Listen to the ‘perfect Christmas song’ created by AI: Software trained on existing tracks can now write its own hits with festive titles like ‘Cinnamon Hollybells’
- Team called Made by AI trained neural network by inputting Christmas tunes
- Catchy tunes by the system include Syllabub Chocolatebell and Cocoa Jollyfluff
- Neural networks are a type of program that imitate the way that brains learn
- It was taught using more than one hundred digital files of festive songs
Catchy Christmas songs can now be created by a special songwriting AI, taught by studying existing festive tunes.
The system came up with catchy jingles with names like ‘Syllabub Chocolatebell’, ‘Peaches Twinkleleaves’ and ‘Cocoa Jollyfluff’.
Researchers from Made by AI trained a neural network by inputting one hundred Christmas tunes in the form of Musical Instrument Digital Interface (MIDI) files.
It then picked out recurring themes, motifs, instruments and rhythms to generate its own hits.
Play the video to hear the Christmas hit made by AI
Scientists have trained an AI system to write its own catchy Christmas songs by teaching it existing festive tunes. The system came up with catchy tunes with names like ‘Syllabub Chocolatebell, ‘Peaches Twinkleleaves’ and ‘Cocoa Jollyfluff’ (stock image)
A spokesperson from Made by AI said: ‘The tunes generated with AI have similarities with classical memorable Christmas music – and that is the point – otherwise it would be a random tune with no connection to the upcoming Holidays.’
Neural networks are a type of computer program that imitate the way that brains learn to solve problems.
MIDI versions of Christmas songs provided a broad base for the team to train the software in order for it to imitate Christmas songs.
Netflix is testing an instant-replay feature that lets you…
In the footsteps of giants: ‘Treasure trove’ haul of 145…
People who hold radical political views both on the far left…
Artificial islands built in the Netherlands will bring…
Share this article
Neural networks tend to pick up tone and vocabulary easily, but struggle with making sense.
So, the carol-trained network learned to produce a lot of lines that sounded joyful.
The software was able to pick out themes synonymous with Christmas to come up with songs of its own.
It generated five tunes which are now available on the Made by AI Soundcloud page.
The AI also came up with the names for the tunes.
The team noted that they only had limited time to work on the network but said with ‘more work and adherence to musical concepts such as structure and repetition, it might be possible to generate something a little more catchy’.
Researchers from Made by AI trained a neural network by inputting one hundred Christmas tunes in the form of Musical Instrument Digital Interface (MIDI) files. MIDI versions of Christmas songs provided a broad base for the team to train the model for it to imitate Christmas songs (stock image)
HOW DID THE TEAM TRAIN THEIR NEURAL NETWORK TO IMITATE ITS OWN CHRISTMAS SONGS?
The first step was to analyse the dataset by going through the existing tunes and storing the notes, chords and the sequences that were used in each tune.
They used a long short-term memory network (LSTM).
The training consists of modelling the LSTM network that would ‘learn’ these sequences of notes and chords, with the goal of being able to generate its own sequences when fully learned.
The training took approximately 3 hours with the Graphics Processing Unit instance, used for deep learning purposes.
This was by any means more computation than needed, but we knew we had little time and could run into the need to optimise hyperparameters in the model, which we also did with trying out different batch sizes on the model.
Source: Read Full Article