61
The deepest problem with deep learning
Some reflections on an accidental Twitterstorm, the future of AI and deep learning, and what happens when you confuse a schoolbus with a…
62
Playing Mortal Kombat with TensorFlow.js
While experimenting with enhancements of the prediction model of Guess.js, I started looking at deep learning. I’ve focused mainly on recurrent neural networks (RNNs), specifically LSTM because of their “unreasonable effectiveness” in the domain of Guess.js. In the same time, I started playing with convolutional…
63
A Long Peek into Reinforcement Learning
<meta http-equiv=
64
65
Dissecting Reinforcement Learning
Explaining the basic ideas behind reinforcement learning. In particular, Markov Decision Process, Bellman equation, Value iteration and Policy Iteration algorithms, policy iteration through linear algebra methods. It includes full working code written in Python.
66
67
68
Interactive GPU Programming, Part 2: Hello OpenCL
This is really the same article as Part 1 - Hello CUDA, but with the focus on OpenCL, so I'll skip most of the narration and just show you the code.
69
Recurrent Neural Network Tutorial for Artists (2017)
Recurrent Neural Network Tutorial for Artists
70
Do neural networks dream of electric sheep?
If you’ve been on the internet today, you’ve probably interacted with a neural network. They’re a type of machine learning algorithm that’s used for everything from language translation to finance modeling. One of their specialties is image recognition. Several companies - including Google, Microsoft, IBM, and…
71
A New Angle on L2 Regularization
Deep neural networks have been shown to be vulnerable to the adversarial example phenomenon: all models tested so far can have their classifications dramatically altered by small image perturbations . The following predictions were for instance made by a state-of-the-art network trained to recognize celebrities :
72
A Full Hardware Guide to Deep Learning
In this guide I analyse hardware from CPU to SSD and their impact on performance for deep learning so that you can choose the hardware that you really need.
73
Foundations of ML and AI: Book recommendations
Here's what I recommend to programmers if they want to understand machine learning. Six fantastic books that you can work with in your spare time to build yo...
74
AI and Blockchain: An Introduction
At the kind invitation of Rob May and the Botchain team, I had the opportunity recently to keynote Brains and Chains, an interesting conference in New York expl
75
Building a Program Synthesizer
Build a program synthesis tool, to generate programs from specifications, in 20 lines of code using Rosette.
76
Imaginary worlds dreamed by BigGAN
These are some of the most amazing generated images I’ve ever seen. Introducing BigGAN, a neural network that generates high-resolution, sometimes photorealistic, imitations of photos it’s seen. None of the images below are real - they’re all generated by BigGAN.
77
AI Nationalism
For the past 9 months I have been presenting versions of this talk to AI researchers, investors, politicians and policy makers. I felt it was time to share these ideas with a wider audience. Thanks to the Ditchley conference on Machine Learning in 2017 for giving me a fantastic platform to get ear
78
Multi-Armed Bandits, Conjugate Models and Bayesian Reinforcement Learning
This is the first of a two-part series about Bayesian bandit algorithms. Check out the second post here. Let’s talk about Bayesianism. It’s developed a reputation (not entirely justified, but not entirely unjustified either) for being too mathematically sophisticated or too computationally intensive to work at scale…
79
Machine Learning Trick of the Day: Instrumental Thinking
· Read in · 1499 words · All posts in series · [dropcap]This[/dropcap] trick is unlike the others we’ve conjured. It will not reveal a clever manipulation of a probability or an integral or a…
80
Troubling Trends in Machine Learning Scholarship
By Zachary C. Lipton* & Jacob Steinhardt* *equal authorship Originally presented at ICML 2018: Machine Learning Debates [arXiv link] Published in Communications of the ACM 1 Introduction Collectively, machine learning (ML) researchers are engaged in
81
Audio Adversarial Examples
Figure from our paper: given any waveform, we can modify it slightly to produce another (similar) waveform that transcribes as any different target phrase.
82
83
The Math Required for Machine Learning
For the past year, I’ve been working on implementing well known model architectures and building web applications, so I have a fair amount…
84
Regarding the Most Recent Amazon Mechanical Turk Data Contamination
Suggested Citation: Bai, H. (2018) Evidence that A Large Amount of Low Quality Responses on MTurk Can Be Detected with Repeated GPS Coordinates. Retrieved from: https://www.maxhuibai.com/blog/evidence-that-responses-from-repeating-gps-are-random
85
How should we evaluate progress in AI?
Improving artificial intelligence research with scientific testing, design practice, and meta-rational choice of methods and criteria
86
Why collision avoidance is harder for an AI-based system
Recent incident where Uber car was involved in an accident where a cyclist got killed has sparked many comments and opinions. Many blame…
87
88
89
Interactive computational neuroscience – part 1, spiking neurons
An interactive introduction to spiking neuron models -- all in the browser. Learn about modeling neurons as dynamical systems and the computational properties of neurons.
90
Intro to Quantum Information Science
Someone recently wrote that my blog is “too high on nerd whining content and too low on actual compsci content to be worth checking too regularly.” While that’s surely one of the…