The technology that argues ChatGPT has the potential to do more than just talk. Linxi Jim Fan, an AI researcher at chip maker Nvidia, worked with some colleagues to devise a way to set up the powerful GPT-4 language model, the brains behind ChatGPT and a growing number of other apps and services, to untangle all inside the Minecraft block video game.
The Nvidia team, which included Anima Anandkumar, the company’s director of machine learning and a professor at Caltech, created a Minecraft bot called Voyager that uses GPT-4 to troubleshoot in-game problems. The language model generates goals that help the agent navigate the game and code that improves the bot’s ability to play the game over time.
Voyager doesn’t play as a person, but can read the game state directly, via an API. She might see a fishing rod in her inventory and a river nearby, for example, and use GPT-4 to suggest a goal to fish to gain experience. She will then use this goal to have GPT-4 generate the code needed for the character to achieve it.
Courtesy of NVIDIA
The most innovative part of the project is the code generated by GPT-4 to add behaviors to Voyager. If the initially suggested code doesn’t work perfectly, Voyager will try to improve it using error messages, in-game feedback, and a description of the GPT-4 generated code.
Over time, Voyager builds a library of code to learn how to build more and more complex things and explore more parts of the game. A graph created by the researchers shows how capable she is compared to other Minecraft agents. Voyager gets more than three times as many items; she explores more than double; and creates tools 15x faster than other AI agents. Fan says the approach could be improved in the future by adding a way for the system to incorporate visual information from the game.
While chatbots like ChatGPT have wowed the world with their eloquence and seeming knowledge, even though they often invent things, Voyager shows the huge potential of language models to perform useful actions on computers. Using language models in this way could perhaps automate many routine office tasks, potentially one of the technology’s biggest economic impacts.
Courtesy of NVIDIA
The process Voyager uses with GPT-4 to figure out how to do things in Minecraft could be adapted for a software assistant that solves how to automate tasks via the operating system on a PC or phone. OpenAI, the startup behind ChatGPT, has added plugins to the bot that allow it to interact with online services like grocery delivery app Instacart. Microsoft, which owns Minecraft, is also training AI programs to play it, and the company recently announced Windows 11 Copilot, an operating system feature that will use machine learning and APIs to automate certain tasks. It might be a good idea to experiment with this type of technology within a game like Minecraft, where the faulty code can do relatively little harm.
Video games have long been a testing ground for AI algorithms, of course. AlphaGo used a technique called reinforcement learning, which trains an algorithm to play a game by giving it positive and negative feedback, such as from your score within a game.
It’s harder for this method to guide an agent in an open-ended game like Minecraft, where there is no score or set of goals, and where a player’s actions may not pay off until much later. Whether or not you believe we should be preparing to contain the existential threat of AI right now, Minecraft looks like an excellent playground for the technology.
#put #GPT4 #Minecraft #discovered #potential