Florian Stimberg

Research Engineer at @DeepMind

[Florian Stimberg @FStimberg]

Great thread from David on our newly released report on SynthID-image and all the lessons we learned working on SynthID for the last few years.

[David Stutz @davidstutz92]

We released our report on SynthID-image, outlining what it takes to watermark over ten billion images generated with @GoogleDeepMind AI models and making this technology available to external partners. I really enjoyed writing this report – a thread:

https://t.co/65Z9e0A8AS

[Florian Stimberg @FStimberg]

SynthID, the project I've been working on for more than a year will now also work on text and video.

[Google DeepMind @GoogleDeepMind]

SynthID will now expand to 2️⃣ new modalities: text and video.

While this tool isn’t a silver bullet for identifying AI-generated content, it’s an important building block for helping millions of people understand the provenance of synthetic content. → https://t.co/CS0s9NaAfX

[Google DeepMind @GoogleDeepMind]

SynthID will now expand to 2️⃣ new modalities: text and video.

While this tool isn’t a silver bullet for identifying AI-generated content, it’s an important building block for helping millions of people understand the provenance of synthetic content. → https://t.co/CS0s9NaAfX

[Michael Chang @mmmbchang]

Gemini and I also got a chance to watch the @OpenAI live announcement of gpt4o, using Project Astra! Congrats to the OpenAI team, super impressive work!

[Michael Chang @mmmbchang]

It's such an honor to work on Project Astra with such an amazing team from across Gemini and Google DeepMind! While the #GoogleIO keynote was happening we had a last minute idea of watching the keynote with Project Astra. Check it out!

[Olivia W @OliviaW47557022]

Student researcher position applications are open at Google Deepmind!

I'm hosting a SR in the intersection of bias and generative models. If you're an interested PhD student please reach out!

https://t.co/dKPbGByGEb…

[Florian Stimberg @FStimberg]

Happy to see the project I worked on for most of the year finally launched!

[Google DeepMind @GoogleDeepMind]

We’re excited to launch 𝗦𝘆𝗻𝘁𝗵𝗜𝗗 today with @GoogleCloud: a digital tool to watermark and identify AI-generated images. 🖼️

It will be available on Imagen, one of @Google’s latest text-to-image models.

Here’s how it works. 🧵 https://t.co/p8s5XK5Ekr #GoogleCloudNext

[Google DeepMind @GoogleDeepMind]

As society becomes more digital, it’s critical to improve the code powering the world's computing.

Today in @Nature, we present AlphaDev, an AI system using reinforcement learning to discover enhanced computer science algorithms.

How does it work? 🧵 https://t.co/2ukGaPey5w

[Florian Stimberg @FStimberg]

@tvaziri David Simon wrote a long blog post about the work that went into the remake of The Wire and the change in aspect ratio: https://t.co/64txomUsHQ

[Sven Gowal @sgowal]

DeepMind is dedicated to bringing Safe, Reliable & Trustworthy AI to the real world. Consider applying and joining our amazing team.

https://t.co/WiTJ9RSqCR

https://t.co/vGr0gty07c

[Alan Karthikesalingam @alan_karthi]

So happy to share #MedPaLM2 - our team's evolution of Med-PaLM. A new state of art for medical question-answering!

Med-PaLM 2 scores 86.5% on MedQA-USMLE, exceeding Med-PaLM's score by >19% 🤯, & 81.8% on PubMedQA...

More here: https://t.co/J83vdrpxFt

[Florian Stimberg @FStimberg]

@karamcnair @ZoeSchiffer @TerriGerstein Benevity does donations for Google in the UK as well.

[Florian Stimberg @FStimberg]

None of the models we tested achieved more than 62% accuracy on the worst case over the 3 hold-out obfuscations. How far can this be pushed, especially for smaller models?

[Florian Stimberg @FStimberg]

How easy is it for adversaries to hide image content from classifiers through obfuscations? Our new benchmark allows you to evaluate this!
Dataset and evaluation code: https://t.co/kWKcqwFTMy
Paper: https://t.co/U1zXMy1Zng.
Joint work between @DeepMind, @Google & @GoogleAI 🧵

[Florian Stimberg @FStimberg]

@DeepMind @Google @GoogleAI We hope this benchmark will add a new dimension to evaluating robustness and will help research on how to make models generalize better to obfuscations.

[Cosmin Paduraru @CauseMean]

Excited to share the details of our work at @DeepMind on using reinforcement learning to help large-scale commercial cooling systems save energy and run more efficiently: https://t.co/i2M1cghANJ.

Here’s what we found 🧵

[Google DeepMind @GoogleDeepMind]

Our team tested a number of methods under a variety of distribution shifts. To learn the conclusions or test your own method, join the #ICLR2022 poster session today or the oral presentation on Wednesday.

Poster: https://t.co/pMYGAAjTHu
Oral: https://t.co/kdyiYtb0I4 1/