yingzhen
@liyzhen2
teaching machines 🤖 to learn 🔍 and fantasize 🪄
now 🇬🇧 @ImperialCollege @ICComputing 🇸🇬 @NTUsg @NTU_ccds
ex 🇬🇧 @MSFTResearch @CambridgeMLG
now 🇬🇧 @ImperialCollege @ICComputing 🇸🇬 @NTUsg @NTU_ccds
ex 🇬🇧 @MSFTResearch @CambridgeMLG
164 Following 4.5K Followers 702 Posts
Latest posts 5
The bitter lesson is never about brute force, but rather about advancing methods that can leverage compute **efficiently**
Brute force computation is not enough for intelligence. Clever new methods are needed. The science of AI is far from finished. IMHOView quoted post on X
Congrats to Ed!
Been co-supervising a PhD student together, I can tell Ed is definitely world class researcher in robot learning, and he’s very much down to the technical details
I always find his lab’s robot demos of “instant” learning with very high data efficiency stunning
Been co-supervising a PhD student together, I can tell Ed is definitely world class researcher in robot learning, and he’s very much down to the technical details
I always find his lab’s robot demos of “instant” learning with very high data efficiency stunning
Today we’re out of stealth, and we’ve raised a $40m seed round co-led by @creandum and @usv, to bring Instant Robot Learning to life.
For the past few months, we‘ve been busy building a new company.
Machenta.
This follows results from a new kind of robot foundation model we’ve built, which I’m so excited to share with you today!
Usually, it takes hours/days of data collection and training to teach a robot.
We’ve built a model that learns tasks instantly: one demonstration, no further training.
A model trained entirely in simulation, without a single pixel of real-world data.
A model where teaching robots is so quick and easy, it feels like physical vibe coding.
A model for robots that learn like humans do, from someone like you.
Welcome to the era of Instant Robot Learning.
Thank you @Kathi_Wilhelm, @rebeccakaden, @hkanji, @jonno_evans and all of our angels for your support!
@Machenta_
Learn more: machenta.com/blog/instant-r…
See below for what you can do with Instant Robot Learning!View quoted post on X
This is now also accepted at #NeurIPS2026
Our own answer: structured coupling
arxiv.org/abs/2605.07676
- flow matching with VAE-based coupling
- VAE encoder & flow sharing networks
- VAE decoder init. + flow refinement for sampling
flow matching 🤝 VAEs -> good representation & sample quality🚀View quoted post on X