Smarter Phones In Your Hacks With TensorFlow Lite

November 23, 2017 by Steven Dufresne 11 Comments

One way to run a compute-intensive neural network on a hack has been to put a decent laptop onboard. But wouldn’t it be great if you could go smaller and cheaper by using a phone instead? If your neural network was written using Google’s TensorFlow framework then you’ve had the option of using TensorFlow Mobile, but it doesn’t use any of the phone’s accelerated hardware, and so it might not have been fast enough.

Google has just released a new solution, the developer preview of TensofFlow Lite for iOS and Android and announced plans to support Raspberry Pi 3. On Android, the bottom layer is the Android Neural Networks API which makes use of the phone’s DSP, GPU and/or any other specialized hardware to speed up computations. Failing that, it falls back on the CPU.

Currently, fewer operators are supported than with TensforFlor Mobile, but more will be added. (Most of what you do in TensorFlow is done through operators, or ops. See our introduction to TensorFlow article if you need a refresher on how TensorFlow works.) The Lite version is intended to be the successor to Mobile. As with Mobile, you’d only do inference on the device. That means you’d train the neural network elsewhere, perhaps on a GPU-rich desktop or on a GPU farm over the network, and then make use of the trained network on your device.

What are we envisioning here? How about replacing the MacBook Pro on the self-driving RC cars we’ve talked about with a much smaller, lighter and less power-hungry Android phone? The phone even has a camera and an IMU built-in, though you’d need a way to talk to the rest of the hardware in lieu of GPIO.

You can try out TensorFlow Lite fairly easily by going to their GitHub and downloading a pre-built binary. We suspect that’s what was done to produce the first of the demonstration videos below.

Continue reading “Smarter Phones In Your Hacks With TensorFlow Lite” →

Prototyping, Making A Board For, And Coding An ARM Neural Net Robot

November 21, 2017 by Steven Dufresne 1 Comment

[Sean Hodgins]’s calls his three-part video series an Arduino Neural Network Robot but we’d rather call it an enjoyable series on prototyping, designing a board with surface mount parts, assembling it, and oh yeah, putting a neural network on it, all the while offering plenty of useful tips.

In part one, prototype and design, he starts us out with a prototype using a breadboard. The final robot isn’t on an Arduino, but instead is on a custom-made board built around an ARM Cortex-M0+ processor. However, for the prototype, he uses a SparkFun SAM21 Arduino-sized board, a Pololu DRV8835 dual motor driver board, four photoresistors, two motors, a battery, and sundry other parts.

Once he’s proven the prototype works, he creates the schematic for his custom board. Rather than start from scratch, he goes to SparkFun’s and Pololu’s websites for the schematics of their boards and incorporates those into his design. From there he talks about how and why he starts out in a CAD program, then moves on to KiCad where he talks about his approach to layout.

Part two is about soldering and assembly, from how he sorts the components while still in their shipping packages, to tips on doing the reflow in a toaster oven, and fixing bridges and parts that aren’t on all their pads, including the microprocessor.

In Part three he writes the code. The robot’s objective is simple, run away from the light. He first tests the photoresistors without the motors and then writes a procedural program to make the robot afraid of the light, this time with the motors. Finally, he writes the neural network code, but not before first giving a decent explanation of how the neural network works. He admits that you don’t really need a neural network to make the robot run away from the light. But from his comparisons of the robot running using the procedural approach and then the neural network approach, we think the neural network one responds better to what would be the in-between cases for the procedural approach. Admittedly, it could be that a better procedural version could be written, but having the neural network saved him the trouble and he’s shown us a lot that can be reused from the effort.

In case you want to replicate this, [Sean]’s provided a GitHub page with BOM, code and so on. Check out all three parts below, or watch just the parts that interest you.

Continue reading “Prototyping, Making A Board For, And Coding An ARM Neural Net Robot” →

Artificial Intelligence At The Top Of A Professional Sport

October 20, 2017 by Lauren Faris 20 Comments

The lights dim and the music swells as an elite competitor in a silk robe passes through a cheering crowd to take the ring. It’s a blueprint familiar to boxing, only this pugilist won’t be throwing punches.

OpenAI created an AI bot that has beaten the best players in the world at this year’s International championship. The International is an esports competition held annually for Dota 2, one of the most competitive multiplayer online battle arena (MOBA) games.

Each match of the International consists of two 5-player teams competing against each other for 35-45 minutes. In layman’s terms, it is an online version of capture the flag. While the premise may sound simple, it is actually one of the most complicated and detailed competitive games out there. The top teams are required to practice together daily, but this level of play is nothing new to them. To reach a professional level, individual players would practice obscenely late, go to sleep, and then repeat the process. For years. So how long did the AI bot have to prepare for this competition compared to these seasoned pros? A couple of months.

Continue reading “Artificial Intelligence At The Top Of A Professional Sport” →

Nematoduino: A Roundworm Neural Model On An Arduino

October 13, 2017 by Dan Maloney 13 Comments

When it comes to building a neural network to simulate complex behavior, Arduino isn’t exactly the first platform that springs to mind. But when your goal is to model the behavior of an organism with only a handful of neurons, the constraints presented by an Arduino start to make sense.

It may be the most important non-segmented worm you’ve never heard of, but Caenorhabditis elegans, mercifully abbreviated C. elegans, is an important model organism for neurobiology, having had its entire nervous system mapped in 2012. [Nathan Griffith] used this “connectome” to simulate a subset of the diminutive nematode’s behaviors, specifically movements toward attractants and away from obstacles. Riding atop a small robot chassis, the Arduino sends signals to the motors when the model determines it’s time to fire the virtual worm’s muscles. An ultrasonic sensor stands in for the “nose touch” neurons of the real worm, and when the model is not busy avoiding a touch, it’s actively seeking something to eat using the “chemotaxis” behavior. The model is up on GitHub and [Nathan] hopes it provides an approachable platform for would-be neuroroboticists.

This isn’t the first time someone has modeled the nematode’s connectome in silico, but kudos to [Nathan] for accomplishing it within the constraints an Arduino presents.

Continue reading “Nematoduino: A Roundworm Neural Model On An Arduino” →

Neural Network Gimbal Is Always Watching

October 10, 2017 by Tom Nardi 16 Comments

[Gabriel] picked up a GoPro to document his adventures on the slopes and trails of Montreal, but quickly found he was better in front of the camera than behind it. Turns out he’s even better seated behind his workbench, as the completely custom auto-tracking gimbal he came up with is nothing short of a work of art.

There’s quite a bit going on here, and as you might expect, it took several iterations before [Gabriel] got all the parts working together. The rather GLaDOS-looking body of the gimbal is entirely 3D printed, and holds the motors, camera, and a collection of ultrasonic receivers. The Nvidia Jetson TX1 that does the computational heavy lifting is riding shotgun in its own swanky looking 3D printed enclosure, but [Gabriel] notes a future revision of the hardware should be able to reunite them.

In the current version of the system, the target wears an ultrasonic emitter that is picked up by the sensors in the gimbal. The rough position information provided by the ultrasonics is then refined by the neural network running on the Jetson TX1 so that the camera is always focused on the moving object. Right now the Jetson TX1 gets the video feed from the camera over WiFi, and commands the gimbal hardware over Bluetooth. Once the Jetson is inside the gimbal however, some of the hardware can likely be directly connected, and [Gabriel] says the ultrasonics may be deleted from the design completely in favor of tracking purely in software. He plans on open sourcing the project, but says he’s got some internal house keeping to do before he takes the wraps off it.

From bare bones to cushy luxury, scratch-built camera gimbals have become something of a right of passage for the photography hacker. But with this project, it looks like the bar got set just a bit higher.

Continue reading “Neural Network Gimbal Is Always Watching” →

Decoding Enigma Using A Neural Network

August 11, 2017 by Steven Dufresne 21 Comments

[Sam Greydanus] created a neural network that can encode and decode messages just as Enigma did. For those who don’t know, the Enigma machine was most famously used by the Germans during World War II to encrypt and decrypt messages. Give the neural network some encrypted text, called the ciphertext, along with the three-letter key that was used to encrypt the text, and the network predicts what the original text, or plaintext, was with around 96-97% accuracy.

The type of neural network he used was a Long Short Term Memory (LSTM ) network, a type of Recurrent Neural Network (RNN) that we talked about in our article covering many of the different types of neural networks developed over the years. RNNs are Turing-complete, meaning they can approximate any function. [Sam] noticed the irony in this, namely that Alan Turing both came up with the concept of Turing-completeness as well as played a big part in breaking the Enigma used in World War II.

How did [Sam] do it?

Continue reading “Decoding Enigma Using A Neural Network” →

Neural Nets In The Browser: Why Not?

August 4, 2017 by Al Williams 16 Comments

We keep seeing more and more Tensor Flow neural network projects. We also keep seeing more and more things running in the browser. You don’t have to be Mr. Spock to see this one coming. TensorFire runs neural networks in the browser and claims that WebGL allows it to run as quickly as it would on the user’s desktop computer. The main page is a demo that stylizes images, but if you want more detail you’ll probably want to visit the project page, instead. You might also enjoy the video from one of the creators, [Kevin Kwok], below.

TensorFire has two parts: a low-level language for writing massively parallel WebGL shaders that operate on 4D tensors and a high-level library for importing models from Keras or TensorFlow. The authors claim it will work on any GPU and–in some cases–will be actually faster than running native TensorFlow.

Continue reading “Neural Nets In The Browser: Why Not?” →