WEEK1
Here's the first AI lesson, again like the other lesson a simple rundown on what to expect as well as an intro explanation of what was AI.
Before that, I had to research as a task the difference between A.I, M.L, and D.L. This is what found out,
For A.I I knew what it was so here is my definition, "It is a type of intelligence that is built, that acts on and produces results with information and intelligence we give it"
For M.L I looked it up, saw two definitions, and combined them as a reference into my take, " It's within the studies of AI, and its main purpose is to "LEARN" and "ADAPT" without using Instruction given to it, what it does is use Algorithms and statistical models to help analyze and have a better understanding from the pattern in data."
And D.L, Here is what I found "It's AI that teaches computers to think like a human. It uses Images and texts to help train DANN".
Task 2: Understanding AI's Role in Gaming
1) There a a lot of games that AIn that replicate realism.
Ex 1, NPC reacting to the sound of your footstep, notice, in general, you made while trying to be a stealth - You will find it in games like AC, GTA, Hitman, Watchdogs, and any survival game.
Ex 2, Blending in the crowd and getting spotted when you leave in an area you not supposed to be - Games like AC, Hitman.
Ex3, Petting an animal(dog) - any that has this feature.
Ex4, When I hit an NPC everyone will react including the NPC that is getting attacked. - Every open world
2)The classic game I will be choosing is Zelda and the modern I will be choosing will be any adventure survival ( BOTW, AC, Fenyx).
For a platforming game like Zelda, the AI was simple, if I walk in an area that has enemies they follow me, and they will give me their attack sequence, and it will all stop until destroy them. Some might interact with the environment objects I think. In modern Game open-world/adventure Survival games, You can sneak past them, without making a sound if you do they will be alerted. Adding on that you have to be in their view for them to see you. They follow you until you quickly hide until they go back to what they were doing. If I attack one it will alert the other npc. But both do have NPC dialog. For environment AI behavior the NPC can use mother nature as weapons. Another AI day and night cycle, weather, or random events. From the amount of memory, technology upgraded, and mindset improved, you can see we went into details built on that over the years.
3) I guess the First issue is overcomplicating your AI idea and not taking your time/polishing it. The reason is building an AI behavior is not easy and you will end up with issues and a lot of troubleshooting, especially trying something you haven't done before which can't afford to mess up. So you will need to plan to continue testing for anything unusual. And for behavior, you have done and confident test it to make sure it's ok. The point of AI in an open world is to make illusion realism, sometime the process takes a long people will try to cut corners which just ruins the game itself
Unrelated. Another issue with AI is that AI is popular so people will use these AI to help build their game not being sure if it's going to work or not. And take away the creativity.
4) One scenario that is the stealth area which includes enemy patrol is interrogation/hold-ups. So If my players walk up to an enemy put it in hold up, AI behavior kicks in NPC gets startled and afraid begging to stop and you will get an option on how you will interrogate. I got the influence from Mgs etc.
I believe it will enhance the game experience it feels realistic and makes you feel like you doing to someone in real time. And it can easily be upgraded and improved upon making it more interesting. We can have where while you trying to interrogate NPC it's trying to wiggle out of the situation if you take long or do nothing while you hold them hostage.
You could easily send a Companion to interrogate on your behalf and come back to you and tell you what they have found.
WEEK 2
Today was a simple session, we had a small Unreal AI demonstration from a teacher assistant ROB, and we discussed AI, specifically unreal AI tools. He presented blueprints blackboard going AI behavior and such. The prof showed us an AI PowerPoint presentation to read through while discussing what was on it. We talked about different types of AI that are used in games. Like the navigation aspect for AI moving around, sensing like hearing and sights, decision making, and reacting to an object.
WEEK 3
(OCT) 10/08 - 10/09/2024
Went over the different types of movement AI that can be used in Unreal, During the week had to try out different ways to use it. I have done AI before but simple ones it was done in a blueprint like these example tutorials that have shown little different results. Self explanatory easy to replicate.
WEEK 4
10/15/2024 -
I spent the whole lesson continuing with my Egg catcher game. There were some complications when came to the Anvil. Continuing from the last lesson of that day. Which was spawning randomly between anvil or egg, when caught anvil you get no point, and the basket gets destroyed and you lose. I implemented most stuff just when I almost got it still ain't working, I got some help and realized I needed if statements or (||). Another issue I came across was the Retry UI but I'm still working, I have no idea why I'm having trouble. Somehow I got auto restart working, but now I need to get it attached to the UI.
Today Homework
Was wrote my definition of these terms:
BlackBoard
behavior tree
AI controller
behavior tree
AI controller
Ai Controller - is self-explanatory it is a controller that has an AI function that helps a pawn get controlled with the use of AI.
Blackboard = is a graph AI tree that contains AI behaviors, similar to blueprint graphs but it focuses on AI behaviors. It is used to set up put AI plan on unreal or unity etc, on how AI should work.
Behavior tree = is a visual diagram that models the AI decision and how it will execute the plan within that space.
Blackboard = is a graph AI tree that contains AI behaviors, similar to blueprint graphs but it focuses on AI behaviors. It is used to set up put AI plan on unreal or unity etc, on how AI should work.
Behavior tree = is a visual diagram that models the AI decision and how it will execute the plan within that space.
WEEK 7
I started a new document tutorial prof gave to us it will be about creating AI opponents, and combining my AI class and his networking class.
So far it started well until had a problem with typical unreal engine sensitivity. Need proof to check it out.