Mauri's interests

Mauri's posts

Post has attachment

Public

Amazon Alexa "DX Cluster" skill proof of concept. This demo shows how to listen the latest DX Cluster spots and asking Alexa to set the Elecraft KX3 radio remotely on the frequency to listen the ham radio activity.

Post has attachment

Public

Mauri Niininen commented on a post on Blogger.

Hi Douglas

There is in total some 30+ articles in my ham radio blog.

There is in total some 30+ articles in my ham radio blog.

Post has attachment

Public

My second book is now in editing phase. Here is a short summary of the creative process I am using and what I have learned in the last 12 months about writing books.

Post has attachment

A simple weekend project: Remote control of Elecraft KX3 with audio streaming using Raspberry Pi 2 and Android phone.

Post has attachment

Post has attachment

A new LSTM RNN based Morse decoder built using TensorFlow. This learns Morse code from bit patterns and is able to translate noisy patterns directly to characters. Without any specific optimization the decoder achieved 96.3% accuracy with a training set of 114,000 samples.

Post has shared content

I am trying to use Keras LSTM Recurrent Neural Network to decode noisy Morse Code (CW). Since Morse code is a human language decoding it has similar challenges than automatic speech recognition or handwriting recognition - speed and rhythm variances, signal noise, interference or fading etc.

I posted a simple proof of concept on my blog with Python sources. I am looking for feedback and comments from machine learning experts as I am trying to learn how to apply these algorithms to solve this problem. Thanks!

I posted a simple proof of concept on my blog with Python sources. I am looking for feedback and comments from machine learning experts as I am trying to learn how to apply these algorithms to solve this problem. Thanks!

I created LSTM RNN neural network using Keras framework to experiment with noisy Morse code. Read more details from my blog.

Any comments and feedback would be much appreciated.

Any comments and feedback would be much appreciated.

Post has shared content

I am trying to use Keras LSTM Recurrent Neural Network to decode noisy Morse Code (CW). Since Morse code is a human language decoding it has similar challenges than automatic speech recognition or handwriting recognition - speed and rhythm variances, signal noise, interference or fading etc.

I posted a simple proof of concept on my blog with Python sources. I am looking for feedback and comments as I am trying to learn how to apply these algorithms to solve this problem. Thanks!

I posted a simple proof of concept on my blog with Python sources. I am looking for feedback and comments as I am trying to learn how to apply these algorithms to solve this problem. Thanks!

I created LSTM RNN neural network using Keras framework to experiment with noisy Morse code. Read more details from my blog.

Any comments and feedback would be much appreciated.

Any comments and feedback would be much appreciated.

Post has shared content

I am trying to use Keras LSTM Recurrent Neural Network to decode noisy Morse Code (CW). Since Morse code is a human language decoding it has similar challenges than automatic speech recognition or handwriting recognition - speed and rhythm variances, signal noise, interference or fading etc.

I posted a simple proof of concept on my blog with Python sources. I am looking for feedback and comments from deep learning experts as I am trying to learn how to apply these algorithms to solve this problem. Thanks!

I posted a simple proof of concept on my blog with Python sources. I am looking for feedback and comments from deep learning experts as I am trying to learn how to apply these algorithms to solve this problem. Thanks!

I created LSTM RNN neural network using Keras framework to experiment with noisy Morse code. Read more details from my blog.

Any comments and feedback would be much appreciated.

Any comments and feedback would be much appreciated.

Post has attachment

Any comments and feedback would be much appreciated.

Wait while more posts are being loaded