I’m interested in what attracts you to these systems? Sone early network science came out of looking at ecological networks, and as far as I’m aware a lot of behaviour modelling in those networks used agent based modelling, so the the intellectually roots of these systems with things like LLMs are fairly close, but what can these systems uniquely help us with, do you think?
Certainly the emergence aspect is intriguing to me. The way that a relatively small set of rules and primitives results in complex and surprising behavior.
Ok, so here’s and idea, I see that LLMs are also exhibiting emergent behaviour (presumably from the way information can be encoded in the network structure of the model). There must be some constraints in the kinds of emergence that can happen e.g. energy, or pseudo-energy constrains. I wonder if studying such constraints in these systems might give guidance on what kind of constraints might hold in more complex systems (kind of like how you put constraints on some Monte Carlo approximation methods)
Great stuff, thank you for writing! 💚 🥃
Thanks so much!
I’m interested in what attracts you to these systems? Sone early network science came out of looking at ecological networks, and as far as I’m aware a lot of behaviour modelling in those networks used agent based modelling, so the the intellectually roots of these systems with things like LLMs are fairly close, but what can these systems uniquely help us with, do you think?
Certainly the emergence aspect is intriguing to me. The way that a relatively small set of rules and primitives results in complex and surprising behavior.
Ok, so here’s and idea, I see that LLMs are also exhibiting emergent behaviour (presumably from the way information can be encoded in the network structure of the model). There must be some constraints in the kinds of emergence that can happen e.g. energy, or pseudo-energy constrains. I wonder if studying such constraints in these systems might give guidance on what kind of constraints might hold in more complex systems (kind of like how you put constraints on some Monte Carlo approximation methods)
Intriguing idea! Will have to think about this more.