This sort of dovetails off a question I asked on the AI Exchange: Is There A Need For Stochastic Inputs To Mimic Real-World Biology And Environment?

I figured this site would give more insight on what I'm trying to accomplish neurologically/biologically.

My question is this: in order for artificial intelligence to more accurately simulate consciousness, bodily processes that may influence thoughts, and more complex response to stimuli that may not always be deterministic, would a new library similar to TensorFlow or PyTorch be necessary, or are these libraries sufficient enough?

This paper actually used FFNN to accomplish hunger/satiety: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4936290/

Would a new math be required, such as linear-time logic or fuzzy logic that would be better implemented in a completely new software library? Perhaps used in-tandem with something like TensorFlow, in order to reproduce the variations in feelings like hunger or random thoughts and creative inspiration?

For example, I'm thinking of an application similar to a tensor field, where over time a mood (happy/sad/angry) could be (literally) bit-by-bit changed, affecting surrounding tensors in a stochastic way, but it may not lead to a full mood swing at all.

Am I focused too specifically on a software library to accomplish the features of consciousness/moods/thoughts rather than the holistic Holy Grail everyone is chasing of accomplishing AI consciousness and artificial general intelligence? I was thinking of starting this as an open source software project but I may be on a wild goose chase.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.