About Me

I am a Research Scientist at Spotify Research in New York City, working on personalization and recommendation algorithms. Prior to Spotify, I was a Postdoctoral Fellow at the Spiegel Research Center at Northwestern University working on personalization and recommender systems for media and news. I obtained my Ph.D. in Information & Computer Science at the University of Colorado Boulder in 2020 under the supervision of Prof. Robin Burke. With over 10 years of experience in designing recommender systems and conducting research in this area, I am happy to apply my expertise to build systems that benefit hundreds of millions of users. In particular, I am known for my work on popularity bias in recommender systems and for developing algorithms to tackle such bias from the perspective of multiple stakeholders. My research in popularity bias and multi-stakeholder recommender systems has garnered more than 1600 citations since 2017, indicating the influence of my work in the research community. In addition to being active in the research community, I also have experience in building end-to-end large-scale Machine Learning and recommendation algorithms in the industry.

I enjoy yoga, calligraphy, mountains, walking in New York City, cooking, traveling around the world, and exploring new coffeeshops & restaurants.

For speaking, consultation, and educational opportunities in areas including machine learning & artificial intelligence, recommender systems & personalization, and large language models (LLMs) please reach out to me via the email listed in the contact section.

Research

Calibrated Recommendations as a Minimum-Cost Flow Problem

Himan Abdollahpouri, Zahra Nazari, Alex Gain, Clay Gibson, Maria Dimakopoulou, Jesse Anderton, Benjamin Carterette, Mounia Lalmas, Tony Jebara (WSDM 2023)

Calibration in recommender systems has recently gained significant attention. In the recommended list of items, calibration ensures that the various (past) areas of interest of a user are reflected with their corresponding proportions. For instance, if a user has watched, say, 80 romance movies and 20 action movies, then it is reason-able to expect the recommended list of movies to be comprised of about 80% romance and 20% action movies as well. In this paper, we propose a novel approach based on the minimum-cost flow problem for generating calibrated recommendations. We demonstrate the superior performance of our proposed approach compared to the state-of-the-art in generating relevant and calibrated recommendation lists.
[Paper] [Blog Post]

Multistakeholder Recommender Systems

Himan Abdollahpouri and Robin Burke (Recommender Systems Handbook 2022)

Multistakeholder recommendation is the term applied when a recommender system is designed, implemented and/or evaluated taking into account the perspectives of multiple stakeholder groups. Published studies of recommender systems largely concentrate on optimizing only for user experience. However, a user-centric approach does not account for system objectives such as fairness, balance, and profitability, and obscures concerns that might come from other stakeholders, such as the providers or sellers of items being recommended. This chapter describes the characteristics of multistakeholder recommendation and particular considerations involved in creating and evaluating multistakeholder recommender systems, and demonstrates these ideas in a multistakeholder evaluation of existing algorithms.
[Paper]

Multistakeholder recommendation: Survey and research directions

Himan Abdollahpouri and Robin Burke (Recommender Systems Handbook 2022)

Academic studies of recommender systems have largely concentrated on optimizing for user experience when designing, implementing and evaluating their algorithms and systems. However, this concentration on the user has meant that the field has lacked a systematic exploration of other aspects of recommender system outcomes. A user-centric approach limits the ability to incorporate system objectives, such as fairness, balance, and profitability, and obscures concerns that might come from other stakeholders, such as the providers or sellers of items being recommended. Multistakeholder recommendation has emerged as a unifying framework for describing and understanding recommendation settings where the end user is not the sole focus. This article outlines the multistakeholder perspective on recommendation.
[Paper]

The Unfairness of Popularity Bias in Recommendation

Himan Abdollahpouri and Robin Burke (Recommender Systems Handbook 2022)

Recommender systems are known to suffer from the popularity bias problem: popular (i.e. frequently rated) items get a lot of exposure while less popular ones are under-represented in the recommendations. In this paper, we look at this problem from the users' perspective: we want to see how popularity bias causes the recommendations to deviate from what the user expects to get from the recommender system. We define three different groups of users according to their interest in popular items (Niche, Diverse and Blockbuster-focused) and show the impact of popularity bias on the users in each group. Our experimental results on a movie dataset show that in many recommendation algorithms the recommendations the users get are extremely concentrated on popular items even if a user is interested in long-tail and non-popular items showing an extreme bias disparity.
[Paper]

A Graph-Based Approach for Mitigating Multi-Sided Exposure Bias in Recommender Systems

Masoud Mansoury, Himan Abdollahpouri, Mykola Pechenizkiy, Bamshad Mobasher, Robin Burke (ACM Transactions on Information Systems (TOIS) 2021)

Fairness is a critical system-level objective in recommender systems that has been the subject of extensive recent research. A specific form of fairness is supplier exposure fairness, where the objective is to ensure equitable coverage of items across all suppliers in recommendations provided to users. In this article, we introduce FairMatch, a general graph-based algorithm that works as a post-processing approach after recommendation generation to improve exposure fairness for items and suppliers. The algorithm iteratively adds high-quality items that have low visibility or items from suppliers with low exposure to the users’ final recommendation lists. We show that FairMatch, although it significantly improves exposure fairness and aggregate diversity, maintains an acceptable level of relevance of the recommendations.
[Paper]

User-centered Evaluation of Popularity Bias in Recommender Systems

Himan Abdollahpouri, Masoud Mansoury, Robin Burke, Bamshad Mobasher, Edward Malthouse (UMAP 2021)

Recommendation and ranking systems are known to suffer from popularity bias. Prior research has examined various approaches for mitigating popularity bias and enhancing the recommendation of long-tail, less popular, items. The effectiveness of these approaches is often assessed using different metrics to evaluate the extent to which over-concentration on popular items is reduced but not much attention is give to how different users with different levels of interest towards popular items are affected by such algorithms. In this paper, we show the limitations of the existing metrics to evaluate popularity bias mitigation and we propose new metrics that can address these limitations. In addition, we present an effective approach that mitigates popularity bias from the user-centered point of view.
[Paper]

News

(Sep 22nd 2022) Co-organizer of RecSys MORS2022 workshop on Multi-objective Recommender Systems.

(Oct 3 2021) Moved to New York City.

(Apr 11 2022) Invited Talk at Norwegian School of Economics, Norway.

(Sep 25 2021) Co-organizer of RecSys MORS2021 workshop on Multi-objective Recommender Systems.

(Sep 9 2021) Started working at Spotify Research.

(May 6 2021) Invited Talk at Center for Data Science

(Aug 2020) Started my postdoc at Spiegel Research Center at Northwestern University.

(Jul 9th 2020) Defended My PhD (University of Colorado Boulder).

Get in touch