Emergent Tool Use from Multi-Agent Interaction
1–10 of 64 posts
Re: Emergent Tool Use from Multi-Agent Interaction
#2I'm completely amazed by that. The hint of a simulated world seems so matrix-like as well, imagine some intelligent thing evolving out of that. Wow.
Re: Emergent Tool Use from Multi-Agent Interaction
#3Amazing. Very cool to see this sort of emergent behavior.
I also very much enjoyed this section:
"We propose using a suite of domain-specific intelligence tests that target capabilities we believe agents may eventually acquire. Transfer performance in these settings can act as a quantitative measure of representation quality or skill, and we compare against pretraining with count-based exploration as well as a trained from scratch baseline."
Along with the videos, I can't help but get a very 'Portal' vibe from it all. "Thank you for helping us help you help us all." - GLaDOS
Re: Emergent Tool Use from Multi-Agent Interaction
#4What is the size of these “strategies”, measured in weights,bytes, or whatever measurement you look at?
Re: Emergent Tool Use from Multi-Agent Interaction
#5Finally Auotcirricula gets some love! Discussed in some detail in https://www.talkrl.com/episodes/natasha-jaques
Re: Emergent Tool Use from Multi-Agent Interaction
#6What is the size of these “strategies”, measured in weights,bytes, or whatever measurement you look at?
You mean what is actually stored in "memory" to play those actions? It's usually the trained model, which is a graph with many layers, each containing many nodes and being connected to other layers. Depending on the model size, it can take anywhere from a few MB to tens or hundreds of GBs, but usually the smaller the better (as having a too large model will lead to over-fitting, meaning that it has enough data to only learn the strategies needed for the current problem and not generalize the solving of such problems).
Re: Emergent Tool Use from Multi-Agent Interaction
#7One plausible, perhaps optimal strategy in the second arena is for the hiders to build a shelter around the seekers and lock them in place, circumventing the whole cat and mouse over ramps and ramp surfing (which the seekers would never be able to access). I wonder why this strategy is not arrived at.
Re: Emergent Tool Use from Multi-Agent Interaction
#8Has any intelligence arisen without multi-agent interaction?
Probably belongs in our definition of intelligence.
Re: Emergent Tool Use from Multi-Agent Interaction
#9What is the size of these “strategies”, measured in weights,bytes, or whatever measurement you look at?
You mean what is actually stored in "memory" to play those actions? It's usually the trained model, which is a graph with many layers, each containing many nodes and being connected to other layers. Depending on the model size, it can take anywhere from a few MB to tens or hundreds of GBs, but usually the smaller the better (as having a too large model will lead to over-fitting, meaning that it has enough data to onl…
Yup, this is what I’m talking about. I’ve been following googles WANNs and other methods for reducing the size of these “strategies” drastically. In the WANN paper they have an example of going from About 2500 Parmams to 40. I’ve got my own hunch as to where it is going, but I’m wondering if openai is studying this at all.
Like there are simple primitives in civil engineering like the level and pulley, I expect we find a zoo of tiny primitives which are the things that these agents are learning.
Re: Emergent Tool Use from Multi-Agent Interaction
#10One plausible, perhaps optimal strategy in the second arena is for the hiders to build a shelter around the seekers and lock them in place, circumventing the whole cat and mouse over ramps and ramp surfing (which the seekers would never be able to access). I wonder why this strategy is not arrived at.
There are multiple seekers, and the seekers may not be placed close together.