My two rap-related projects, Raplyzer, which analyzes the rhyme density of different rappers, and DeepBeat, which is a rap lyrics generating AI, were widely covered in the media last year. But with the fame come the haters. The purpose of this post is to prove that my haters are wrong! (For real: I honestly don’t consider anyone a hater, nor will there be any proofs in this post. Rather, I’ll present some quantitative evidence for the validity of the algorithms but also discuss their limitations.)
Most of us are aware that the websites we visit can reveal quite a lot about ourselves – and that these revelations are highly sought after by advertisers. Nearly all companies today are eager to access data about our online behavior so they can show us more relevant content (those Facebook ads can sometimes be scarily accurate).
We did – in our paper titled “Beyond rankings: comparing directed acyclic graphs” (pdf) which I’ll be presenting at the ECML PKDD conference in Portugal next month. This was the first project of my PhD, but there’s also something else that makes it fundamentally different from the other research projects I’ve been involved with.
Typically, when I undertake a research project, I have a concrete question, like what is the next location a person will visit, to which I start looking for different solutions. In other words, I begin with a nail and start looking for a suitable hammer. However, this time we started by developing a cool new hammer with some neat theoretical properties before we had any idea if a suitable nail even exists. Continue reading
“Men lie, women lie, numbers don’t” – Jay Z
Among the many things rappers like to boast about, some are relatively easy to quantify, like money, whereas rhyming skills are something that have been very difficult to measure – up till now. In this post, I’ll present Raplyzer, a computer program which automatically detects rhymes from rap lyrics and which is used to rank popular rappers based on their average Rhyme factor. I’ll also present another program called BattleBot, which is a search engine for rhyming rap lines based on the algorithm used in Raplyzer.
Usually, when I think I’ve come up with a great idea, I wait until the next day to see if it still seems as good. Most of the time it doesn’t. However, when Arno Solin first told me about the HisKi database, which contains digitized church records (births, deaths, marriages, migration) from Finland spanning from the 1600s to the late 1800s, and the analysis possibilities this data could offer, I immediately felt compelled to start working on it, and the next day I was even more excited.
Eventually, I told my professor about the idea and we decided that I would start my PhD research around the questions arising from the HisKi dataset. Other people also liked the idea and so I was chosen to present my research in a pitching competition called Falling Walls Lab in Berlin a few months ago (Aalto news wrote about my trip here). Here’s a video of my two and a half minutes presentation where I explain what kind of research questions I’m aiming to address.
(An English summary for this and the previous post can be found here.)
Viime postauksen yhteydessä sain lukijoiltani paljon hyviä ideoita Raplysaattoriin liittyen. Yksi ideoista oli selvittää, millä yksittäisellä kappaleella on kovin riimikerroin. Tässä postauksessa julkaisen yksittäisten kappaleiden top 10 -listan. Continue reading
“Puolet räppäreist ei tajuu rimmaamisest mitään / ennen mikkiin päästämistä pitäis kirjalliset pitää”
Näin toteaa suomiräpin epäilemättä tämän hetken tunnetuin nimi, Cheek, kappaleessaan Kuka muu muka. Tässä kirjoituksessa kuvailen, miten tietokoneella voidaan löytää lyriikoissa esiintyviä riimejä automaattisesti ja tutkin, löytyykö edellä mainitulle Cheekin väitteelle katetta analysoimalla Suomen tunnetuimpien räppäreiden sanoituksia toteuttamallani tietokoneohjelmalla. Ohjelma laskee tunnistamiensa riimien pituuksia sekä arvioi artistin sanavaraston kokoa. Continue reading