But whereas researchers had previously used models like Codex to generate tens or hundreds of candidates, DeepMind had AlphaCode generate up to more than 1 million. When presented with a fresh problem, AlphaCode generates candidate code solutions (in Python or C++) and filters out the bad ones. For example, a problem might ask for a program to determine the number of binary strings (sequences of zeroes and ones) of length n that don't have any consecutive zeroes. Then, they trained it to translate problem descriptions into code, using thousands of problems collected from programming competitions. Like the Codex researchers, they started by feeding a large language model many gigabytes of code from GitHub, just to familiarize it with coding syntax and conventions. Slashdot reader sciencehabit shares a report from Science Magazine:ĪlphaCode's creators focused on solving those difficult problems. The findings have been published in the journal Science. DeepMind's new artificial intelligence system called AlphaCode was able to "achieve approximately human-level performance" in a programming competition.
0 kommentar(er)
