RF-Sleep learns to predict sleep stages from radio measurements without any attached sensors on subjects. We introduce a new predictive model that combines convolutional and recurrent neural networks to extract sleep-specific subject-invariant features from RF signals and capture the temporal progression of sleep.
DeepMind's AI Learns Imagination-Based Planning
Károly Zsolnai-Fehér's new video | Two Minute Papers #178
AI Learns Semantic Style Transfer
A new Two Minute Papers #177
Introducing Movidius Neural Compute Stick
Movidius Neural Compute Stick is a modular artificial intelligence (AI) accelerator in a standard USB 3.0 stick form factor. Designed for product developers, researchers and makers, Movidius Neural Compute Stick aims to reduce barriers to developing, tuning and deploying deep learning applications at the edge by delivering dedicated high-performance deep neural network processing
Refocusing Videos With Neural Networks | Two Minute Papers #173
The paper "Light Field Video Capture Using a Learning-Based Hybrid Imaging System" and its implementation is available here
The story of Replika, the AI app that becomes you
Replika is a chatbot that creates a digital representation of you. It's strange and fascinating -- but the story behind it is even better.
Convolutional Neural Networks - The Math of Intelligence
Convolutional Networks allow us to classify images, generate them, and can even be applied to other types of data. We're going to build one in numpy that can classify and type of alphanumeric character and it will run in a Flask web app.
AI Learns Visual Common Sense With New Dataset | Two Minute Papers #169
Károly Zsolnai-Fehér's new video
DeepMind's AI Learns Superhuman Relational Reasoning | Two Minute Papers #168
Károly Zsolnai-Fehér's new video The paper "A simple neural network module for relational reasoning" is available here: https://arxiv.org/abs/1706.01427
Published on Jul 6, 2017
The Brain: A neural network built entirely in Quartz Composer
Mike Matas made this demonstration using Apple's QC environment.
Feb 16, 2016
Neural Network 3D Simulation
Denis Dmitriev created this 3D animation for cybercontrols.org
Nov 15, 2016
Tutorial On Programming An Evolving Neural Network In C# w/ Unity3D
C# Tutorial by Derek Banas. Tutorial on programming an evolving neural network (MLP).
Evolving Neural Networks Of Joint Segmented Line Creatures
JSLC or Joint Segmented Line Creatures evolve to run to the right as fast as possible. Creatures that move to the right the fastest will have the higher fitness. Each JSLC can be created by the user by drawing and can consist of a number of joints. Each joint has 3 inputs to the brain. Previous joint speed, max angle flag, and current joint angle. the output of the neural network controls the speed of the joints.
Intro - The Math of Intelligence
Welcome to The Math of Intelligence! In this 3 month course, we'll cover the most fundamental math concepts in Machine Learning. In this first lesson, we'll go over a very popular optimization technique called gradient descent to help us predict how many calories a cyclist would burn given just their distance traveled.
The Power of the Blockchain
You won't find this data structure in your computer science textbooks yet, but it will soon underpin the way the entire Internet works. Let's talk about what the blockchain is and how it can be used to improve our AI.
Jun 9 2017
Phase-Functioned Neural Networks for Character Control
We present a real-time character control mechanism using a novel neural network architecture called a Phase-Functioned Neural Network. In this network structure, the weights are computed via a cyclic function which uses the phase as an input.
AI Learns To Create User Interfaces (pix2code) | Two Minute Papers
The paper "pix2code: Generating Code from a Graphical User Interface Screenshot" is available
How to Deploy a Tensorflow Model to Production
Siraj Raval has another new high energy video.
Case Study: TensorFlow in Medicine - Retinal Imaging (TensorFlow Dev Summit 2017)
Diabetic retinopathy is the fastest growing cause of blindness. Learn from Lily Peng how TensorFlow was trained to analyze retinal fundus images to diagnose this condition. She describes the project steps: from acquiring a dataset, training a deep network, and evaluating of the results.
What are Future Jobs and Work in Era of Artificial Intelligence(AI)?
From Accenture and Skinome
AI Learns to Synthesize Pictures of Animals | Two Minute Papers
Károly Zsolnai-Fehér has another marvelous production.
The Rise of Artificial Intelligence through Deep Learning | Yoshua Bengio | TEDxMontreal
A revolution in AI is occurring thanks to progress in deep learning. How far are we towards the goal of achieving human-level AI? What are some of the main challenges ahead?
TensorFlow Tutorial #03-B Layers API
Magnus Erik Hvass Pedersen Has produced a new video.
How to use the Layers API to simplify the implementation of a Convolutional Neural Network in TensorFlow. This is used for recognizing handwritten digits from the MNIST data-set.
GTC 2017: Nvidia gpu technology conference Tesla V100 Volta
This is a recording of the GTC 2017 livestream.
(From TED 2013) Andrew McAfee What will future jobs look like?
Note: Although this is from 2013, the talk is even more applicable now in 2017. Economist Andrew McAfee suggests that, yes, probably, droids will take our jobs -- or at least the kinds of jobs we know now. In this far-seeing talk, he thinks through what future jobs might look like, and how to educate coming generations to hold them.
Differentiable Neural Computer Siraj Raval Live May 10 2017
The Differentiable Neural Computer is an awesome model that DeepMind recently released. It's a memory augmented network that can perform meta-learning (learning to learn). We'll go over it's architecture details and implement it ourselves in Tensorflow.
How to Learn from Little Data - Intro to Deep Learning #17v
Siraj Raval has a new video. One-shot learning! In this last weekly video of the course, i'll explain how memory augmented neural networks can help achieve one-shot classification for a small labeled image dataset. We'll also go over the architecture of it's inspiration (the neural turing machine).
TensorFlow Tutorial #16 Reinforcement Learning
Magnus Erik Hvass Pedersen has a new video.
How to implement Reinforcement Learning in TensorFlow. This is a version of Q-Learning that is somewhat different from the original DQN implementation by Google DeepMind. Demonstrated on the Atari game Breakout. April 24,2017
The incredible inventions of intuitive AI | Maurice Conti
The title isn't overstated. This video is seminal.
From Cold Fusion TV - A.I. is Progressing Faster Than You Think!
Cold Fusion TV has a new video which follows-up on a piece done in 2016.
How to Generate Video - Intro to Deep Learning #15
Siraj Raval Hit another home run with this video on GANs.
Generative Adversarial Networks. It's time. We're going to use a Deep Convolutional GAN to generate images of the alien language from the movie arrival that we can then stitch together to animate into video. I'll go over the architecture of a GAN and then we'll implement one ourselves! April 20,2017
Algorithmic bias,Joy Buolamwini
Algorithmic bias like human bias can result in exclusionary experiences and discriminatory practices. Find out how below.
How to Generate Images - Intro to Deep Learning #14
Siraj's latest video is his words: We're going to build a variational autoencoder capable of generating novel images after being trained on a collection of images.
TensorFlow Dev Summit 2017
The 2017 TesnorFlow Summit was held in February. The Summit's 18 talks are archived and available on YouTube.
Shape2vec: Understanding 3D Shapes With AI
New Two Minute Papers: Shape2vec: Understanding 3D Shapes With AI
Stable Neural Style Transfer | Two Minute Papers
Breaking DeepMind's Game AI System | Two Minute Papers
Feb 19 2017 A new vide titled AI Builds 3D Models From Images With a Twist has been posted by Two Minute Papers aka Károly Zsolnai-Fehér'
Feb 17,2017 Siraj @sirajraval has also published a video titled How to Make an Image Classifier - Intro to Deep Learning #6
We want to mention four sources that will be mentioned often.
9 Cool Deep Learning Applications | Two Minute Papers
9 Cool Deep Learning Applications | Two Minute Papersis another one of 130 videos most of which touch on Neural Networks. Here is a full list of Two Minute Videos. we've sent my robot out and have downloaded all 130 video and the SRT transcript files. One of our goals is to cross index all of the good Neural Net content.
Siraj Raval is another source of Neural Network video series. 'High energy' is an understatement. He presents topics with examples which have open source code available for study and experimentation. Big pluses when you want to elevate your understanding.
Next in the roster of good Neural Network video series Deep Learning TV by Jagannath Rajagopal.
Neural Net Weekly is on the air. For just over one month, we have been spending evenings looking with amazement at what is happening. The potential for impact, disruption and consequence outstrips the web's beginning. So when we found that the domain name was available, we wanted to try document our learning travels. Neural Net Weekly will attempt to capture the week in neural networks. We will announce the best curated content for learning and experimenting with artificial networks. Keeping up with just this segment of neural networks will be a big task. We will use collective curation to achieve these goals. The video Understand collective curation in under 90 seconds helps to describe the modus operandi.