More brainy give aways - Thanks everyone here for the TIPS! What if a computer vision system, is just detecting events. And its not learning anything. It has 1 condition, per event, in its finite little brain system. But you give a man a length of video, and he comes back shaking his head its too hard. Would 3d do any better? How about turning everything into 3d stickmen. Now maybe its easier to have basic conditional measurements be more accurate to activate events? But then what kind of context do you need to detect these events appropriately.
lets look at some examples. If I wanted to tell if someone was jumping, id detect if feet arent touching the ground. If someone is going to score a goal - He would be lowering in proximity. (But that would take looking over time maybe.) If someone was getting punched, Id detect a fist in close proximity to a head. I could then relate these events to the pattern store and give it more functionality, and the patterns could predict what events are coming up in the future. But I think the best way to do it, is if its playing monkey ball - write the states it needs to play it properly, then just pretend all the humans playing are it. So its a boolean model, detecting off the camera. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Ta9ef92db7ce9c030-M381c4b3a7780cb22f717e105 Delivery options: https://agi.topicbox.com/groups/agi/subscription
