Industry automation
Test & measurement
Solar energy
Training , Trade fairs, Evens
Interesting video
AR236.B & AR232.B
APAR AR236.B & AR232.B radio loggers
JZ-500 and JZ-500-C black
JZ-500 and JZ-500-C black cables manufac
HF41F subminiature power relays from HON
GOLDTOOL toolsets in cases
Intel® Neural Compute Stick 2
RS Components announces availability of
MC4 solar connector sets
Multifunctional WIHA LED Torch
G2RV-SR interface relays from OMRON
LUP-20-LED-SMD desktop magnifier
Power Integrations Unveils High-Efficien

Artificial Intelligence on the Edge of Your Network

Artificial Intelligence on the Edge of Your Network

This is not the topic of the future; with AAEON hardware you can handle the deployment of artificial intelligence now. Read the article to find out how it all works and what you need for that.

Why now?

In the present state of technology we can train deep neural network (DNN) for a specific tasks like object and human face detection and recognition, speech recognition, language translation, games (chess, go, etc.), autonomous vehicle driving, sensor status monitoring and predictive machine maintenance decisions, evaluation of X-ray images in healthcare etc. For such specialized tasks, a DNN can reach or even exceeds human capabilities.

Why Use Artificial Intelligence on the Edge

For example, the modern building contains a lot of sensors, HVAC devices, elevators, security cameras, etc. connected to the internal network. For the security, latency or robustness reasons, it is more appropriate to run artificial intelligence tasks locally, on the edge of a local network, and send only anonymized data that are necessary for taking global decisions to the cloud.

Hardware on the Edge

To deploy DNN on the edge we need a device with enough computing power and simultaneously low power consumption. The current state of technology offers a combination of low power CPU and VPU accelerator (x86 CPU SBC+ Intel Myriad X VPU) or CPU + GPU accelerator (ARM CPU + Nvidia GPU).

The easiest way to start DNN is using UP Squared AI Vision X Developer Kit version B. It is based on UP Square SBC with Intel Atom®X7-E3950 processor with 8GB RAM, 64 GB eMMC, AI Core X module with Myriad X MA2485 VPU and USB camera with resolution 1920 x 1080 and manual focus. The kit is preinstalled with Ubuntu 16.04 (kernel 4.15) and OpenVINO toolkit 2018 R5.

Toolkit contains precompiled demo applications in /home/upsquared/build/intel64/Release and pre-trained models in /opt/intel/computer_vision_sdk/deployment_tools/intel_models. To see help for any demo application, run it in terminal with –h option. It is recommended to initialize OpenVINO environment before running demo application by command source /opt/intel/computer_vision_sdk/bin/setupvars.sh.

Besides UP Squared AI Vision X Developer Kit, Aaeon also offers:

1. Myriad X MA2485 VPU based modules: AI Core X (mPCIe full-size, 1x Myriad X), AI Core XM 2280 (M.2 2280 B+M key, 2x Myriad X), AI Core XP4/ XP8 (PCIE [x4] card, 4 or 8x Myriad X).
2. BOXER-8000 series based on Nvidia Jetson TX2 module.
3. BOXER-8320AI with Core i3-6100U or Celeron 3955U processor and two AI Core X modules.
4. Boxer-6841M series with the motherboard for Intel 6th / 7th generation of Core-I or Xeon processor for socket LGA1151 and 1x PCIe [x16] or 2x PCIe [x8] slots for GPU with power consumption max. 250W.

Hardware for Learning To train DNN we need high computing power. For example, on the ImageNet competition in 2012, the winning team used the convolutional neural network AlexNet. 1.4 ExaFLOP=1,4e6 TFLOP operations were required for learning. It took 5 to 6 days on two Nvidia GTX 580 GPUs, where each has the 1.5 TFLOP computing performance.

The table below summarizes the theoretical peak performance of the hardware.

For comparison, the high-end Intel Xeon Platinum 8180 processor
? has 28 cores with 2 AVX-512 & FMA units per core
? AVX-512 turbo frequency 2.3GHz if all cores are active
? costs USD 10 000

It offers theoretical peak performance: # of cores * frequency in GHz * AVX-512 DP FLOPS/Hz * # of AVX-512 units * 2 = 2060.8 GFLOPS in double precision (DP) › 4.1216 TFLOPS in single (32-bit).

As you can see from the table above, GPU provides far more performance for neural networks learning. It is necessary to note that the number of operations per second is not the only parameter that affects learning performance. Factors as RAM size, the data transfer rate between CPU and RAM, GPU and GPU RAM or between GPU also influence the learning speed.


The OpenVINO (open visual inference and neural network) toolkit is free software that allows quick deployment of applications and solutions that emulate human vision.

The OpenVINO toolkit:
? Uses CNN (convolution neural network)
? Can split computation between Intel CPU, integrated GPU, Intel FPGA, Intel Movidius Neural Compute Stick and vision accelerators with Intel Movidius Myriad VPUs
? Provides an optimized interface to OpenCV, OpenCL, and OpenVX
? Supports Caffe, TensorFlow, MXNet, ONNX, Kaldi frameworks

https://software.intel.com/en-us/openvino-toolkit, https://docs.openvinotoolkit.org

TensorFlow is an open-source library for numerical computation and machine learning. It provides a convenient front-end API for building applications in Python programming language. However, the TensorFlow-generated application itself is converted to optimized C ++ code that can run on a variety of platforms such as CPUs, GPUs, local machine, a cluster in the cloud, embedded devices at the edge and the like.

Other Useful Software
Jupyter Lab / Notebook

How it Works

Simplified Neuron Model
Simple neuron model – the perceptron was first time described by Warren McCulloch and Walter Pitts in 1943 and it is still the standard of reference in the field of neural networks.

Activation function f () adds nonlinearity to perceptron. Without non-linear activation function in the neural network (NN) of perceptrons, no matter how many layers it had, it would behave just like a single-layer perceptron, because summing these layers would give you just another linear function. The most often used activation function is rectified linear unit – ReLU.

y = f(x) = max (0, x), for x < = 0, y = 0, for x ? 0, y=x

Inference (Forward pass)

The figure above shows a deep neural network (DNN) because it contains multiple layers between input and output layers. As you can notice, DNN requires matrix multiplications and additions. Special hardware optimized for this task, such as GPU (graphics processing unit) and VPU (vision processing unit), is much faster than general-purpose CPU (central processing unit, processor) and has lower power consumption.

Learning (Backward pass)

Let’s say we want to teach DNN to recognize orange, banana, apple, and raspberry (object classes) in the image.

1. We prepare a lot of images of fruits above and divide it into the training set and validation set. The training set contains images and correct, required outputs for these images. DNN will have 4 outputs. The first output provides a score (probability) that fruit in the picture is orange, the second provides the same for a banana, etc.

2. We set initial values for all weights w_i and biases b_i. Random values are typically used.

3. We pass the first image through DNN. The network provides scores (probability) on each output. Let’s say that the first image depicts orange. Network outputs can be y= (orange, banana, apple, raspberry) = (0.5, 0.1, 0.3, 0.1). Network “says” that input is orange with a probability of 0.5.

4. We define a loss (error) function that quantifies the agreement between the predicted scores and the correct scores for each class. Function E = 0.5*sum (e_j) ^2, where e_j = y_j - y_real_j and j is number of images in training set is often used. E_1_orange = 0.5*(0.5-1)^2=0.125, E_1_banana =.0.5*(0.1-0)^2 = 0.005 E_1_apple = 0.5*(0.3-0)^2 = 0.045, E_1_raspberry = 0.5*(0.1-0)^2 = 0.005 E_1 = (0.125, 0.005, 0.045, 0.005)

5. We pass all the remaining images from the training set through DNN and calculate E (E_orange, E_banana, E_apple, E_raspberry) loss function value for the whole training set.

6. To modify all weights w_i and biases b_i for the next training pass (epoch), we need to know the influence of each parameter w_i and b_i to loss function for each class. If the increase of parameter value causes the increase of loss function value, we need to decrease this parameter and vice versa. But how to calculate required parameters increase or decrease?

Let’s try a simple example.

We have three points with coordinates (x, y): (1, 3), (2.5, 2), (3.5, 5). We want to find a line y = w.x + b for which loss function E = 0.5*sum (e_j) ^2 , where e_j = y_j – y_real_j, j=1, 2, 3 is minimal. To make task as simple as possible, let’s say that w=1.2 and we need to find optimal value for b only. We pick the initial value for b=0.

Let’s calculate loss function: E = 0.5*sum (e_j) ^2 = 0.5*(e_1^2 + e_2^2 + e_3^2), e_1=1.2*1 + b -3, e_2 = 1.2*2.5 + b – 2, e_3 = 1.2*3.5 + b – 5.

Loss function E is simple, we can find a minimum of E by solving equation: ?E/?b = 0. It is a compound function, to calculate ?E/?b we apply the chain rule.

?E/?b=0.5*((?E/?e_1)*(?e_1/?b) + (?E/?e_2)*(?e_2/?b) + (?E/?e_3)*(?e_3/?b)) = 0.5*(2*e_1*1 + 2*e_2*1 + 2*e_3*1) = (1.2*1 + b – 3) + (1.2*2.5 + b – 2) + (1.2*3.5 + b – 5) = 0 => b = 0.53333

In practice, where number of parameters w_i and b_i can reach a million or more, it is not practical to solve equation ?E/?b_i = 0 and ?E/?b_i = 0 directly, iterative algorithm is used instead of that.

We started with b = 0, next value will be b_1 = b_0 – ?*?E/?b, where ? is learning rate (hyper-parameter) and -?*?E/?b is step size. We stop learning when step size reaches the defined threshold, in practice 0.001 or smaller.

For ? = 0.3, b_1 = 0.48, b_2 = 0.528, b_3 = 0.5328 and b_4 = 0.53328 and b_5 = 0.5533328. After 5 iterations the step size dropped to 4.8e-5 and we stop learning here. The value of b obtained by this algorithm is practically the same as the value obtained by solving the equation ?E/?b=0.

This method is called gradient descend.

Learning rate is an important hyper-parameter. If it is too small, it takes a lot of steps to find a minimum loss function; if it is high, the algorithm can fail. In practice, improved variants of an algorithm such as  Adam are used.

7. We repeat steps 5 and 6 until loss function value decreases to the required value.

8. We pass the validation set through trained DNN and evaluate accuracy.

In the present time, DNN learning is a highly experimental work. Many DNN architectures are known, each of them is well suited for a particular range of tasks. Every DNN architecture has its own set of hyper-parameters that influence the behavior of DNN. Arm yourself with patience and the result will come soon.

For further information about AAEON products, please do not hesitate to contact us at aaeon@soselectronic.com

2019102102 / 24.10.2019 / Various / SOS electronic s.r.o. /

Apacer Brings New Technologies to SSD
Apacer Brings New Technologies to SSD
New Apacer’s SSDs improve already high data integrity, longevity, security, and survivability. Let’s have a look at how they did it.
Get more space on PCB with TEL 10WI
TRACO POWER 10 Watt DC/DC converter for space-limited applications.
Get more space on PCB with TEL 10WI
Artificial Intelligence on the Edge of Your Network
Artificial Intelligence on the Edge of Your Network
This is not the topic of the future; with AAEON hardware you can handle the deployment of artificial intelligence now. Read the article to find out how it all works and what you need for that.
More than 100 new active tips for each soldering joint
The largest Weller product range expansion includes three soldering handles and more than 100 tips with active heating, suitable for microscopic work to solder coolers and LEDs on PCB with aluminum base.
More than 100 new active tips for each soldering joint
Oscilloscopes that will surprise you with superior features already in the basic class
Oscilloscopes that will surprise you with superior features already in the basic class
The vast array of useful features and the concept of future growth; this is the Rohde&Schwarz RTB 200x series of oscilloscopes.
Replace the good relay with an even better one
Finder has found a new, versatile and cheaper solution. The well-known relays of 44.52 and 44.62 series will be replaced by the new one - 40.62.
Replace the good relay with an even better one
WINSTA – Perfect in Details
WINSTA – Perfect in Details
Winsta is the pluggable connection system from WAGO. The system has been particularly developed for building installations and offers connections for virtually any building requirement from power supply to bus lines.
Sensirion – The Sensor Company
Overview of the basic information and our tips for CO2, VOC, humidity, temperature, liquid and gas flow sensors. If you are looking for a solution, Sensirion will not disappoint you.
Sensirion – The Sensor Company
SOS electronic webinar: How to set up PICO-PI-IMX8M-MINI-2G-DEV for your first Qt application?
SOS electronic webinar: How to set up PICO-PI-IMX8M-MINI-2G-DEV for your first Qt application?

SOS electronic - Term: 2019.09.11 10:00 - 11:00
Quartz Crystals and Quartz Crystal Oscillators Ansen
Without crystal oscillators, there would be no wireless communication, GPS receivers or Ethernet. Therefore, we are expanding the portfolio of suppliers with the renowned ANSEN manufacturer. Although new technologies such as MEMS oscillators exist, crystal oscillators remain the dominant source of accurate and stable frequency.
Quartz Crystals and Quartz Crystal Oscillators Ansen
Can penguin live in desert?
Can penguin live in desert?
Not likely. Similarly, neither an ordinary networking cable wouldn´t survive in a harsh industrial environment without harm. You can prevent undesired secondary costs and ensure the safety for your network by installing a robust cable.
W6100 is ready for IPv6
With the enormous growth of IoT devices connected to the networking, it became evident that far more addresses would be needed to connect devices than the IPv4 address space had available. Wiznet has a solution.
W6100 is ready for IPv6 
Webinar recording: LPWA technology
Webinar recording: LPWA technology
We are excited about your interest in our “SOS Webinar with Quectel: LPWA (NB-IoT + LTE Cat M1)”. Such high participation in technically oriented webinars is not always a matter of course.
We are expanding the portfolio of IoT products with the world brand USR IoT
Our new supplier, USR IoT, is an expert in industrial communication IoT products with a focus on serial data networking and M2M applications. ISO9001 quality management with more than 20 years of experience in business guarantees its expertise.
We are expanding the portfolio of IoT products with the world brand USR IoT
Building ESP32 products fast
Building ESP32 products fast
ESP-Jumpstart is Espressif’s new reference guide which aims to help developers convert their ideas into actual products as quickly as possible.
SVM30 Module - a complete solution for HVAC applications
The SVM30 measures concentration of VOC, relative humidity and temperature and calculates CO2eq from measured concentration of H2. It uses SGP30 and SHTC1 sensors. At first glance, however, it seems to be quite big, check it out with us.
SVM30 Module - a complete solution for HVAC applications
RTC-1010: The Optimal Combination of Performance, Durability and Interfaces
RTC-1010: The Optimal Combination of Performance, Durability and Interfaces
The rugged 10,1” tablet is a perfect tool for industrial field service and maintenance or mobile HMI. It offers industry standard interfaces and connectors, hot-swappable battery and optional barcode reader, as well as the fact that it has been designed to allow customization.
THL15WI Series operates at 15W power in smaller case
The higher energy efficiency, maintaining the quality and a lower cost. These parameters were the cornerstones for the new THL15WI family from TRACO in a 1 x 1 x 0.4″ metal case.
THL15WI Series operates at 15W power in smaller case
Company of the week

Keysight Technologies
Address Book

Keysight Technologies

XP Power Limited

Yamaichi Electronics USA Inc.

Fischer Connectors SA


TE Connectivity

akYtec GmbH

SOS electronic s.r.o.


Taiwan Semiconductor

ROHM Semiconductor

ČEZ a.s.

Pickering Interfaces Ltd


Mouser Electronics

Foremost Electronics

RS Components GmbH



Willow Technologies


Visual Communications Company, LLC

ACS Motion Control


Transfer Multisort Elektronik Sp. z o.o.

Hradil Spezialkabel GmbH

Sonderhoff Holding GmbH

Radio Frequency Systems

Apacer Technology Inc.

Pico Technology
productronica 2019, Munich, DE, 12.-15.11.2019
NEPCON Japan, Tokio, 15.1.-17.1.2020
DistribuTECH, 28.1.-30.1.2020, San Antonio, TX
BATTERY JAPAN, Tokyo, 26.2.-28.2.2020
BATTERY JAPAN, Osaka, 9.9.-11.9.2020
IMTS2020, 14.9.-19.9.2020, Chicago, IL
ISE 2021, Barcelona, Spain, 2.2.-5.2.2021
DistribuTECH, 9.2.-11.2.2021, San Diego, CA
BATTERY JAPAN, Tokyo, 3.3.-5.3.2021
BATTERY JAPAN, Osaka, 29.9.-1.10.2021
DistribuTECH, 25.1.-27.1.2022, Dallas, TX
DistribuTECH, 7.2.-9.2.2023, San Diego, CA
DistribuTECH, 23.1.-25.1.2024, Indianapolis, IN
DistribuTECH, 11.2.-13.2.2025, Dallas, TX
Interesting video
The ISS Design Challenge ...
Interesting video
Mouser Electronics Warehouse Tour with Grant Imahara
naše portály dle jazyka:

česko/slovenská jazyková verze:

anglická jazyková verze:

polská jazyková verze:

ruská jazyková verze:
naše portály dle zaměření:




MALUTKI media s.r.o.
Těrlická 475/22
735 35 Horní Suchá
tel. 00420-603531605
e-mail: info@malutki-media.com

All trademarks are the property of their respective owners.
ISSN 1801-3813