[ML News] Google introduces Pathways | OpenAI solves Math Problems | Meta goes First Person

Subscribers:
284,000
Published on ● Video Link: https://www.youtube.com/watch?v=G7-fRGaCZts



Duration: 36:46
58,629 views
0


#pathways #mlnews #ego4d

Your irregular dose of Machine Learning News.

OUTLINE:
0:00 - Intro
0:20 - Sponsor: Weights & Biases
2:10 - Google Introduces Pathways AI Architecture
6:30 - OpenAI trains Language Models to do High School Math
8:25 - Sam Altman says Neural Networks truly learn
9:35 - Google AI researchers frustrated with lawyers
12:10 - DeepMind RL Lecture Series 2021
12:40 - Fashion Store sells Adversarial Patches
13:15 - A viable method to remove the GIL from CPython
15:05 - BigScience Workshop releases T0
17:40 - Huggingface Hub Dataset Viewer
18:10 - Scite classifies scientific citations
19:25 - Facebook AI Ego4D dataset & challenges
21:50 - Tesla Dojo Configurable Floating Point Spec
23:10 - Windows releases PyTorch-DirectML for Deep Learning on DirectX GPUs
23:50 - Helpful Things
33:00 - Traders use ML to analyze CEOs' language
34:20 - Cadbury creates DeepFake ads for local Indian businesses
35:25 - This Shoe Does Not Exist

Sponsor: Weights & Biases
https://wandb.com

References:
Google Introduces Pathways AI Architecture
https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture/?utm_source=pocket_mylist

OpenAI trains Language Models to do High School Math
https://openai.com/blog/grade-school-math/
https://arxiv.org/abs/2110.14168

Sam Altman says Neural Networks truly learn
https://twitter.com/sama/status/1450857134648823809?s=09&t=KazQPHo6Epn0M6ihs4DqHg&utm_source=pocket_mylist

Google AI researchers frustrated with lawyers
https://archive.ph/lsQJJ#selection-2855.0-2855.294

DeepMind RL Lecture Series 2021
https://deepmind.com/learning-resources/reinforcement-learning-series-2021

Fashion Store sells Adversarial Patches
https://twitter.com/naotokui/status/1450673712722702340

A viable method to remove the GIL from CPython
https://lwn.net/Articles/872869/

BigScience Workshop releases T0
https://bigscience.huggingface.co/
https://arxiv.org/abs/2110.08207
https://huggingface.co/bigscience/T0pp

Huggingface Hub Dataset Viewer
https://twitter.com/huggingface/status/1454079471154257923

Scite classifies scientific citations
https://scite.ai
https://direct.mit.edu/qss/article/doi/10.1162/qss_a_00146/102990/scite-A-smart-citation-index-that-displays-the

Facebook AI Ego4D dataset & challenges
https://ai.facebook.com/blog/teaching-ai-to-perceive-the-world-through-your-eyes

Tesla Dojo Configurable Floating Point Spec
https://tesla-cdn.thron.com/static/SBY4B9_tesla-dojo-technology_OPNZ0M.pdf?xseo=&response-content-disposition=inline%3Bfilename%3D%22tesla-dojo-technology.pdf%22

Windows releases PyTorch-DirectML for Deep Learning on DirectX GPUs
https://devblogs.microsoft.com/windowsai/introducing-pytorch-directml-train-your-machine-learning-models-on-any-gpu/

Helpful Things
https://github.com/achaiah/pywick?utm_source=pocket_mylist
https://github.com/orybkin/lexa-benchmark?utm_source=pocket_mylist
https://orybkin.github.io/lexa/
https://twitter.com/danijarh/status/1438137568688807942?utm_source=pocket_mylist
https://github.com/RobertTLange/mle-hyperopt
https://keras.io/examples/vision/mobilevit/?utm_source=pocket_mylist
https://twitter.com/osanseviero/status/1451929248231563265?utm_source=pocket_mylist
https://huggingface.co/spaces/flax-community/image-captioning
https://huggingface.co/transformers/master/model_doc/visionencoderdecoder.html
https://github.com/facebookresearch/bitsandbytes
https://arxiv.org/abs/2110.11216
https://arxiv.org/pdf/2110.11216.pdf
https://github.com/facebookresearch/xformers
https://superbbenchmark.org/
https://arxiv.org/abs/2110.07731
https://github.com/BaguaSys/bagua?utm_source=pocket_mylist
https://github.com/cgarciae/treex
https://jax.readthedocs.io/en/latest/pytrees.html

Traders use ML to analyze CEOs' language
https://www.reuters.com/technology/ai-can-see-through-you-ceos-language-under-machine-microscope-2021-10-20/

Cadbury creates DeepFake ads for local Indian businesses
https://www.bgr.in/entertainment/shah-rukh-khan-not-just-a-cadbury-ad-twitter-diwali-celebration-1016913/

This Shoe Does Not Exist
https://www.thisshoedoesnotexist.com/


Links:
TabNine Code Completion (Referral): http://bit.ly/tabnine-yannick
YouTube: https://www.youtube.com/c/yannickilcher
Twitter: https://twitter.com/ykilcher
Discord: https://discord.gg/4H8xxDF
BitChute: https://www.bitchute.com/channel/yannic-kilcher
Minds: https://www.minds.com/ykilcher
Parler: https://parler.com/profile/YannicKilcher
LinkedIn: https://www.linkedin.com/in/ykilcher
BiliBili: https://space.bilibili.com/1824646584

If you want to support me, the best thing to do is to share out the content :)

If you want to support me financially (completely optional and voluntary, but a lot of people have asked for this):
SubscribeStar: https://www.subscribestar.com/yannickilcher
Patreon: https://www.patreon.com/yannickilcher




Other Videos By Yannic Kilcher


2021-12-24[ML News] AI learns to search the Internet | Drawings come to life | New ML journal launches
2021-12-21[ML News] DeepMind builds Gopher | Google builds GLaM | Suicide capsule uses AI to check access
2021-11-27Implicit MLE: Backpropagating Through Discrete Exponential Family Distributions (Paper Explained)
2021-11-25Peer Review is still BROKEN! The NeurIPS 2021 Review Experiment (results are in)
2021-11-24Parameter Prediction for Unseen Deep Architectures (w/ First Author Boris Knyazev)
2021-11-20Learning Rate Grafting: Transferability of Optimizer Tuning (Machine Learning Research Paper Review)
2021-11-18[ML News] Cedille French Language Model | YOU Search Engine | AI Finds Profitable MEME TOKENS
2021-11-15Gradients are Not All You Need (Machine Learning Research Paper Explained)
2021-11-12[ML News] Microsoft combines Images & Text | Meta makes artificial skin | Russians replicate DALL-E
2021-11-10Autoregressive Diffusion Models (Machine Learning Research Paper Explained)
2021-11-05[ML News] Google introduces Pathways | OpenAI solves Math Problems | Meta goes First Person
2021-11-03EfficientZero: Mastering Atari Games with Limited Data (Machine Learning Research Paper Explained)
2021-10-31[YTalks] Siraj Raval - Stories about YouTube, Plagiarism, and the Dangers of Fame (Interview)
2021-10-29[ML News] NVIDIA GTC'21 | DeepMind buys MuJoCo | Google predicts spreadsheet formulas
2021-10-29[ML News GERMAN] NVIDIA GTC'21 | DeepMind kauft MuJoCo | Google Lernt Spreadsheet Formeln
2021-10-27I went to an AI Art Festival in Geneva (AiiA Festival Trip Report)
2021-10-24Symbolic Knowledge Distillation: from General Language Models to Commonsense Models (Explained)
2021-10-21I took a Swiss train and it was awesome! Train Seat Review - SBB InterCity 1 - Geneva to St. Gallen
2021-10-20[ML News] Microsoft trains 530B model | ConvMixer model fits into single tweet | DeepMind profitable
2021-10-07[ML News] DeepMind does Nowcasting | The Guardian's shady reporting | AI finishes Beethoven's 10th
2021-10-06Grokking: Generalization beyond Overfitting on small algorithmic datasets (Paper Explained)



Tags:
deep learning
machine learning
arxiv
explained
neural networks
ai
artificial intelligence
paper
google
google ai
google pathways
jeff dean
pathways model
sparse neural network
meta
meta ai
ego4d
sam altman
openai
openai math
language model math
t0
tzero
bigscience
bigsciencew
deepmind
deepmind lecture series
huggingface
huggingface hub
dataset viewer
machine learning news
tech news