61
Atlas robot does parkour
By training Atlas to maneuver its way through complex parkour courses, our engineers push the humanoid robot to its limits.
62
Useful algorithms that are not optimized by Jax, PyTorch, or TensorFlow
In some previous blog posts we described in details how one can generalize automatic differentiation to give automatically stability enhancements and all sorts of other niceties by incorporating graph transformations into code generation. However, one of the things which we didn’t go into too much is the limitation of…
63
GPT-3 tries pickup lines
Once upon a time I decided to train a neural net to generate pickup lines. Once I started collecting the training data I began to regret it when I saw how awful the existing lines were. Turns out I needn’t have worried. The neural net I used was so small and clueless that
64
High-performance speech recognition with no supervision at all
To enable speech recognition technology for many more languages spoken around the globe, Facebook AI is releasing wav2vec Unsupervised, a new method to train models with no supervision whatsoever. Wav2vec Unsupervised rivals the performance of the best supervised systems from just a few years ago.
65
66
Are deep neural networks dramatically overfitted? (2019)
<meta http-equiv=
67
Machine learning’s crumbling foundations
Technological debt is insidious, a kind of socio-infrastructural subprime crisis that's unfolding around us in slow motion. Our digital infrastructure is built atop layers and layers and layers of code that's insecure due to a combination of bad practices and bad frameworks.
68
Science fiction hasn’t prepared us to imagine machine learning
It resembles the Library of Babel more than HAL.
69
Ideas in statistics that have powered AI
Andrew Gelman, a statistics professor at Columbia, and Aki Vehtari, a computer science professor at Finland’s Aalto University, recently published a list of the most important statistical ideas in the last 50 years. Here, they break it down in easy-to-understand terms.
70
Game theory as an engine for large-scale data analysis
Modern AI systems approach tasks like recognising objects in images and predicting the 3D structure of proteins as a diligent student would prepare for an exam. By training on many example problems, they minimise their mistakes over time until they achieve success. But this is a solitary endeavour and only one of the…
71
Machine learning saves us $1.7M a year on document previews
Recently, we translated the predictive power of machine learning (ML) into $1.7 million a year in infrastructure cost savings by optimizing how Dropbox generates and caches document previews. Machine learning at Dropbox already powers commonly-used features such as search , file and folder suggestions , and OCR in…
72
73
74
PyTorch 1.10
We are excited to announce the release of PyTorch 1.10. This release is composed of over 3,400 commits since 1.9, made by 426 contributors. We want to sincerely thank our...
75
How image search works at Dropbox
Photos are among the most common types of files in Dropbox, but searching for them by filename is even less productive than it is for text-based files. When you're looking for that photo from a picnic a few years ago, you surely don't remember that the filename set by your camera was 2017-07-04 12.37.54.jpg .
76
Intro to Deep Learning
I just sat down this morning and organized all deep learning related videos I recorded in 2021. I am sure this will be a useful reference for my future…
77
Why machine learning struggles with causality
Machine learning systems thrive on finding and exploiting statistical regularities. This is why they can't do causal reasoning.
78
Hopfield Networks Is All You Need
Blog post
79
80
Gopher – A 280B parameter language model
Language, and its role in demonstrating and facilitating comprehension - or intelligence - is a fundamental part of being human. It gives people the ability to communicate thoughts and concepts, express ideas, create memories, and build mutual understanding. These are foundational parts of social intelligence. It’s…
81
Data scientists shouldn’t need to know Kubernetes
[Hacker News discussion, Twitter thread]
82
Self-supervised learning: The dark matter of intelligence
How can we build machines with human-level intelligence? There’s a limit to how far the field of AI can go with supervised learning alone. Here's why self-supervised learning is one of the most promising ways to make significant progress in AI.
83
Protein complex prediction with AlphaFold-Multimer
Build AI responsibly to benefit humanity
84
Fun and dystopia with AI-based code generation using GPT-J-6B
At the least, AI-generated code is much more readable than the average human’s.
85
Baking with machine learning (2020)
Like many people, I’ve been entertaining myself at home by baking a ton. I’ve gotten pretty good at following recipes, but I decided I wanted to take things one step further and understand the science behind what differentiates a cake from a bread or a cookie. I also like machine learning so I thought: what if I could…
86
Building a recommendation engine inside Postgres with Python and Pandas (2020)
Learn how you can leverage Python and Pandas from directly inside PostgreSQL to build your own recommendation engine.
87
88
89
Gaussian Processes from Scratch (2019)
This post explores some concepts behind Gaussian processes, such as stochastic processes and the kernel function. We will build up deeper understanding of Gaussian process regression by implementing them from scratch using Python and NumPy.
90