Re: DeepMind AlignNet

A related  article by the same team which imho touches on key issues for KR
and what we do/would like to do here

https://deepmind.com/research/publications/Artificial-Intelligence-Values-and-Alignment







On Mon, Jul 27, 2020 at 9:38 AM Paola Di Maio <paola.dimaio@gmail.com>
wrote:

> This looks also in the KR direction,
>
> DeepMind’s AlignNet Learns Stable Object Representations Across Time
>
> https://medium.com/syncedreview/deepminds-alignnet-learns-stable-object-representations-across-time-50b88ea8a9b0
>
>
> What do we learn about AI KR from this example?
>
> Glad to see this direction in AI
>