What’s the practical use of this technology right now?
Our short-term objective with this project has been to help the DICE team scale up its quality assurance and testing, which would help the studio to collect more crash reports and find more bugs.
In future titles, as deep learning technology matures, I expect self learning agents to be part of the games themselves, as truly intelligent NPCs that can master a range of tasks, and that adapt and evolve over time as they accumulate experience from engaging with human players.
When do you think we will see self-learning AI becoming a mainstream technology in games?
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.
Will self-learning agents ever beat professional FPS-players? If so, when?
With the risk of going out on a limb with a crazy prediction, I think it’s reasonable to expect AI agents capable to defeat human players in a limited competitive game mode—one that features smaller maps, focused teams and clear objectives—in a couple of years from now. However, at SEED we’re not necessarily out to build AI that will defeat human players. Our aim is to help create new experiences that enhance games and make them more fun. Getting owned by a superior AI isn’t necessarily that fun for players in the long run.
jdreyer wrote on Mar 23, 2018, 17:41:Thanks!Dev wrote on Mar 23, 2018, 06:31:jdreyer wrote on Mar 23, 2018, 01:45:maddog wrote on Mar 23, 2018, 00:08:jdreyer wrote on Mar 22, 2018, 18:57:
Very cool. I've been looking for a BF2 AIX replacement for a decade. This shows promise.
I made navmeshes for some of the maps in the AIX mod (bot name pc_load_letter). Glad to see people are still playing and enjoying the mod.
Was sad that the team broke up shortly after the release.
AIX is the best non-scripted single player experience I've had. I love it and have been looking for a replacement for a decade.
If I've never heard of this, where can I find info, and will it work with my old steam version of BF2?
Hey Dev, AIX 2.0 is at moddb. I'm using it with the Steam version of BF2.
eRe4s3r wrote on Mar 22, 2018, 18:44:
I feel like de ja vue... I swear I saw the same dumb AI back in CS 10 years ago in a mod AI......
Dev wrote on Mar 23, 2018, 06:31:jdreyer wrote on Mar 23, 2018, 01:45:maddog wrote on Mar 23, 2018, 00:08:jdreyer wrote on Mar 22, 2018, 18:57:
Very cool. I've been looking for a BF2 AIX replacement for a decade. This shows promise.
I made navmeshes for some of the maps in the AIX mod (bot name pc_load_letter). Glad to see people are still playing and enjoying the mod.
Was sad that the team broke up shortly after the release.
AIX is the best non-scripted single player experience I've had. I love it and have been looking for a replacement for a decade.
If I've never heard of this, where can I find info, and will it work with my old steam version of BF2?
jdreyer wrote on Mar 23, 2018, 01:45:maddog wrote on Mar 23, 2018, 00:08:jdreyer wrote on Mar 22, 2018, 18:57:
Very cool. I've been looking for a BF2 AIX replacement for a decade. This shows promise.
I made navmeshes for some of the maps in the AIX mod (bot name pc_load_letter). Glad to see people are still playing and enjoying the mod.
Was sad that the team broke up shortly after the release.
AIX is the best non-scripted single player experience I've had. I love it and have been looking for a replacement for a decade.
maddog wrote on Mar 23, 2018, 00:08:jdreyer wrote on Mar 22, 2018, 18:57:
Very cool. I've been looking for a BF2 AIX replacement for a decade. This shows promise.
I made navmeshes for some of the maps in the AIX mod (bot name pc_load_letter). Glad to see people are still playing and enjoying the mod.
Was sad that the team broke up shortly after the release.
Will self-learning agents ever beat professional FPS-players? If so, when?Eh?? Surely building a bot to beat a human is not that hard. Even the most basic bot would have near-zero reaction time and perfect aim.
With the risk of going out on a limb with a crazy prediction, I think it’s reasonable to expect AI agents capable to defeat human players in a limited competitive game mode—one that features smaller maps, focused teams and clear objectives—in a couple of years from now.
jdreyer wrote on Mar 22, 2018, 18:57:
Very cool. I've been looking for a BF2 AIX replacement for a decade. This shows promise.
Slick wrote on Mar 22, 2018, 20:08:
This is the right approach to take.
Traditional AI bots rely on a small map with preset pathing for the AI to take. It's why AI as far back as Quakebots have been able to "seem" passably realistic.
But anyone who played Bad Company 1 knows that once you bring this into an open world, and ESPECIALLY if you add things like destructable environments, the AI can barely make heads or tails of what's happening, as was evidenced by your 3 friendly AI teammates in the campaign.
DICE lamented how hard it was to do the AI back then, and this change towards machine learning is the right move, and should keep Frostbite as the industry-leading engine.