Chengkun Li @CVPR✈️

🇨🇭phd student @EPFL_en, previous sr at google, bytedance | open science

[Chengkun Li @CVPR✈️ @ckli85]

@arnie_hacker whoa! how long are you staying?

[Chengkun Li @CVPR✈️ @ckli85]

@ludovico_bessi couldn't agree more... every trip away just makes me appreciate switzerland more

[Chengkun Li @CVPR✈️ @ckli85]

MuscleMimic was just featured in the @AITinkerers newsletter! Check out the write-up and other top AI demos here: https://t.co/BbiU9R2v8J

link to project: https://t.co/SiiWeOt9Tg

kudos to all collaborators @EPFL @amathislab @mcgillu @nvidia

[Chengkun Li @CVPR✈️ @ckli85]

@breadli428 @GoogleDeepMind congrats! let’s meet in london!

[Michael Andregg @michaelandregg]

We've uploaded a fruit fly. We took the @FlyWireNews connectome of the fruit fly brain, applied a simple neuron model (@Philip_Shiu Nature 2024) and used it to control a MuJoCo physics-simulated body, closing the loop from neural activation to action.

A few things I want to say about what this means and where we're going at @eonsys. 🧵

[Tanishq Mathew Abraham, Ph.D. @iScienceLuvr]

Chinese New Year is rapidly becoming the AI researcher's favorite holiday

[Andrii Maksai @AMaksai]

For a while now, I wanted to check how well frontier models can look at an image and handwrite on top of it. So I did a thing:

InkSlop: Vibe-coded benchmark for Spatial Reasoning with Digital Ink

Link and key points below👇

[Chengkun Li @CVPR✈️ @ckli85]

@breadli428 @NeurIPSConf when is your workshop?

[Chengkun Li @CVPR✈️ @ckli85]

@breadli428 @UnitreeRobotics @leggedrobotics @IrvingRobotics wow, pretty good sim-to-real! just that too bad it's blind 😂

[Horace He @cHHillee]

Lots of sympathy to the Anthropic team 🙏🙏🙏

[Alex Hägele @haeggee]

Long in the making, finally released: Apertus-8B and Apertus-70B, trained on 15T tokens of open data from over 1800 languages. Unique opportunity in academia to work on and train LLMs across the full-stack. We managed to pull off a pretraining run with some fun innovations, ...

[About Switzerland @AbtSwitzerland]

Switzerland has just released Apertus, its first large-scale open-source language model. Developed by EPFL, ETH Zurich and the Swiss National Supercomputing Centre, Apertus sets a new standard for transparent, multilingual and trustworthy AI.

ℹ️ https://t.co/Tsiwtl6eIs

[Prime Intellect @PrimeIntellect]

Introducing the Environments Hub

RL environments are the key bottleneck to the next wave of AI progress, but big labs are locking them down

We built a community platform for crowdsourcing open environments, so anyone can contribute to open-source AGI

[Chengkun Li @CVPR✈️ @ckli85]

Intelligence is grounded in the physical reality of movement

The body can actively compute and simplify complex problems

Scaling embodied experience, not just data

True AI requires interaction, not just observation

[Chengkun Li @CVPR✈️ @ckli85]

Great talk on generative ai and agentic ai by @sirbayes, very inspiring especially the future work section!

[Chengkun Li @CVPR✈️ @ckli85]

@chrisoffner3d maybe joint vision-action modeling better align with animal example here? something like v-jepa ac by @ylecun 🤔

[Chengkun Li @CVPR✈️ @ckli85]

@_k_sridhar Impressive work, it’s super exciting to see the progress! QQ, in this scene, is it possible to interact with the machine or the dough?

[Google DeepMind @GoogleDeepMind]

What if you could not only watch a generated video, but explore it too? 🌐

Genie 3 is our groundbreaking world model that creates interactive, playable environments from a single text prompt.

From photorealistic landscapes to fantasy realms, the possibilities are endless. 🧵

[Chengkun Li @CVPR✈️ @ckli85]

The more you use JAX, the more you feel its power.. like wielding the sharpest knife in your toolbox 🗡️