Mikhail Jacob

Researcher in getting computers to stop being boring.

[Mikhail Jacob @MikhailJacob]

@pcastr Resolution Games, where I led the ML team previously, shipped two VR games - Racket Club and Home Sports (pickleball, hockey, badminton), where opponents/bots were trained using imitation learning and RL respectively. So both games but they provide actual financial value. 😅

[Google DeepMind @GoogleDeepMind]

Introducing Genie 2: our AI model that can create an endless variety of playable 3D worlds - all from a single image. 🖼️

These types of large-scale foundation world models could enable future agents to be trained and evaluated in an endless number of virtual environments. → https://t.co/1dzB2BUlWo

[Sam Devlin @smdvln]

Our team @MSFTResearch is hiring for a 2-year AI Residency role in the area of learning to control embodied agents, with the goal of informing future applications in Gaming and Robotics. For more details and to formally apply, please visit: https://t.co/z4Cv4GRsl6

[Mikhail Jacob @MikhailJacob]

@SarahMillican75 Literally a day after going on sale, it's sold out anywhere close to Vancouver. 70+ scalpers reselling tickets on StubHub for up to 700 dollars. So disappointing! Our ticket sale system is truly broken.

[Alice Ruppert 🐴🎮 @MaliceDaFirenze]

We're making a game where you play a courier rider in 13th century Mongolia

No combat, just you, the horse(s) you've tamed, bred and trained and the vast wilderness.

Would you play this?

[Mikhail Jacob @MikhailJacob]

@UpolEhsan Congratulations Upol! Exciting times ahead. All the best with the new start. 😀

[Mikhail Jacob @MikhailJacob]

The Coalition is hiring! If you're as excited about the future of games and ML as we are, join us and help make it happen! https://t.co/Ov9eQRpVCE

#ML #games #AI #LLM #RL #agents #UnrealEngine #NLP #CV

[elvis @omarsar0]

// Searching for Best Practices in RAG //

Cool paper showing the best practices for building effective RAG workflows.

Proposes strategies that focus on performance and efficiency, including emerging multimodal retrieval techniques.

[AI at Meta @AIatMeta]

📣 New research from GenAI at Meta, introducing Meta 3D Gen: A new system for end-to-end generation of 3D assets from text in <1min.

Meta 3D Gen is a new combined AI system that can generate high-quality 3D assets, with both high-resolution textures and material maps end-to-end, producing results that are superior to existing solutions — at 3-10x the speed of existing work in this space.

Details in the technical report ➡️ https://t.co/lrkShaRCju

[Shreya Shankar @sh_reya]

wrote a very long piece on data flywheels for LLM applications. hope you like it: https://t.co/EWGE9h6oVc

[Xiao-Xiao Long @xxlong0]

Extract accurate depth from your 3D gaussians! Fast, accurate, and easy to use.

Code of RaDe-GS is released!
project page:https://t.co/6NxCxw0ykR
code:https://t.co/3EinZkt7PX
paper:https://t.co/z38zSkfGKs

No need to change your 3D GS representations, we propose a play-and-plug rasterizer to extract accurate depth and normal from 3D gaussians, leading to the possibility of high-quality geometry reconstruction.

It's quite efficient and fast, achieving 0.69mm errors on DTU dataset in only 5 minutes.
Very easy to adopt the rasterizer in your any 3dgs projects.

[Xin Kong @XinKong_IC]

Tired of single image to 3D? Check out EscherNet tomorrow @CVPR that can take flexible number of views for 3D generation!

THURSDAY, JUNE 20
ORAL: 9:00-10:30, SUMMIT BALLROOM (TOP FLOOR)
POSTER: 10:30-12:00, ARCH 4A-E, #69

Try our @Gradio online demo
https://t.co/PAaJdgRaZt

[Pratyush Maini @pratyushmaini]

1/We've nailed a framework to reliably detect if an LLM was trained on your dataset: LLM Dataset Inference.

After over a year of thinking of writing about how hard this is, we had a breakthrough that made me quite literally jump from my seat!

📝: https://t.co/B9jvMsoNGW Long🧵

[Luma @LumaLabsAI]

🧵1/10: It’s been 2 days since #LumaDreamMachine launched and we’re blown away by the outpouring of creative expression we’ve seen from the community. Here are a few of our favorite creations that have inspired us. Like what you see? Try Dream Machine today for free → https://t.co/F96dyjWcoD 🐘: @mantlehood

[Runway @runwayml]

Introducing Gen-3 Alpha: Runway’s new base model for video generation.

Gen-3 Alpha can create highly detailed videos with complex scene changes, a wide range of cinematic choices, and detailed art directions.

https://t.co/YQNE3eqoWf

(1/10)

[JD | RoyalCities @RoyalCities]

I spent the past few days testing &amp; finetuning the new StableAudio model and I'm at the point that I can say this will truly open the doors to musicians basically having an infinite sample generator.

Here is what I've found.

[bit of a long 🧵- Spoilers it gets sorta wild👇]:

[Charity Majors @mipsytipsy]

Because writing code is the easiest part of software engineering, and just the start of software systems.

https://t.co/zW7YUa3rfs

Big thanks to @copyconstruct and @hillelogram, whose pieces I linked, and all of the many people who read drafts of this and made it better. &lt;3

[Lovis Odin @OdinLovis]

With @Dogstudio comfyUi crazy 3D animation workflow One prompt &gt; SDXL image génération &gt; instantmesh and zero123 to generate 3D &gt; upscale of texture &gt; animation of orbit caméra &gt; vid2vid with SD 1.5 LCM &gt; latent upscale &gt; boummm haha
With Amazing node of @MrForExample

[AK @_akhaliq]

Look Once to Hear

Target Speech Hearing with Noisy Examples

In crowded settings, the human brain can focus on speech from a target speaker, given prior knowledge of how they sound. We introduce a novel intelligent hearable system that achieves this capability, enabling target speech hearing to ignore all interfering speech and noise, but the target speaker. A naive approach is to require a clean speech example to enroll the target speaker. This is however not well aligned with the hearable application domain since obtaining a clean example is challenging in real world scenarios, creating a unique user interface problem. We present the first enrollment interface where the wearer looks at the target speaker for a few seconds to capture a single, short, highly noisy, binaural example of the target speaker. This noisy example is used for enrollment and subsequent speech extraction in the presence of interfering speakers and noise. Our system achieves a signal quality improvement of 7.01 dB using less than 5 seconds of noisy enrollment audio and can process 8 ms of audio chunks in 6.24 ms on an embedded CPU. Our user studies demonstrate generalization to real-world static and mobile speakers in previously unseen indoor and outdoor multipath environments. Finally, our enrollment interface for noisy examples does not cause performance degradation compared to clean examples, while being convenient and user-friendly. Taking a step back, this paper takes an important step towards enhancing the human auditory perception with artificial intelligence.

[Mikhail Jacob @MikhailJacob]

@MatthewGuz @AmiiThinks I saw an ad for this event in the elevator at work. Have a lovely time!