How to generate game character behaviors using AI and ML - Unite Copenhagen

[@wassilchoujaa3478]

For that deep of a subject He handle the presentation pretty well

[@nalankadi1654]

Please add more documentation and examples for these kits. They sound fascinating during these presentations, but when you look them up, there is almost no information on how to use them. AI Planning especially, I can't find any discussion on it besides the small documentation page, the Unite presentation videos, and the forum discussion post.

[@conscarcdr]

French accent is now a mandatory requirement to do AI research :)

[@hitagiNcode]

ahahahah

[@theconfusedllama]

And not being French is a mandatory requirement for doing anything

[@lennoxhunter7338]

Sorry to be offtopic but does anyone know of a tool to log back into an instagram account? I was dumb forgot the account password. I love any tips you can offer me!

[@kanejaxxon599]

@Lennox Hunter instablaster :)

[@457Deniz457]

Future has arrived ! :) Gj guys :)

[@adamz8314]

I wish if they show more of incoming feature for AI pleaner., but I like the point when AI Planner, able to answer three questions,

[@joshua.jebadurai]

His breathing gave me anxiety, hope some ML agent gives him a collar mic next time. Great content though.

[@MalikenGD]

I have a few questions. The first half of the talk, that doesnt really need ML does it? It just sounds like a behavior tree meets a search algorithm. Where is the neural net ? In the second half, around 24 30 he talks about difficulty tuning. We've had difficulty tuning for years, is this portion trying to convince people to use ML instead of hard coding it?

[@milanstevic8424]

1) It doesn't need NN. That talk is just crammed up in the same talk session, because the whole presentation is an ongoing update on several projects, all being done on a vaguely similar topic -- the AI and related stuff. I watched the first guy some time ago in a dedicated talk, which was much more verbose and had a proper introduction to what they're trying to accomplish with it. 2) Yes, because that slide is all about the potential uses for the ML Agents solution. I don't think they're trying to convince anyone, difficulty tuning is simply just a valid example for how one can use this technology, and of course this depends on the actual project.

[@milanstevic8424]

@42:48 you gotta love that crash

[@81gamer81]

Imagine standing on stage holding a presentation on an advanced topic for hundreds your peers. Its recorded. Now imagine doing it in another language you dont often speak

[@milanstevic8424]

Now imagine you're blind, have no mouth, and you're held hostage in front of an SS committee that's in a bit of a hurry. Yes, I'm joking.

[@iWhisperASMR]

NEAT

[@malcomxpetabeta776]

nice animation s

[@ThePCxbox]

You know using ML agents are fun and all but actually figuring out how to install the damn thing and use it is such a pain in the ass. Atleast make a straight forward reference doc instead of making us use an autogenerated one :/

[@tatoforever]

Don't use ML if you know what your agent should be doing. Unless you want to spend a large amount of time doing the same thing a hand made algorithm can compute a lot faster.

[@mohokhachai]

🎉

[@BNuwan123]

Talker is also an AI

[@erobusblack4856]

Im trying to find a programmer that can make a one super ai it is the key point of the game

[@dreamingacacia]

KEKW

[@DerXavia]

oof thick french accent

[@INeatFreak]

Biavia

[@tatoforever]

Another abandonned plugin to just pump virtually the value of Unity.