Avi

Turning ideas into reality 🔁 | Human Perception | Ex NYU Courant & Center for Data Science, CTO @ Youth HealthTech, Google AR/VR, Samsung Research

[Prof. Feynman @ProfFeynman]

We have this terrible struggle to try to explain things to people who have no reason to want to know.

[Avi @Being_Avinav]

Misinformation vs Information. #ResponsibleAI #AIPerception

[Ars Technica @arstechnica]

AI Overviews hallucinates that Airbus, not Boeing, involved in fatal Air India crash https://t.co/CqWCfo2Ljv

[BBC News (World) @BBCWorld]

First view of the Sun's south pole filmed by spacecraft https://t.co/jTiSYiPbGg

[DailyPapers @HuggingPapers]

New paper!

QuickVideo accelerates long-video understanding with a system-algorithm co-design, achieving up to 3.5x speedup!

[Avi @Being_Avinav]

@ai_for_success @VraserX It feels like monopolization through freeware distribution.

[Avi @Being_Avinav]

@MKBHD Hi Marques
How does it perform with non fitting/ loose/ layered clothes for both reference and generated clothes?

[DailyPapers @HuggingPapers]

New method for robust and efficient LLM-based NLG evaluation

An inversion learning method that learns effective reverse mappings from model outputs back to their input instructions, enabling the automatic generation of highly effective, model-specific evaluation prompts.

[Tom Yeh @ProfTomYeh]

With the threat of layoffs everywhere, my students asked me: "How do we make ourselves invaluable on an AI/ML team?" 🤔 I told them 3 things: Hint: It's not coding. It's not knowing the latest API frameworks either. 👇

[Amin Nejatbakhsh @aminejat]

Excited to announce that our paper on "Comparing noisy neural population dynamics using optimal transport distances" has been selected for an oral presentation in #ICLR2025 (1.8% top papers). Check the thread for paper details (0/n).

Presentation info: https://t.co/31IcGuWSiA.

[Avi @Being_Avinav]

Always question never doubt :)

[Ricky T. Q. Chen @RickyTQChen]

We are presenting 3 orals and 1 spotlight at #ICLR2025 on two primary topics:

On generalizing the data-driven flow matching algorithm to jump processes, arbitrary discrete corruption processes, and beyond.

And on highly scalable algorithms for reward-driven learning settings.

[ScienceAlert @ScienceAlert]

One of the earliest warnings of cognitive decline.
https://t.co/cHPL3H9Nxu

[Towards Data Science @TDataScience]

Beyond pattern recognition... can CNNs reason? @SalvatoreRaieli dives into the limitations of standard CNNs and how meta-learning empowers them to learn abstract concepts, leading to better generalization.

https://t.co/7RPeR1wbLX

[Avi @Being_Avinav]

@GoogleHealth @GoogleAI @Google How do we know that this technology is non-invasive?

[Justus Thies @JustusThies]

You are working on photo-realistic human avatars? Then, join the #CVPR2025 challenge and show how good your avatars are!

Submission instructions:
https://t.co/ArnmOgqdW2

@CVPR

[Kosta Derpanis @CSProfKGD]

I remember seeing backprop as a senior in undergrad in the late 90s, frankly I couldn’t follow. Much of the prezi was tied to a specific architecture with sigmoids. It was only years later when I saw the modular approach did it make sense!

[Gabriel Peyré @gabrielpeyre]

Oldies but goldies: J Koenderink, The structure of images, 1984. Introduces (with A. Witkin) the notion of scale space, i.e., the multi-scale analysis of signals and images, which is at the heart of many vision architectures. https://t.co/VwSdEI2o7Y

[Avi @Being_Avinav]

@elonmusk Try Try and you will succeed!!

[Sam Bowman @sleepinyourhat]

I’m starting an AI safety research group at NYU. Why? (🧵)