Is ContinualAI focused on DL?

Hi, I just joined the slack server where a message encouraged asking questions on the forum, so here I am :slight_smile: I’m just discovering ContinualAI and in the Wiki the introduction has sections on DL and catastrophic forgetting, so I am wondering if ContinualAI is focused on DL? I did see one post here on ART but no discussion.

Hi! ContinualAI is focused on building AI that learns continually! Period.
Sure, much of this is influenced by DL, primarily because to date, DL has shown tremendous promise in fulfilling many of the goals of stronger AI. Also, since most of the AI field is seemingly pushing DL forward, those trends outside of CLAI will bleed in a bit.
But that said, we certainly aren’t only focused on DL. As an example, many of us (and myself included) are quite interested in the neuroscience of how our brains continually learn.
If there are other methods outside of DL that you are interested in, or think are promising, we would love to hear them!

1 Like

All sounds good, thanks!

Yep, I absolutely agree with @kwcooper. For sure ContinualAI has a strong grounding within the DL community but I think the field is growing fast in terms of scope and objectives!

Maybe the intro page could give a feeling for the breadth of approaches explored. I’ve seen some mention of ART and it would be interesting to know of other approaches being explored in addition to DL. I guess approaches within neuroevolution would be another angle. Personally I’m interested in EPANN (I would include EPRNN within that) and ART.