AI Business is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 3099067.

How Can Machine Learning Help You Find New Podcasts?

by Ed Lauder
Article Image

Acast's Johan Billgren and Caitlin Thompson have detailed how their recommendation engine, powered by machine learning, will help you to find new podcasts.

Finding podcasts has always been a bit of a struggle. People tend to rely on ‘Best of’ lists, word of mouth, the related tab that comes up on iTunes… However, a company called Acast has applied machine learning to the problem. In an interview with Forbes, Acast’s Johan Billgren and Caitlin Thompson detailed how the platform works.

“In layman's terms, you could say that machine learning is all about vast amounts of data,” explained Billgren. “And what you do is create a training set, and you have a standard algorithm or a machine learning algorithm, and it tries different patterns like an equation that has different parameters.”

According to Billgren, they tested the new algorithm on 20% of their users to see whether it worked, and the results were very positive, so much so that they launched the full update to all users in February 2017.

Billgren told Forbes’ Sarah Rhea Werner that they “used six months worth of anonymous listening data and trained our algorithm to basically make recommendations based on a large number of users and user behavior instead of the standard categories and tags and keywords.”

“And it turns out that it's very efficient and gives you really good recommendations,” he finished. He went on to claim that their recommendation engine is “going to get better and better over time, and we will also find more uses for it.”

At the moment, many podcasts listeners are finding new material to listen to via iTunes’ own suggestions feature, yet Billgren explained how Acast’s machine learning powered recommendation engine differs from Apple’s. “In our examples, we get a lot of results and recommendations that are not obvious, but still correct,” he began. “They are simply recommendations you could not have predicted from categories, descriptions or even basic user behavior.”

Acast has been developing their algorithm to concentrate on the user rather than the podcast itself, which means that you may end up getting recommendations that are wildly different to the podcast you are currently listening to.

“Basically, the recommendation engine will recommend your show if it fits the person, because we have unique recommendations for every user. It's basically up to the user behavior of many users,” he highlighted.

So, through Acast’s machine learning-powered recommendation engine, you’ll be able to listen to the podcasts that suit you, which might include our very own AI Business Podcast. You can listen to it via iTunes and SoundCloud.

Practitioner Portal - for AI practitioners

Story

IBM donates AI fairness and explainability tools to the Linux Foundation

6/29/2020

Three projects move under the wing of the open source organization

Story

Scoping machine learning projects: The six questions each analytics translator has to know

6/26/2020

Innovative ML projects can only succeed if they manage to transform a business problem into clear tasks data scientists can work on

Practitioner Portal

EBooks

More EBooks

Upcoming Webinars

Experts in AI

Partner Perspectives

content from our sponsors

Research Reports

9/30/2019
More Research Reports

Infographics

AI tops the list of most impactful emerging technologies

Infographics archive

Newsletter Sign Up


Sign Up