Friday April 9 2021

The program consists of a series of plenary talks and contributed talks.

The recorded PeRSonAl workshop can be found here using SlidesLive.

If you are attending PeRSonAl at MLSys 2021 please, optionally, add yourself to this Google spreadsheet. The sheet helps us understand the background of participants and organize future workshops.


Time (EST)TopicSpeaker
9:15am – 9:30am Welcome to the 3rd PeRSonAl WorkshopCarole-Jean Wu/FAIR; Udit Gupta/FAIR-Harvard
9:30am – 10:00am Explainable ML for Recommender Systems: Challenges and OpportunitiesHima Lakkaraju/Harvard University
10:00am – 10:30amA Memory-centric Approach in Designing System Architectures for Personalized RecommendationsMinsoo Rhu/KAIST
10:30am – 10:45amMERCI: Efficient Embedding Reduction on Commodity Hardware via Sub-Query Memoization.
Yejin Lee, Seong Hoon Seo, Hyunji Choi, Hyoung Uk Sul, Soosung Kim, Jae W. Lee, Tae Jun Ham (Seoul National University)
10:45am – 11:00amErasure Coding Based Fault Tolerance for Recommendation Model TrainingKaige Liu (Facebook), Jack Kosaian, Rashmi Vinayak (CMU)
11:00am – 11:15amElliot: A Comprehensive and Rigorous Framework For Reproducible Recommender Systems Evaluation
Vito Walter Anelli (Polytechnic University of Bari), Alejandro Bellogín (Universidad Autónoma de Madrid), Antonio Ferrara, Daniele Malitesta, Felice Antonio Merra, Claudio Pomo (Polytechnic University of Bari), Francesco Maria Donini (Università della Tuscia), Tommaso Di Noia (Polytechnic University of Bari)
11:15am – 11:30amOptimizing Deep Learning Recommender SystemsTraining on CPU Cluster ArchitecturesDhiraj Kalamkar, Evangelos Georganas, Sudarshan Srinivasan, Jianping Chen, Mikhail Shiryaev, and Alexander Heinecke (Intel)
11:30am – 11:45amMain-Memory Acceleration for Bandwidth-Bound Deep Learning InferenceBenjamin Cho, Jeageun Jeung, Mattan Erez (UT Austin)
11:45am – 12:00pmDeepRecSys: A System for Optimizing End-To-End At-scale Neural Recommendation InferenceUdit Gupta (Harvard University/FAIR), Samuel Hsia (Harvard University), Vikram Saraph, Xiaodong Wang, Brandon Reagen (Facebook), Gu-Yeon Wei (Harvard University), Hsien-Hsin S. Lee (FAIR), David Brooks (Harvard University), Carole-Jean Wu (FAIR)
12:00pm – 1:00pmCoffee/Lunch Break
1:00pm – 2:00pm Keynote: From Recommender Systems to Natural Language Processing and Back AgainJulian McAuley/UCSD
2:00 — 2:30 pmRevisiting Recommender Systems on the GPUEven Oldridge/NVIDIA
2:30pm – 3:00pm Coffee/Lunch Break
3:00pm – 3:30pmLow-Precision Hardware Architectures Meet Recommendation Model Inference at ScaleSummer Deng/Facebook
3:30pm – 4:00pm Pushing the Limits of Recommender Training Speed:  An MLPerf ExperienceTayo Oguntebi/Google
4:00pm – 4:15pmCross-Stack Workload Characterization of Deep Recommendation SystemsSamuel Hsia, Udit Gupta, Mark Wilkening (Harvard University), Carole-Jean Wu (FAIR), Gu-Yeon Wei, David Brooks (Harvard University)
4:15pm – 4:30pmAccelerated Learning by Exploiting Popular ChoicesMuhammad Adnan, Yassaman Ebrahimzadeh Maboud, Divya Mahajan (Microsoft), Prashant Nair (University of British Columbia)
4:30pm – 4:45pmTowards Disaggregated Memory RecommendersTalha Imran (Penn State), Nadav Amit, Irina Calciu (VMWare Research)
4:45pm – 5:00pmScalability, Latency, Flexibility: The Case for Similarity Search as a ServiceAmir Sadoughi, Edo Liberty, Lior Ehrenfeld, Ron Begleiter, Fei Yu, Mark Chew, Jack Pertschuk, Roei Mutay, Greg Kogan, Beni Ran (Pinecone)
5:00pm – 5:15pm Capacity-Driven Scale-Out Neural Recommendation: Enabling the Growing Scale of RecommendationMike Lui (Drexel University, Facebook), Yavuz Yetim, Oz Ozkan, Zhuoran Zhao, Shin-Yeh Tsai, Carole-Jean Wu (Facebook), Mark Hempstead (Tufts Unviersity)
5:15pm – 5:30pmTraining with Multi-Layer Embeddings for Model ReductionBenjamin Ghaemmaghami, Zihao Deng, Benjamin Cho, Leo Orshansky (UT Austin), Ashish Kumar Singh (E2OPEN), Mattan Erez, Michael Orshansky (UT Austin)
5:30pm – 5:45pmTowards Automated Neural Interaction Discovery for Click-Through Rate PredictionQingquan Song (Texas A&M University), Dehua Cheng, Hanning Zhou, Jiyan Yang, Yuandong Tian (Facebook), Xia Hu (Texas A&M University)
5:45pm – 6:00pmClosing sessionCarole-Jean Wu/FAIR; Udit Gupta/FAIR-Harvard

Details for the contributed talks, including the abstracts and speaker bio’s can be found here.

Call for Talk Participation

Personalized recommendation is the process of ranking and recommending content based on users’ personal preferences. Recommendation algorithms are central to providing personalized search results, marketing strategies, e-commerce product suggestions, and entertainment content. Given the pervasive use of personalized recommendations across many Internet services, state-of-the-art recommendation algorithms are using increasingly more sophisticated machine learning approaches. These advances have led to personalized recommendation algorithms consuming a large fraction, and in many cases the majority, of AI cycles and datacenter capacity. Thus, the unique demands of recommendation algorithms must be met with innovative solutions across the computing stack.

The PeRSonAl workshop invites submissions across all sub-areas in algorithms, datasets, and systems and hardware related to personalized recommendation. Topics of interest include but are not limited to:

  • Emerging algorithms for personalized recommendation
  • Datasets to train and test recommendation algorithms
  • Specialized systems and hardware
  • Novel applications of recommendation algorithms
  • Case studies and prototypes of training and deploying recommendation systems

As the workshop will be hosted virtually it will comprise pre-recorded, 30-minute presentations based on authors’ submissions. Submissions can be up to 2 pages excluding reference (following the same formatting guidelines as the conference). Submissions should be sent to and

Important Dates

  • Paper submission deadline: March 26, 2021
  • Paper notification: March 29, 2021
  • Pre-recorded presentation deadline: April 5, 2021
  • PeRSonAl workshop: April 9, 2021