Borrowed from the brain

Your brain is built from cells called neurons. Each one collects signals from its neighbors, and if those signals add up to enough, it fires a signal of its own down a long fiber to the next cell. Nothing about a single neuron is smart — the intelligence comes from billions of them, wired together.

An artificial neuron is a deliberately crude copy of that idea, built entirely from numbers. It keeps the shape of the original — gather signals, combine them, decide whether to pass something on — and throws away everything else.Biological neurondendritescell bodyaxonsynapsesArtificial neuroninputsx1x2x3?weighted sumoutputThe two do the same job. Dendrites become numeric inputs, the cell body’s summing becomes a weighted sum, and the axon’s fire-or-don’t-fire becomes a single output number.

02 — The building block

One neuron, deciding something small

A single artificial neuron does three things: it multiplies each input by a weight (how much that input matters), adds a bias (a built-in nudge), and passes the result through a rule that turns the sum into an output. Here’s one deciding whether to go outside.1Sunny?value = 11Free this evening?value = 10Raining?value = 0weight 0.7weight 0.5weight ?0.9?bias ?0.5sum = 0.7Go outside0.7 × 1 + 0.5 × 1 ? 0.9 × 0 ? 0.5 = 0.7. Because the result is above zero, this neuron’s rule says “fire” — output: go outside. Change the weights and it would weigh the same facts differently.

The weights and the bias are the only things a network actually contains. Everything you’ve heard about a network “knowing” something really means: its weights happen to be set to useful values.

03 — The network

Wiring thousands of them together

One neuron can only weigh a few facts. Real problems — a photo, a sentence, a sound clip — need many neurons arranged in layers: an input layer that takes in raw data, one or more hidden layers that combine it in increasingly abstract ways, and an output layer that gives the final answer.Input layerHidden layerOutput layerCatNot catEvery line is its own weight. This small example has thirty of them; a real image-recognition network has millions. Each hidden neuron combines the whole input layer in its own slightly different way, so together the layer can notice edges, then shapes, then whole features.

04 — Learning

Where do the weights come from?

Nobody sets millions of weights by hand. Instead, the network starts with random ones — it guesses badly on purpose — and then trains: it repeats a short loop, thousands or millions of times, each time getting slightly less wrong.Show a labeled example”this photo is a cat”Network guesses”73% cat, 27% not”Measure the errorhow far off was that?Nudge every weighta little, to lower the errorrepeat, millions of timesThis is the whole trick. The technique for calculating exactly how to nudge each weight is called backpropagation — it traces the error backward through the network, layer by layer, so every single weight gets its own precise correction.

Given enough labeled examples and enough loops, the weights settle into values that make good guesses — not because the network understands cats, but because those particular numbers happen to separate “cat” patterns from everything else.

05 — Putting it together

A trained network, at work

Once training is done, using the network is fast and simple: one forward pass through the layers, no more adjusting. Here’s a (greatly simplified) network reading a handwritten digit.pixels inthe trained networkconfidence per digit0123456789No pixel means “seven” on its own. Each hidden neuron responds to a small pattern of pixels; combined across the layers, that adds up to one output standing far above the rest.

06 — Keeping it honest

What this picture leaves out

Real networks are this idea taken to extremes rather than something fundamentally different. A large language model has many layers and billions of weights instead of thirty; images pass through neurons arranged to scan for local patterns rather than see every pixel at once; and training needs carefully chosen examples, because a network only ever gets as good as what it was shown — bias in the examples becomes bias in the weights.

It’s also worth being plain about what a network is not doing. It has no model of the world, no goals, and no awareness of what it’s looking at — it is a fixed mathematical function, shaped by training, that turns one set of numbers into another. That function can be extraordinarily useful without there being anyone “home” inside it.

A short glossary

NeuronA unit that multiplies its inputs by weights, adds a bias, and passes the sum through an activation rule.WeightA number that scales how much one input matters to a neuron. Training is mostly the process of adjusting these.BiasAn extra number added to the sum, letting a neuron fire more or less easily regardless of its inputs.Activation functionThe rule that turns a neuron’s sum into its output — often something like “pass it through if positive, otherwise send near-zero.”LayerA group of neurons that all take input from the same previous group and hand their output to the same next group.BackpropagationThe method used during training to work out exactly how much to adjust each weight, based on how wrong the final guess was.

https://link.amazon/B0hjl7FoH

“This post contains affiliate links. As an Amazon Associate, I earn from qualifying purchases at no extra cost to you.”

https://youtu.be/bl5FvRUALz0

sudo systemctl enable multi-user.target
sudo systemctl set-default multi-user.target

Mapping between runlevels and systemd targets

If you just want to change runlevel only

sudo systemctl isolate multi-user.target

However, while typing, my thumb inadertently touches the  touch pad and the cursor jumps to some random location.  I am always using an external mouse for navigation  as  I dont like to use t touch pad . So I decided to find out a way to disable the touch pad when I am typing long documents.

Here are the steps for disabling and  enabling  touch pad.

Find out the device id of your touch pad.   Open a terminal and type

$  xinput list


Look for the id of your touch pad. In my case the touch pad is detected as a logitec mouse and the id is 13.  The other mouse shown in the figure is my external mouse. If in doubt, remove the external mouse and re run the command to find out the correct device id.

In order to disable touch pad type the following command.

$ xinput set-prop 13 "Device Enabled" 0

You can eneable the touch pad by changing 0 to 1 in the command above.

Here is what I did.

Removed the existing arduio installation

$ sudo apt-get remove arduino

Download the appropriate arduino software from arduino.cc.  This is available for both 32 bit and 64 bit editions.  Choose the correct version.

The arduino software platfrom is written  in Java. So install java runtime.

$  sudo apt-get install openjdk-7-jre

Unzip the arduino software  and move it to an appropriate location. I moved it to  /opt.

tar -xJf arduino-1.6.4-linux64.tar.xz
sudo mv arduino-1.6.4 /opt

Add appropriate path to .bashrc file.

I added PATH=$PATH;/opt/arduino-1.6.4/bin. Change it according to your arduino location and version.

Open a terminal and execute the new arduino.

Our idea was to put   ubuntu 14.04 and deploy the machines. The machine has an ethernet port and wifi port.  Installation of Ubuntu 15.10 was a smooth affair.  Unfortunately, the broadcom wifi interface did  not work out of the box. Here is what I did to fix the issue.

1. Find out which wifi chip is used.

$lspci

The wifi chip is  Broadcom Corporation BCM43142. You have to download and compile the kernel module for this chip.

2) Make sure that the machine is connected to internet and the apt sources are properly set up

3) Update  the machine

$ sudo apt-get update

4) Finally instal the  bcmwl driver.

$ sudo apt-get install linux-headers-`uname -r` dkms build-essential bcmwl-kernel-source

This will downlad necessary driver  your wifi will be up and running. The above line installs dynamic kernel module support  and compiles the broadcom wireless driver.

For the ham radio home brewer, such a chip which can be programmed over many ham bands offer  numerous possibilities.   The most obvious choice would be  as a VFO. Currently, the analog devices DDS chips are ruling the market. Compared to them the Silicon Labs chips are cheaper and less cumbersome to handle. I am yet to see the noise performance of these two chips compared.

In fact, there are a couple of commercial Si5351A breakout boards   targeted at the  home brewer. They cost around 10$. But the chip itself cost only 1.33$  on the Silicon Labs website and you need only a few  additional components to complete the board. The total cost need not be more than 3 $.  Along with a cheap Arduino clone, Si5351A gives you lot of flexibility. An added advantages is that the SI5351A  has three independent programmable clock generators. You can combine  the  VFO and   BFO  into a single chip.

My idea was to build a single side PCB so that any one can build one via toner transfer. Last week,  I had a long  chat  with my friend and colleague Amal Dev  about designing  a simple board.  The board routing was done by Amal Dev on eagle,  We have  designed an experimental single sided board and it may not be conforming to RF design principles always.

For the interested home brewer a  toner transferable image is  here.    You can  buy the  SI5351 from RS components  or from Mouser.  The board is to be powered from a 5v source.  The SI5351A needs 3.3v for its operation. The LM1117 is used for generating this.  Note that the LM1117  is SMD and is soldered on the copper side.  The two MOSFETs ( BS170)   are used to convert the I2C levels from 5v to 3.3v.  Check the pin outs before you solder.  The board can be directly powered from an arduino.

For programming the board you can try the example code from Hansummers.  Alternately you can try the excellent si5351 arduino library.  Tom Hall AK2B has some  sample sketches using this library. Please note that I2C address of the chip may be different from those found in the above code. Please use an I2C scanner  to find out your chips address and then modify your code accordingly.

The pin out of the SI5351A is shown below.

A special thanks to Amogh Desai who tested the first version  and reported couple of errors.  Further comments and error reports are welcome.

I used an arduino Uno along with a rotary encoder for  controlling the module. The DDS module pin out is shown below.

The connections to arduino and rotary encoder is  shown below.

You can download the code hosted on github from  here. 

The code is for 40m VFO    with IF of 9Mhz.  You can easily modify it for other  bands and IFs. Additional functions can be programmed easily.

The DDS modules  are available on line form from many sources. The rotary encoder can be obtained from element14.