Lightnews — Scholar-powered news

Michiel van de Panne @mvandepanne.bsky.social · 13d

Really interesting and apparently effective ideas here.

Bruno Levy @brunolevy01.bsky.social · 13d

Large-scale semi-discrete optimal transport with distributed Voronoi diagrams,
Final version available here:
www.sciencedirect.com/science/arti...

semi-discrete optimal transport: galaxy clusters and Laguerre diagrams.

1 4

Reposted by Michiel van de Panne

Jan Rosenow @janrosenow.bsky.social · 16d

Critics of renewables often point to land use as a reason not to invest in wind & solar. Land use IS a real issue for all energy assets incl renewables (unless on-site).

But what critics conveniently forget is the huge land impact of fossil fuels. Aerial view of fracking sites in Wickett, Texas. 👇

27 180 520

Reposted by Michiel van de Panne

Electrek @electrek.co · 17d

Tesla influencers tried Elon Musk’s coast-to-coast self-driving, crashed before 60 miles

A duo of Tesla shareholder-influencers tried to complete Elon Musk’s coast-to-coast self-driving ride that he claimed Tesla would be able to do in 2017 and they crashed before making it about 60 miles. more…

electrek.co

6 24 59

Reposted by Michiel van de Panne

madmcg.bsky.social @madmcg.bsky.social · 17d

"global private sector investments last year of more than $2tn in green energy & low-carbon industry, dwarfing the $1tn poured into fossil fuels"

Green is the way forward and ...
... "Gov who choose to ignore this, financial advisers and economists who chose to ignore this, do so at their peril.”

Calgary Climate Hub @calgaryclimatehub.ca · 18d

www.theguardian.com/environment/... #ClimateChange #EnergyTransition

‘Something is working’: UN climate chief optimistic about green transition

Exclusive: Simon Stiell believes economic benefits will compel countries to speed up climate action

www.theguardian.com

3 5

Reposted by Michiel van de Panne

JJ @jjodx.bsky.social · 26d

The myth of optimality in human movement science

www.sciencedirect.com/science/arti...

The myth of optimality in human movement science

The concept of optimality dominates contemporary human movement science, with researchers across biomechanics, motor control, and neuroscience routine…

www.sciencedirect.com

1 22 64

Michiel van de Panne @mvandepanne.bsky.social · Sep 3

I just measured 240 ug/m3 near Alma & 10th, with the tester beeping loudly at me.

Michiel van de Panne @mvandepanne.bsky.social · Sep 3

This is absolutely true -- this is a superb and much-needed consolidation of so much of modern RL. Kevin, inquiring minds want to understand the process you use to put this artwork together! @sirbayes.bsky.social Perhaps this is also the ultimate benchmark for Gemini Deep Research reports. ;-p

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · Sep 3

Reminded again of Kevin Murphy's excellent RL overview: arxiv.org/abs/2412.05265
A lot of the stuff covered here really is at the cutting edge and not compiled so nicely anywhere else

Reinforcement Learning: An Overview

This manuscript gives a big-picture, up-to-date overview of the field of (deep) reinforcement learning and sequential decision making, covering value-based methods, policy-based methods, model-based m...

arxiv.org

1 10

Michiel van de Panne @mvandepanne.bsky.social · Aug 31

There's a detailed study here: ICLR 2021, openreview.net/forum?id=nIA.... TLDR: Input norm is good, value fn norm can be both good or bad; mini-batch Adv norm has little impact. Would be interested to know of any needed revisiting of these.

What Matters for On-Policy Deep Actor-Critic Methods? A Large-Scale...

In recent years, reinforcement learning (RL) has been successfully applied to many different continuous control tasks. While RL algorithms are often conceptually simple, their state-of-the-art...

openreview.net

1

Reposted by Michiel van de Panne

UBC Computer Science @cs.ubc.ca · Aug 29

Congratulations to UBC’s student hacking club Maple Bacon for making history at hacking convention DEF CON 2025! Read more about their win in our new article: www.cs.ubc.ca/news/2025/08...

1 5

Reposted by Michiel van de Panne

Claire (derogatory) @clarissakickass.bsky.social · Aug 27

I don’t need a self driving car. I need a vehicle that runs between major areas and can transport large numbers of people at once to reduce environmental impact…wait it’s a train. I’m describing trains.
Build more goddamned TRAINS, America.

250 950 6.9K

Reposted by Michiel van de Panne

Tom Schaul @schaul.bsky.social · Aug 27

Where do some of Reinforcement Learning's great thinkers stand today?

Find out! Keynotes of the RL Conference are online:
www.youtube.com/playlist?lis...

Wanting vs liking, Agent factories, Theoretical limit of LLMs, Pluralist value, RL teachers, Knowledge flywheels
(guess who talked about which!)

1 23 75

Reposted by Michiel van de Panne

Brent Toderian @brenttoderian.bsky.social · Aug 24

WATCH & SHARE: This is very good. A prominent youtuber is actually giddy (but also very thorough in covering the joys, advantages, issues and challenges) about how trying biking for transportation in London has changed his life.

Easy to watch & more convincing probably than 99% of public campaigns.

I Cycled 2500km in London — Here's How It Changed My Life

YouTube video by Evan Edinger

youtu.be

13 81 280

Michiel van de Panne @mvandepanne.bsky.social · Aug 24

Really like this Dale Schuurmans talk at RLC 2025, touching on what LLMs can and cannot do, from first principles. youtu.be/-2URALk7IfQ?...

Dale Schuurmans, Language Models and Computation - RLC 2025

YouTube video by Amii

youtu.be

7

Michiel van de Panne @mvandepanne.bsky.social · Aug 24

Excellent talk by Joëlle Pineau at the Reinforcement Learning Conference (RLC 2025) youtu.be/PZcaMKxljxc?...

Joelle Pineau, Beyond Bellman's Legacy - RLC 2025

YouTube video by Amii

youtu.be

1 2 18

Reposted by Michiel van de Panne

Craig Reynolds @craigreynolds.bsky.social · Aug 23

My daughter knew I would enjoy the evolutionary dynamics (an evolutionary stable state / Nash equilibrium) described in this 3 minute video from KQED’s excellent Deep Look team:

These Lizards Have Been Playing Rock-Paper-Scissors for 15 Million Years

#evolution #nature

These Lizards Have Been Playing Rock-Paper-Scissors for 15 Million Years | Deep Look

YouTube video by Deep Look

www.youtube.com

1 7 33

Michiel van de Panne @mvandepanne.bsky.social · Aug 23

Waabi aiming to deploy in Texas later this year. Will be interesting to track.

1 1

Michiel van de Panne @mvandepanne.bsky.social · Aug 23

Excellent and fascinating Veritaseum video about the history and science of rubber. youtu.be/AFXLZ7FEJc4?...

This is the natural disaster to worry about

YouTube video by Veritasium

youtu.be

2

Reposted by Michiel van de Panne

Tom Flood @tomflood.bsky.social · Aug 20

Table For One Please.

side of a building with a billboard that says smoking tables next to non-smoking tables used to be normal too and picture of child biking beside a truck on painted bike lane

7 73 310

Reposted by Michiel van de Panne

The Guardian @theguardian.com · Aug 15

Box, run, crash: China’s humanoid robot games show advances and limitations

Beijing is keen to showcase country’s prowess in robotics, say observers – but some are sceptical about real-world use

www.theguardian.com

7 5 12

Michiel van de Panne @mvandepanne.bsky.social · Aug 13

I often view learned results (inc. RL control policies) as memory-based lookups, substituting for compute. Hence “amortized inference” from the Stats viewpoint. And comes with tradeoffs & limitations on generalization. Arguably great for going from microscopic to efficient macroscopic models?

1

Michiel van de Panne @mvandepanne.bsky.social · Aug 13

2nd worst forest fire season thus far in Canada: 70,000 km^2 burned: the area of New Brunswick or Ireland. Action on reducing emissions, please.

2

Michiel van de Panne @mvandepanne.bsky.social · Aug 13

@mark-carney.bsky.social Time to create a car market that supports EVs, i.e., similar to Europe and Australia. Would love to see Renault 5 EVs here. Keep the current EV targets. The country (and the world) is on fire. And the era of the intertwined Canadian-American car manufacturing is over.

6

Michiel van de Panne @mvandepanne.bsky.social · Aug 13

Its more about using a suitable richly-parameterized differentiable architecture (possibly domain specific), using stochastic gradient descent for optimization, more so than using ReLUs. Hence Gaussian-splats, transformers, etc.

1

Michiel van de Panne @mvandepanne.bsky.social · Aug 1

It’s a great example of the level of skill to strive for!

Michiel van de Panne @mvandepanne.bsky.social · Aug 1

Is this particular video teleop? I’m guessing yes?

1 1