we will setup some basic AI that can navigate in our map. We will do this by using Navmesh volumes and also simple move to nodes within ...
DevOps for Game Developers with Microsoft Azure
VSTS and Azure together enable developers to create a customize a highly sophisticated CI/CD pipeline for the backend of their service, their website, etc. and ...
Yelawolf - American You
Pre-order the album Love Story now On iTunes: Google Play: Amazon MP3: .