[SFDXA] Morse Learning Machine Challenge Catching on with Hams

[Bill]

 From Tony N2MFT:


    Morse Learning Machine Challenge Catching on with Hams


    09/09/2014

Experimenter*Mauri Niininen* <mailto:mauri at innomore.com>, AG1L, of 
Lexington, Massachusetts, reports that his*Morse Learning Machine 
Challenge* 
<http://ag1le.blogspot.com/2014/09/morse-learning-machine-challenge.html>has 
been catching on among members of the Amateur Radio community.The goal 
of the competition is to build a machine that can learn how to decode 
audio files containing Morse code — a better “code trap,” if you will. 
Niininen said his project has been approved by*Kaggle* 
<https://inclass.kaggle.com/c/morse-challenge>, which bills itself as 
“the world's largest community of data scientists.” Niininen said that 
it takes humans many months of effort to learn Morse code, and, after 
years of practice, the most proficient operators can decode Morse code 
up to 60 or more words per minute

“Humans have extraordinary ability to quickly adapt to varying 
conditions, speed and rhythm. We want to find out if it is possible to 
create a machine learning algorithm that exceeds human performance and 
adaptability in Morse decoding.”

The computer-generated Morse data for the competition includes various 
levels of added noise. The signal-to-noise ratio, speed, and message 
content of the files vary randomly to simulate real-life ham radio HF 
Morse communication.

“I hope to attract people from the Kaggle community, who are interested 
in solving new, difficult challenges using their predictive data 
modeling, computer science and machine learning expertise,” Niininen added.

During the competition, participants will build a learning system 
capable of decoding Morse code, using development data consisting of 200 
WAV audio files containing short sequences of randomized Morse. Data 
labels are provided for a training set, so participants can self 
evaluate their systems.

“To evaluate their progress and compare themselves with others, they can 
submit their prediction results online to get immediate feedback,” he 
explained. “A real-time Kaggle*leader board* 
<https://inclass.kaggle.com/c/morse-challenge/leaderboard/public>shows 
participants their current standing based on their validation set 
predictions.” Niininen has provided a sample/Python/Morse decoder to 
make it easier to get started.

Niininen said that within the first 24 hours of the competition, he had 
33 downloads. “We have already 53 downloads of the materials for this 
competition,” he said on September 5, “and it is growing by the hour, as 
the word about this challenge is spreading.”

http://www.arrl.org/news/morse-learning-machine-challenge-catching-on-with-hams