|
Adam Geitgey |
https://medium.com/@ageitgey |
Know someone who should be on this list? Tweet @scottontech
See also:
Category Archives: Machine Learning
HS: Machine Learning Driven Programming: A New Programming for a New World
If Google were created from scratch today, much of it would be learned, not coded. Around 10% of Google’s 25,000 developers are proficient in ML; it should be 100% — Jeff Dean
Like the weather, everybody complains about programming, but nobody does anything about it. That’s changing and like an unexpected storm the change comes from an unexpected direction: Machine Learning / Deep Learning.
I know, you are tired of hearing about Deep Learning. Who isn’t by now? But programming has been stuck in a rut for a very long time and it’s time we do something about it.
Lots of silly little programming wars continue to be fought that decide nothing. Functions vs objects; this language vs that language; this public cloud vs that public cloud vs this private cloud vs that ‘fill in the blank’; REST vs unrest; this byte level encoding vs some different one; this framework vs that framework; this methodology vs that methodology; bare metal vs containers vs VMs vs unikernels; monoliths vs microservices vs nanoservices; eventually consistent vs transactional; mutable vs immutable; DevOps vs NoOps vs SysOps; scale-up vs scale-out; centralized vs decentralized; single threaded vs massively parallel; sync vs async. And so on ad infinitum.
It’s all pretty much the same shite different day. We are just creating different ways of calling functions that we humans still have to write. The real power would be in getting a machine to write the functions. And that’s what Machine Learning can do, write functions for us. Machine Learning might just might be some different kind of shite for a different day.
Read the full article: Machine Learning Driven Programming on High Scalability
RW: Intel readies chip to rival NVIDIA for machine learning
After abandoning its own GPU for supercomputers, machine learning, and video games in 2009, Intel has returned to the market with a new 72-core Xeon Phi, to compete with NVIDIA’s growing portfolio of GPUs.
The Xeon Phi ‘Knights Landing’ chip, announced at the International Supercomputing Conference in Frankfurt, Germany last week, is Intel’s most powerful and expensive chip to date and is aimed at machine learning and supercomputers, two areas where Nvidia’s GPUs have flourished.
Inside the chip there is 72-cores running at 1.5GHz, alongside 16GB of integrated stacked memory. The chip supports up to 384GB of DDR4 memory, making it immensely scalable for machine learning programs.
Read the full article: Intel readies chip to rival NVIDIA for machine learning on readwrite
Technological Singularity Timeline
| Year | Event |
| 1958 | Earliest recorded use of the term “Singularity” by mathematician Stanislaw Ulam in his tribute to John von Neumann (1903-1957) |
| 1993 | First public use of the term “Singularity” by Venor Vinge in an address to NASA entitled “The Coming Technological Singularity: How to Survive in the Post Human Era” |
| 2005 | Inaugural “Singularity Summit” hosted by the Singularity Institute for Artificial Intelligence (now MIRI) |
| 2006 | The Singularity Is Near: When Humans Transcend Biology by Ray Kurzweil is published. |
| 2009 | Wired for Thought: How the Brain is Shaping the Future of the Internet by Jeffrey Stibel is published. |
| 2012 | MIRI study “How We’re Predicting AI—or Failing To” |
| 2040 | Median value year predicted for Artificial General Intelligence from the 2012 MIRI study. |
References
- https://docs.google.com/file/d/0B-5-JeCa2Z7hbWcxTGsyU09HSTg/edit?pli=1
- http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19940022855.pdf
- http://www-rohan.sdsu.edu/faculty/vinge/misc/singularity.html
- https://en.wikipedia.org/wiki/The_Singularity_Is_Near
- https://en.wikipedia.org/wiki/Wired_for_Thought
- https://intelligence.org/files/PredictingAI.pdf
Building an Image Processing Pipeline Overview
This post is intended to kick off a multi-part series on building an image processing pipeline. Presented here is an outline which will be updated and modified along the way.
My motivation and the goal is to build a process that ingests photos (currently from a mobile device), sort them into multiple categories, and perform an independent task on each photo after sorting.
The secondary goal is to build a toolkit of scripts for each of these tasks that are reusable in the future.
The programming will be in Python.
Pipeline overview:
- Mobile image capture device (iPhone)
- Program to sync photos with a computer (BitTorrent Sync)
- Directory spy to process newly synced images
- Indexing/de-duplication
- Pre-processing
- Processing
- Final result
Progress so far:
Numbers one and two are easily crossed off the list. I have created a directory spy (item #3) in the past, but it needs to be reviewed and modified to work here.
Verge: Google’s art machine just wrote its first song
Today [6/1/16] , Google’s newest machine learning project released its first piece of generated art, a 90-second piano melody created through a trained neural network, provided with just four notes up front. The drums and orchestration weren’t generated by the algorithm, but added for emphasis after the fact.
It’s the first tangible product of Google’s Magenta program, which is designed to put Google’s machine learning systems to work creating art and music systems. The program was first announced at Moogfest last week.
Along with the melody, Google published a new blog post delving into Magenta’s goals, offering the most detail yet on Google’s artistic ambitions. In the long term, Magenta wants to advance the state of machine-generated art and build a community of artists around it — but in the short term, that means building generative systems that plug in to the tools artists are already working with. “We’ll start with audio and video support, tools for working with formats like MIDI, and platforms that help artists connect to machine learning models,” the team wrote in an announcement. “We want to make it super simple to play music along with a Magenta performance model.”
Read the full article: Google’s art machine just wrote its first song on The Verge
Also see:
Getting Started on Somatic.io
Link: http://www.somatic.io/
Pricing:
- Development $199/mo
- Professional $499/mo
- Production $999/mo
- Enterprise (contact for pricing)
Currently 23 Models:
- Deep Videogame Level Generator – Automatically generate blueprints for videogames levels. Using user-specified building blocks, the model will generate game levels that are meant to steer the player through a sequence of designer-controlled steps.
- Deep Classical Composer – Train this model to create original classical music. Modify the training data to influence the output of the model.
- Colornet – Convert black and white images into full color.
- Deep 3D Plants – Automatically generate 3D plants of different kinds, shapes, and sizes.
- inception-prebuilt – This model allows a data scientist to customize which level in an ANN’s (artificial neural network) hierarchy structure to enhance. Lower levels enhance low level features such as lines and basic shapes. Higher levels enhance actual structures suc…
- neuraltalk2-demo – This vision-to-language model analyzes the contents of an image and outputs an English sentence about what it sees. The model was trained using “storyable” events from Flickr and captions that were generated by crowdsourcers on Amazon’s Mechanical…
- char-rnn-ted – his code implements multi-layer Recurrent Neural Network (RNN, LSTM, and GRU) for training character-level language models. In other words, the model takes one text file as input and trains an RNN that learns to predict the next character in a s…
- DCGAN-faces – A neural network that is able to synthesize fake images based off seeing similar images. This particular model generates faces.
- deep3d – As 3D movies and Virtual Reality become more mainstream, the market for 3D content will grow at an exponential rate. However, producing 3D videos is a challenging and expensive process. This model allows for the conversion of 2D images into stereo…
- Caption to Image Generator – This model generates a strip of images that illustrate a caption.
- GRAN Image Generator – This model can be fed a large image dataset of objects such as faces, cars, chairs, etc. Using the data, it will generate a canvas of brand new objects in this class.
- English to Old English Translator – Translate old English (like the works of Shakespeare) into Modern English and vise versa.
- Context Encoder – This algorithm uses context clues to predict and render a missing part of an image.
- Deep Go – Train bots to play the classic board game Go. Interact with your bot through a web browser.
- Autotag Movie Clips – This model automatically indexes the contents of each frame of a video so that you can search and filter video scenes by objects, setting, actors, and more. Save time searching through hundreds of hours of archived footage by instantly queueing th…
- reverse image search – Give an image, return a list of all similar images from your database. We use a deep neural network architect with 8 layers to be able to compare images.
- Neural Doodle – This model allows you to transfer the style of one image on to another image. By creating a simple doodle map, you can specifically tell the AI how you want the style to be transferred to your target image.
- neural-style-demo – Transfer the style from one image and apply it to a target image. Imagine a self portrait in the style of Van Gogh’s [sic] Starry Night.
- GRAN Cat Generator – This model can be fed a large image dataset of objects such as faces, cars, chairs, etc. Using the data, it will generate brand new objects in this class. This particular model generates cute cats.
- neural-storyteller – This model analyzes an image and produces a story about what it sees. The model can be trained with various data sets in order to alter the story’s voice, tone, and word choice. This particular model was trained using romance novels but you could …
- Video Inception – This model allows for deep dreaming in videos. Train the model to “hallucinate” objects in each frame of a video.
- ClearText – This text editor only allows you to use the 1000 most common words in the English language, forcing you to write more clearly.
- image classifier – given any image, tell you if it belongs to a certain class.
FC Layer – Neural Networks
FC Layer in reference to Neural Networks in machine learning stands for:
Fully Connected Layer
Read more about Fully Conneted layers: http://cs231n.github.io/convolutional-networks/#fc
GRU – Neural Networks
GRU in reference to Neural Networks in machine learning stands for:
Gated Recurrent Unit
Read more about GRU networks: http://arxiv.org/pdf/1412.3555v1.pdf
LSTM – Neural Networks
LSTM in reference to Neural Networks in machine learning stands for:
Long Short Term Memory
Read more about LSTM Networks: http://colah.github.io/posts/2015-08-Understanding-LSTMs/