Robotics have taken the world by storm, they’ve been and will continue being part of our lives in the future. These machines have changed so many aspects of our lives for the better – they’ve become our personal bodyguards, they drive cars for us, cook for us and now… they’re rapping too. You would think that nothing innovators come up with can surprise us anymore but this one definitely caught my attention: an artificial intelligence program that generates rap lyrics and get this, does an even better job than top Hip-Hop artists themselves. I introduce to you, “DeepBeat”.
Researchers from the Aalto University in Finland have invented a machine-learning system, nicknamed “DeepBeat”, which analysed thousands of rap songs from different rappers such as Eminem, Jay-Z and Lil’Wayne. After analysing all the songs from its database, DeepBeat then chooses and puts together lines from the different songs in order to form its own song.
According to the research paper “DopeLearning: A Computational Approach to Rap Lyrics Generation”, Eric Malmi and his colleagues from the university were aiming to offer a method for capturing two aspects of rap- that rap lyrics seek both creativity and excellent lyrical skills in order to create intricate rhyme patterns.
DeepBeat was given just one line to work with but amazingly repeated the process various times until it was able to write a whole song on its own. According to the lead researcher Eric Malmi, DeepBeat not only succeeded in forming a rap song but after observing different rappers’ rhyme density, it performs even better than leading rappers by a whopping 21%.
It is amazing how robots are silently taking over, and will soon be able to perform most human functions. DeepBeat still has a long way to go before it can master the ranks but we’re certainly impressed with what it can do so far. Maybe now we can expect Hip-Hop Awards to add a new category called the “Best Robotic Rapper Award”, or not. Either way, we’ll be routing for DeepBeat. The “dopest” robot alive!