@hardmaru - AI Dynamics - Page 45 of 53

Hyperbolic Deep Reinforcement Learning ICLR2023 Paper and Blog Post

By

–

30 April 2023 16h12

Updated #ICLR2023 final version of the Hyperbolic Deep Reinforcement Learning paper: https://
openreview.net/forum?id=TfBHF
Lgv77
… Michael Bronstein (
@mmbronstein
) wrote a blog post about their fantastic idea: https://
towardsdatascience.com/hyperbolic-dee
p-reinforcement-learning-b2de787cf2f7
…

→ View original post on X — @hardmaru

30 April 2023

Real-world MuJoCo humanoid agents in action

By

@hardmaru

–

30 April 2023 14h35

These are literally IRL versions of MuJoCo humanoid agents…

Looking forward to see the 2 vs 2 version: https://t.co/Glvcr9SUjP
— hardmaru (@hardmaru) 30 avril 2023

These are literally IRL versions of MuJoCo humanoid agents… Looking forward to see the 2 vs 2 version:

→ View original post on X — @hardmaru

30 April 2023

Deep RL Synthesis of Movement Skills for Miniature Humanoid Robot

By

@hardmaru

–

29 April 2023 16h54

This is such a cool project:

“We investigate whether Deep RL is able to synthesize sophisticated and safe movement skills for a low-cost, miniature humanoid robot that can be composed into complex behavioral strategies in dynamic environments.”

Video of behavior and skills → pic.twitter.com/COIdyFpZNR
— hardmaru (@hardmaru) 29 avril 2023

This is such a cool project: “We investigate whether Deep RL is able to synthesize sophisticated and safe movement skills for a low-cost, miniature humanoid robot that can be composed into complex behavioral strategies in dynamic environments.” of behavior and skills →

→ View original post on X — @hardmaru

29 April 2023

DeepMind’s Bipedal Robot Soccer Players with Deep Reinforcement Learning

By

@hardmaru

–

29 April 2023 16h43

Pushing around these little robot soccer players, from DeepMind’s “Learning Agile Soccer Skills for a Bipedal Robot with Deep Reinforcement Learning” paper.https://t.co/KzXryyfhCH https://t.co/VOnhLk0DyY pic.twitter.com/GDcCuRm8KM
— hardmaru (@hardmaru) 29 avril 2023

Pushing around these little robot soccer players, from DeepMind’s “Learning Agile Soccer Skills for a Bipedal Robot with Deep Reinforcement Learning” paper. https://
arxiv.org/abs/2304.13653 https://
sites.google.com/view/op3-soccer

→ View original post on X — @hardmaru

29 April 2023

Try out DeepFloyd IF model on Hugging Face Spaces

By

@hardmaru

–

28 April 2023 19h03

Try out the model here → https://
huggingface.co/spaces/DeepFlo
yd/IF
…

→ View original post on X — @hardmaru

28 April 2023

DeepFloyd IF: Open-source text-to-image model with zero-shot inpainting

By

@hardmaru

–

28 April 2023 7h36

DeepFloyd IF: An open-source text-to-image model by our @DeepfloydAI team @StabilityAI

Check out the examples, with amazing zero-shot inpainting results at high resolution: https://t.co/9TbHDzyGOy pic.twitter.com/YjIAyROCpW
— hardmaru (@hardmaru) 28 avril 2023

DeepFloyd IF: An open-source text-to-image model by our @DeepfloydAI team @StabilityAI Check out the examples, with amazing zero-shot inpainting results at high resolution: https://
github.com/deep-floyd/IF

→ View original post on X — @hardmaru

28 April 2023

Open AI Models as Better Research Baselines Than Closed Systems

By

@hardmaru

–

25 April 2023 18h58

Closed AI Models Make Bad Baselines @annargrs “The more people cite and study the best available open solutions, the more we incentivize open and transparent research, and the more likely it is that the next open solution will be much better.”

→ View original post on X — @hardmaru

25 April 2023