I am pretty convinced that stream/event forecasting is going to be the "next big thing" in AI - and this looks exactly like the kind of work needed to make prog

I am pretty convinced that stream/event forecasting is going to be the "next big thing" in AI - and this looks exactly like the kind of work needed to make progress. This is actually really exciting because I don't see a lot of event stream forecasting tools or research out there and Numenta has totally stalled out as far as I can tell. I need to dig in further to the background papers [1] but from the looks of things in that lab [2] work on complex event recognition ("heterogeneous and interdependent data streams") is exactly in the ballpark that is necessary. [1]https://arxiv.org/pdf/1804.10388.pdf https://arxiv.org/pdf/1804.10388.pdf [2]https://arxiv.org/pdf/1802.04086v1.pdf https://arxiv.org/pdf/1802.04086v1.pdf