Thread by @techreview

For our last chapter of #EmTechDigital sessions, we'll hear from AI researchers and then to close, David Ferrucci from @elementalcog will pull back the curtain on AI.

First up: @agrimgupta92 from Stanford is interested in how the form of a machine changes its ability to learn, shifting the focus away from learning algorithms operating by themselves and onto learning combined with a kind of bodily evolution. #EmTechDigital

@agrimgupta92 .@strwbilly covered how Agrim's work could unlock AI for robots late last year. #EmTechDigital technologyreview.com/2021/10/19/103…

Agrim says there's a connection between generative design and his work. #EmTechDigital

Our next researcher is @ZennaTavares, an associate research scientist at Columbia University and cofounder of Basis. He's interested in giving machines the ability to reason about cause and effect. #EmTechDigital

Why are humans the model for AI?

Humans can understand humans, and humans have achieved things that other animals haven't, Tavares says. #EmTechDigital

What are the fundamental challenges ahead?

Many of them are organizational, Taveres says. Figuring things out when you're not a big company is harder. #EmTechDigital

Our final researcher is @natashajaques, a research scientist at Google Brain and previously TR35 winner, our annual list of 35 researchers under the age of 35 that recognizes outstanding work by young researchers. #EmTechDigital

Natasha is interested in ways to build AI agents that learn not just from data, but from each other—a kind of social learning for AI. #EmTechDigital

Social learning is learning from other intelligent agents in your own environments. Human social environments drive increasingly complex behavior. Maybe multi-agent AI environments can too. #EmTechDigital

Multi-agent social learning can be an effective to improve learning and generalization, Jaques says. #EmTechDigital

Following our researchers who just brought us three new ways of developing and thinking about AI, now we're hearing from our next speaker, David Ferrucci, founder and CEO of @elementalcog. #EmTechDigital

"The last 10 years have been the most exciting 10 years of AI," Ferrucci says. He credits machine learning. #EmTechDigital

Machines need to be understood and trusted--today's models don't offer that. #EmTechDigital

Watson was like a language model, but didn't do it as elegantly, Ferucci says. #EmTechDigital