Answering counterfactual “what-if” questions in a trustworthy and efficient manner
April 20, 2023 12:56 pmAnswering causal questions with machine learning algorithms is a challenging yet critical task.
Categories for Machine Learning
Estimating long-term effects when only short-run experiments are available
April 6, 2023 8:57 amQuantifying cause and effect relationships is of fundamental importance in many fields, from medicine to economics. The gold standard solution to this problem is to conduct randomised controlled trials, or A/B tests.
Gaining confidence in synthetic control causal inference with sensitivity analysis
April 5, 2023 11:13 amUnderstanding cause and effect relationships in Spotify data to inform decision-making is crucial for best serving Spotify’s users and the company.
Users’ interests are multi-faceted: recommendation models should be too
February 22, 2023 10:58 amA new approach to calibrating recommendations to user interests. Users’ interests are multi-faceted and representing different aspects of users’ interest in their recommendations is an important factor for recommender systems....
Survival Analysis Meets Reinforcement Learning
November 25, 2022 10:08 amTL;DR: Survival analysis provides a framework to reason about time-to-event data; at Spotify, for example, we use it to understand and predict the way users might engage with Spotify in the future. In this work, we bring temporal-difference learning, a central idea in reinforcement learning, to survival analysis. We develop a new algorithm that trains a survival model from sequential data by leveraging a temporal consistency condition, and show that it outperforms direct regression on observed outcomes......
Spotify’s Contributions to NeurIPS 2022
November 24, 2022 11:08 amCutting-edge research in Machine Learning, Language Technologies, User Modeling, Audio Intelligence, Search and Recommender Systems are some of the key areas we feel incredibly enthusiastic about at Spotify...
Can we correctly attribute changes among many possible causes when unobserved confounders are present?
November 24, 2022 11:06 amThere are many A/B tests we might like to run, but which are too technically challenging, risky in terms of user impact or even impossible to perform. For instance, in the classic example of whether smoking causes lung cancer, forcing a randomly selected group of people to smoke is unethical if we believe it might damage their health. In the context of technology companies, if we want to understand if app crashes cause users to churn, we would have to randomly select a subgroup of users and crash their apps on purpose – not something we would want to consider given we do not want to break their trust....
Scalable Dynamic Topic Modeling
November 15, 2022 12:51 pmDynamic topic modeling is a well established tool for capturing the temporal dynamics of the topics of a corpus....
Unsupervised Speaker Diarization using Sparse Optimization
September 29, 2022 11:08 amWhat is Speaker Diarization? Speaker diarization is the process of logging the timestamps of when various speakers take turns to talk...
Identifying Promising New Podcasts
September 22, 2022 9:05 amA large number of new podcasts are launched every month on Spotify and other online media platforms. In this work,... View Article
Spotify’s Contributions to RecSys 2022
September 14, 2022 2:12 pmHere at Spotify, we are highly dedicated to cutting-edge research in various areas in Machine Learning, User Modeling, Personalization, and... View Article
The Contribution of Lyrics and Acoustics to Collaborative Understanding of Mood
July 19, 2022 2:49 pmSong lyrics make an important contribution to the musical experience, providing us with rich stories and messages that artists want... View Article
Modeling Users According to Their Slow and Fast-Moving Interests
February 21, 2022 4:10 pmMusic recommendation systems at Spotify are built on models of users and items. They often rely on past user interactions... View Article
Multi-Task Learning of Graph-based Inductive Representations of Music Content
October 29, 2021 5:28 pmPersonalization services at Spotify rely on learning meaningful representations of tracks and users to surface apt recommendations to users in... View Article
Model Selection for Production System via Automated Online Experiments
July 9, 2021 3:32 pmMachine learning practitioners regularly have to select the best model to deploy in production. At Spotify, for instance, we might... View Article
Finding Structure in Users’ Evolving Listening Preferences
April 19, 2021 2:23 pmAt Spotify, we invest into designing recommendation algorithms that allow users to explore the music space more effectively. Recent findings... View Article
Collaborative Classification from Noisy Labels
April 8, 2021 2:23 pmOn online platforms such as Spotify, users have access to large collections of content. Most of the time, these collections... View Article
Shifting Consumption towards Diverse content via Reinforcement Learning
March 8, 2021 2:23 pmAlgorithmically generated recommendations power and shape the bulk of music consumption on music streaming platforms. The ability to shift consumption... View Article
Stochastic Variational Inference for Dynamic Correlated Topic Models
February 17, 2021 6:44 pmTopic models are useful tools for the statistical analysis of data as well as learning a compact representation of co-occurring... View Article
Spotify’s Contributions to NeurIPS 2020
December 7, 2020 7:02 pmThe Thirty-fourth Conference on Neural Information Processing Systems (NeurIPS 2020) begins this week, and we are excited to be a... View Article
The Million Playlist Dataset… Remastered
September 28, 2020 9:52 am Leave your thoughtsIn 2018, Spotify helped organize the RecSys Challenge 2018, a data science research challenge focused on music recommendation, specifically the... View Article
Spotify’s Contributions to RecSys 2020
September 21, 2020 12:45 pm Leave your thoughtsThe 14th ACM Conference on Recommender Systems (RecSys 2020) begins this week, and we are excited to be a Platinum... View Article
Understanding by Unmixing
August 25, 2020 1:19 pm 1 CommentMaking sense of music by extracting and analyzing individual instruments in a song Source Separation at Spotify At the end... View Article
Spotify at RecSys 2019
October 1, 2019 3:54 pm Leave your thoughtsSpotify was out in full force at RecSys in Copenhagen. The ACM Conference on Recommender Systems (RecSys) is the premier... View Article