About this Workshop

$50B dollars and 30 years worth of effort has been invested into self-driving cars, yet it is an open secret that they still require human operators in the loop. The same can be said in terms of chat-bots, and contextual agents. 

​

Why is that?

​

Why is it that machines still struggle with what we as humans find easy?

We believe that paradigmatic level shifts in thinking across a multitude of seemingly unrelated fields - some big and some small - need to be realized. 

This workshop attempts to tie them together through a series of thought provoking questions, observations, and proposals. 

​

Example items we will touch upon: 

​

• How does learning in natural systems differ from artificial ones?

• Does it make sense to talk about artificial intelligence without artificial life first? Why or why not?

• Should we bypass simulations via hybrid wetware (ie biological neural-nets in-a-vat)?

• Should we stop thinking about robots, and start thinking about synthetic-organisms?

• What do we miss if we artificially replicate an intelligence, without the natural lineage that intelligence came from? 

• Is interaction in a/the world decisive in bypassing passive big data?

• What would replacing task based paradigms and replacing them with open-endedness gain us?

• Is there an "atom" of intelligence/learning?

• What is the MVO (minimally viable organism) of natural intelligence? 

• Do you believe that embodiment can be replaced with data acquired from the Internet and/or simulation?

• Is statistical learning a good language to deal with world complexity?

• Is it correct to say that current struggles of AI are stemming from an attempt to use "short-tailed" statistical methods to "long-tail" data?

• If you had Google/FB/Microsoft compute at your disposal, what would you do with it?

• In 1981 the Neocognitron was proposed as a biologically inspired framework that resulted in convolutional neural nets and consequently the current wave of AI. Is there anything equivalent today that might create a similar wave of research in years to come?

• How do you see interaction with environment as a short-circuit for big data? Is "big data" a cul-de-sac of AI? Examples?

• How do you see your proposal/agenda be implemented using current paradigms if at all?

​