Cover story
Figure 2: Machine learning – training the machine
to either the inputs or to the next layers; see Figure 1. There are several types of neural networks. The
more we move to deep learning, the more complex the networks. Deep learning also incorporates some feedback mechanisms, whereas simple ML models have simple forward actions, moving from the data to the output or result.
How can a machine be trained? The fi rst step is data collection; see Figure 2. As we focus on supervised learning, we collect labelled data, so that patterns can be found correctly. The quality of this data will determine how accurate the model is. We need to put it together and make it random, as, if it is too organised, models will not be created correctly, rendering bad algorithms.
The second step is to clean and remove unwanted data.
Any set where features are missing should be removed. Any states that are unknown or where the data is not needed should also be removed. Data must be then separated into two parts, one for training and the other for testing.
The third step is training the algorithm. This is split it into
three steps. The fi rst step is to choose the machine learning classifi cation algorithm. Several ones are available and are suitable for different types of data. Example of machine learning classifi cation algorithm are: • Bonsai; • Decision Tree Ensemble; • Boosted Tree Ensemble; • TensorFlow Lite for Microcontrollers; • PME.
It is important to choose the right model composition as this determines the output after running the ML algorithm
on the collected data. This may need some data scientist skills but could also be left to the automatic engine provided by several model creation tools. The next step is the model training operation, which consists of running several iterations to improve the weights of the different layers and the overall accuracy of the model. We then need to evaluate the model, which is done by testing the model with a subset of data – the one that was kept for testing and evaluation. This set of data is unknown to the model. We can then compare the model output to the well-known results. Once these steps have been completed, we can use
the created model and validate the results by performing inference on targets. It’s best to take the model in the fi eld, provide it with real inputs and see if the results are correct.
Solutions by Microchip Microchip’s microcontrollers and microprocessors support many applications such as smart embedded vision. They are also a good fi t for predictive maintenance based on either vibration, power measurement or sound monitoring. Microchip portfolio can be used in gesture recognition and, coupled with touch capabilities, makes it easier to control human machine interfaces. Microchip provides high-performance PCI switches, which enable the interconnection of GPUs and help with model training. Data collection can be done using microcontrollers, microprocessor units, FPGAs and sensors. Data validation and inference operation can both be done on microcontrollers, microprocessors or FPGAs. Overall, these solutions make machine learning easy to implement. For more information go to:
https://www.microchip.com/en-us/solutions/ technologies/machine-learning
www.electronicsworld.co.uk May 2024 07
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48
