Some very wise purchasers of "The Atomic Human" sensibly choosing to buy "The AI Mirror" by @ShannonVallor Recommended by an algorithm … (what could possibly go wrong?) Here's the link to Shannon's book: https://
global.oup.com/academic/produ
ct/the-ai-mirror-9780197759066
…
@lawrennd
-
Wise Readers Choosing Shannon Vallor’s AI Mirror Book
By
–
-
Open Source Definitions and AI: Critical Issues for UK Future
By
–
Important piece from @AmandaBrockUK on the issues around open source, the importance of definitions, history, emerging issues around AI. Vital for the future of the UK that we get this right. Thank you Amanda and @openuk_uk for leading on this.
-
UK Government Misunderstanding of AI Safety Investment and Regulatory Funding
By
–
Yes, likely both. In the UK case it seems the government (or their advisors) just didn't/don't understand. With 10x more for AISI than the 10 million for the 19 regulators. But it seems odd that same is true throughout Europe … effective lobbying?
-
GPLVM superiority in low data regime versus modern generative AI
By
–
That's why GPLVM works so well on low data … low dim inputs. I don't think there's a "modern" Gen AI approach that competes in (very) low data regime. My favourite demonstration was @baxissimo's Latent Doodle Space paper.
-
Davos AI House Panel Discussion with Leading AI Researchers
By
–
Points made are reminiscent of the panel convened at Davos AI House by @ai_cam_mission and featuring Jessica Montgomery, @ylecun @bschoelkopf and self.
-
TechCrunch Article on Open Source and AI Definitions
By
–
TechCrunch article on definitions od open source and AI.
-
PASCAL Visual Object Challenge Inspiration for ImageNet
By
–
He was very active in compiling the PASCAL Visual Object Challenge which was one of the inspirations for ImageNet.
-
Random Fourier Features Bridge Gaussian Processes and Neural Networks
By
–
Agreed, they're very widely used there. Although if you start using e.g. random fourier features approximations to fit them then you start getting NN like models … and all the techniques of NNs can be imported to help with scale.
-
Neural Networks as Gaussian Processes in the Infinite Width Limit
By
–
The sort of mathematics we can do with NNs is to take limits of layer sizes such that the start to behave like GPs. Then we understand how the NN is working in that limit.
-
Gaussian Processes: Mathematical Clarity vs Neural Network Complexity
By
–
The mathematical understanding of the model. It's extremely well characterised because everything's Gaussian. The mathematics of NNs is so complex that modern "ML experts" go on TV and tell everyone that "we don't understand how they work". We do understand how GPs work.