November 2019

Why does F = ma?

⋅

By Michael Nielsen, November 23, 2019 Note: Rough working notes, me thinking out loud. Thoughtful, well-informed further ideas and corrections welcome. The laws of physics don’t need justification, they just are. In that sense, “why does F = ma?” is a ridiculous question. Certainly, it can’t be proved, it’s not a mathematical theorem, or in any sense inevitable. Indeed, it’s easy to imagine universes in which F = ma is not true: we live in such a universe, […]

Ver mais

Liked Liked
The impact of direct air carbon capture on climate change

The impact of direct air carbon capture on climate change

⋅

By Michael Nielsen, November 21, 2019 Note: Rough and incomplete working notes, me thinking out loud. I’m not an expert on this, so the notes are tentative, certainly contain minor errors, and probably contain major errors too, at no extra charge! Thoughtful, well-informed further ideas and corrections welcome. In these notes I explore one set of ideas for helping address climate change: direct air capture (DAC) of carbon dioxide – basically, using clever chemical reactions to pull CO2 […]

Ver mais

Liked Liked
Meta-Learning Millions of Hyper-parameters using the Implicit Function Theorem

Meta-Learning Millions of Hyper-parameters using the Implicit Function Theorem

⋅

Last night on the train I read this nice paper by David Duvenaud and colleagues. Around midnight I got a calendar notification “it’s David Duvenaud’s birthday”. So I thought it’s time for a David Duvenaud birthday special (don’t get too excited David, I won’t make it an annual tradition…) Jonathan Lorraine, Paul Vicol, David Duvenaud (2019) Optimizing Millions of Hyperparameters by Implicit Differentiation Background I recently covered iMAML: the meta-learning algorithm that makes use of implicit gradients to […]

Ver mais

Liked Liked

Self-Supervised Representation Learning

⋅

[Updated on 2020-01-09: add a new section on Contrastive Predictive Coding]. [Updated on 2020-04-13: add a “Momentum Contrast” section on MoCo, SimCLR and CURL.] [Updated on 2020-07-08: add a “Bisimulation” section on DeepMDP and DBC.] [Updated on 2020-09-12: add MoCo V2 and BYOL in the “Momentum Contrast” section.] [Updated on 2021-05-31: remove section on “Momentum Contrast” and add a pointer to a full post on “Contrastive Representation Learning”]

Ver mais

Liked Liked

↑

Digitado © 2025