Arth Dharaskar

Curious about Machine Learning, Neuroscience and Everything in between ✨

[Andrej Karpathy @karpathy]

There's too much happening right now, so here's just a bunch of links

GPT-4 + Medprompt -> SOTA MMLU
https://t.co/Jkp96izfec

Mixtral 8x7B @ MLX nice and clean
https://t.co/75StzY5AHe

Beyond Human Data: Scaling Self-Training for Problem-Solving with Language Models
https://t.co/gOCWjfY7ec

Phi-2 (2.7B), the smallest most impressive model
https://t.co/Fps8tI5QVi

LLM360: Towards Fully Transparent Open-Source LLMs
https://t.co/l6E16GfdIN

Honorable mentions
https://t.co/7GQqiCGHRH
https://t.co/3GZrYPp9KP
https://t.co/Su8iiDksMZ

[Jeff Dean @JeffDean]

I’m very excited to share our work on Gemini today! Gemini is a family of multimodal models that demonstrate really strong capabilities across the image, audio, video, and text domains. Our most-capable model, Gemini Ultra, advances the state of the art in 30 of 32 benchmarks, including 10 of 12 popular text and reasoning benchmarks, 9 of 9 image understanding benchmarks, 6 of 6 video understanding benchmarks, and 5 of 5 speech recognition and speech translation benchmarks. Gemini Ultra is the first model to achieve human-expert performance on MMLU across 57 subjects with a score above 90%. It also achieves a new state-of-the-art score of 62.4% on the new MMMU multimodal reasoning benchmark, outperforming the previous best model by more than 5 percentage points.

Gemini was built by an awesome team of people from @GoogleDeepMind, @GoogleResearch, and elsewhere at @Google, and is one of the largest science and engineering efforts we’ve ever undertaken. As one of the two overall technical leads of the Gemini effort, along with my colleague @OriolVinyalsML, I am incredibly proud of the whole team, and we’re so excited to be sharing our work with you today!

There’s quite a lot of different material about Gemini available, starting with:

Main blog post: https://t.co/NzSycJl7aE

60-page technical report authored by th Gemini Team: https://t.co/CEdMRyYSLo

In this thread, I’ll walk you through some of the highlights.

[Stas Bekman @StasBekman]

This is the first pass on the new chapter for ML Engineering:

The AI Battlefield Engineering - What You Need To Know

https://t.co/6BqzG2PEDb

This a WIP and your feedback for improvement is always welcome.

[Ilya Shabanov @Artifexx]

This is how I use AI (ChatGPT) to improve my academic research.

"AI Prompt Templates" supercharge any task from PDF annotations to writing a paper.

Step by step tutorial with powerful prompts for researchers:
👇

[Arth Dharaskar @_ArthDh]

@hubermanlab Great information! At the start of the video it was mentioned "many people out there think they have a deviated septum ... little bit later we'll talk about how to repair a deviated septum ...", was this part covered? It seems I might have missed it. Thank you! @hubermanlab

[Harrison Kinsley @Sentdex]

A project to keep an eye on: https://t.co/wvyriUffEP

The goal: A truly open ChatGPT like assistant, and beyond, currently targeting a ~24GB-memory-to-run model. You can help right now with the RLHF stage: https://t.co/3RzhRLZXAK

[Zachary Nado @zacharynado]

Excited to announce our Deep Learning Tuning Playbook, a writeup of tips & tricks we employ when designing DL experiments. We use these techniques to deploy numerous large-scale model improvements and hope formalizing them helps the community do the same! https://t.co/vDhSwZyHJm

[Riley Goodside @goodside]

POV: You're a Senior Data Engineer at Twitter. Elon asks what you've done this week. You've done nothing.

Frantically, you open ChatGPT.

[Karen X. Cheng @karenxcheng]

1/ Now you can create "drone" shots from your phone footage, thanks to NeRF
Collab with @justlv / software @nerfstudioteam / see below for our process
#NeRF #neuralrendering #artificialintelligence #instantNeRF

[Mark Tenenholtz @marktenenholtz]

Python 3.11 is out! 🎉

This is one of the most exciting releases in a while, including significant speed upgrades and better error messages.

Here's what's new:

[Shreya Shankar @sh_reya]

I probably should have written this years ago, but here are some MLOps principles I think every ML platform (codebase, data management platform) should have: 1/n

[PinataHub_Bot @PinataHub_Bot]

If you don't remember retweeting this, it means that you have leaked your Twitter Access Token in a public GitHub repository. Not the best practice, right?
For details, read our latest article: https://t.co/6WBC6DRNDS
#InfoSec #CyberSecurity #GitHub

[Google DeepMind @GoogleDeepMind]

Today with @emblebi, we're launching the #AlphaFold Protein Structure Database, which offers the most complete and accurate picture of the human proteome, doubling humanity’s accumulated knowledge of high-accuracy human protein structures - for free: https://t.co/vtBGmTkKhy 1/

[Janelle Shane @JanelleCShane]

bias laundering edition

[François Fleuret @francoisfleuret]

I wrote a piece of code that, given an input and output tensor sizes and a depth, computes all the sequences of kernel sizes / strides that do it!

No padding. I hate paddings.

https://t.co/2m2JdMiCKR

@PyTorch

[David Duvenaud @DavidDuvenaud]

Gradient descent in differentiable games rotates around solutions instead of converging. For instance, in GANs. We solve this with a simple trick: complex momentum damps the oscillations.

https://t.co/6ql2bqSX21
With @jonLorraine9 @davidjesusacu @PaulVicol

[Keenan Crane @keenanisalive]

Excited to share a new #SIGGRAPH2021 paper with @markgillesie81 and Boris Springborn that is a pretty big breakthrough in mesh parameterization: https://t.co/M0EZGuIGfL

In short: no matter how awful your mesh is, we compute beautiful high-quality texture coordinates.

(1/n)

[Animesh Garg @animesh_garg]

As @NeurIPSConf #NeurIPS2021 deadline approaches, I've consolidated some technical writing tips.

Succinct answers to each of the questions,
Voila, you have a crisp Introduction!

Expand on prior work, intuition/tech details, & expts for the rest ensuring the why is always clear

[Aran Komatsuzaki @arankomatsuzaki]

TPU VM is out of alpha!

Cloud TPU VMs run on the TPU host machine and offer significantly better performance and usability when working with TPUs (e.g. you can store data directly at TPU VMs).

https://t.co/fZW2Pv69nT

[AK @_akhaliq]

LASR: Learning Articulated Shape Reconstruction from a Monocular Video
pdf: https://t.co/0oB93bnY6j
abs: https://t.co/gi1lfVL1q4
project page: https://t.co/AK9BxWgMqa