EA is looking to create self-learning AI through its Search For Extraordinary Experience Division: SEED. This program is designed to supersede current bots and static AI, with EA wanting to build on self-learning algorithms to extend far beyond competing against humans; they also want to utilize the tech in multiplayer and procedural services.
Before diving into what this algorithm aims to do in detail, you can see what EA, DICE, and SEED have cooked up for GDC 2018.
According to SEED technical director Magnus Nordin, the team’s short-term goal is to help DICE scale up the quality of its faulty AI, evolving it beyond what was featured in the campaign of Battlefield 1. As for long-term goals, the team is looking to integrate this tech into future games and have NPCs accumulate experience and reflect it on player counterparts.
Nordin also made mention that…
“I have no doubt in my mind that neural nets will start to gradually make their way into games in the years to come. Self-learning agents aren’t just a good replacement for old-fashioned bots, you can also apply machine learning to a number of fields, such as procedurally generated content, animation, voice generation, speech recognition and more.”
A running joke around this system coming from folks leery of EA, is that this SEED tech will be used in some sort of multiplayer live service system given that the team is looking to integrate it in other programs and services.
There’s no telling if the running joke holds any weight, but it is certain that EA wants to expand on SEED tech beyond just players and AI fighting it out for virtual supremacy.
As of right now, SEED is in a very early state, as seen above in the video, which is an indicator that this system needs further improvement if it’s to be applied in a real world setting for practical use.
You can head on over to ea.com to read the full story about SEED and its progress.