Photo

    Maksims Volkovs, PhD

    Layer 6 AI
    Email: maksims [dot] volkovs [at] gmail [dot] com
    scholar  github  linkedin



ABOUT

I received my PhD at the University of Toronto where I was a member of the Machine Learning Group working with Richard Zemel. My research interests broadly span the field of machine learning with emphasis on recommender systems, information retrieval and more recently computer vision. I co-founded Layer 6 AI which was acquired by the TD Bank.

I actively participate in machine learning competitions on Kaggle where I achieved the status of grandmaster, and other platforms. Some recent highlights:

  • 1'st place 2017 and 2018 ACM RecSys Challenges. I led teams that won both 2017 and 2018 ACM RecSys Challenges organized by XING and Spotify respectively. Summaries of our approaches can be found here and here. We also placed 2'nd in the 2019 and 2020 ACM RecSys Challenges.
  • 1'st place 2019 YouTube-8M Video Understanding Challenge. Video temporal concept localization using the YouTube-8M dataset with over 6M videos.
  • 2'nd place Google Landmark Retrieval Challenge. At the time of the challenge this was the largest publicly available image retrieval dataset. Our solution was based on a novel neighbor graph traversal approach described in the CVPR'19 paper.

PUBLICATIONS

  • Weakly Supervised Action Selection Learning in Video
    Junwei Ma, Satya Krishna Gorti, Maksims Volkovs and Guangwei Yu
    CVPR-2021: Computer Vision and Pattern Recognition
    [pdf][code]
  • Development and Validation of a Machine Learning Model Using Administrative Health Data to Predict Onset of Type 2 Diabetes
    Mathieu Ravaut, Vinyas Harish, Hamed Sadeghi, Kin Kwan Leung, Maksims Volkovs, Kathy Kornas, Tomi Poutanen and Laura Rosella
    JAMA Network Open 4, 5 (2021)
    [pdf][link] [press coverage: 1, 2, 3, 4 ]
  • Predicting Adverse Outcomes Due to Diabetes Complications With Machine Learning Using Administrative Health Data
    Mathieu Ravaut, Hamed Sadeghi, Kin Kwan Leung, Maksims Volkovs, Kathy Kornas, Vinyas Harish, Tristan Watson, Gary F. Lewis, Alanna Weisman, Tomi Poutanen and Laura Rosella
    npj Dgital Medicine 4, 24 (2021)
    [pdf][link]
  • HGCF: Hyperbolic Graph Convolution Networks for Collaborative Filtering
    Jianing Sun, Zhaoyue Cheng, Saba Zuberi, Felipe Pérez and Maksims Volkovs
    WWW-2021: International World Wide Web Conference
    [pdf][code]
  • Development of a Multivariable Model for COVID-19 Risk Stratification Based on Gradient Boosting Decision Trees
    Jahir M. Gutierrez, Maksims Volkovs, Tomi Poutanen, Tristan Watson and Laura Rosella
    [medRxiv]
  • TAFA: Two-headed Attention Fused Autoencoder for Context-Aware Recommendations
    Jin Peng Zhou, Zhaoyue Cheng, Felipe Pérez and Maksims Volkovs
    RecSys-2020: ACM Recommender Systems
    [pdf][code]
  • Predicting Twitter Engagement With Deep Language Models
    Maksims Volkovs, Zhaoyue Cheng, Mathieu Ravaut, Hojin Yang, Kevin Shen, Jin Peng Zhou, Anson Wong, Saba Zuberi, Ivan Zhang, Nick Frosst, Helen Ngo, Carol Chen, Bharat Venkitesh, Stephen Gou and Aidan N. Gomez
    RecSys-2020: ACM Recommender Systems
    [pdf]
  • Improving Transformer Optimization Through Better Initialization
    Xiao Shi Huang, Felipe Pérez, Jimmy Ba and Maksims Volkovs
    ICML-2020: International Conference on Machine Learning
    [pdf][supp][code]
  • Guided Similarity Separation for Image Retrieval (Oral)
    Chundi Liu, Guangwei Yu, Cheng Chang, Himanshu Rai, Junwei Ma, Satya Krishna Gorti and Maksims Volkovs
    NeurIPS-2019: Neural Information Processing Systems
    [pdf][code]
  • Cross-Class Relevance Learning for Temporal Concept Localization
    (1'st Place 2019 YouTube-8M Video Understanding Challenge)
    Junwei Ma, Satya Krishna Gorti, Maksims Volkovs, Ilya Stanevich and Guangwei Yu
    ICCV-2019: Workshop on YouTube-8M Large-Scale Video Understanding
    [pdf]
  • Learning Effective Visual Relationship Detector on 1 GPU
    (1'st Place Open Images 2019 Visual Relationship Challenge)
    Yichao Lu, Cheng Chang, Himanshu Rai, Guangwei Yu and Maksims Volkovs
    ICCV-2019: Open Images 2019 Challenge Workshop
    [pdf]
  • Robust Contextual Models for In-Session Personalization
    Maksims Volkovs, Anson Wong, Zhaoyue Cheng, Felipe Perez, Ilya Stanevich and Yichao Lu
    RecSys-2019: ACM Recommender Systems
    [pdf][code]
  • Explore-Exploit Graph Traversal for Image Retrieval
    Cheng Chang, Guangwei Yu, Chundi Liu and Maksims Volkovs
    CVPR-2019: Computer Vision and Pattern Recognition
    [pdf][code]
  • Noise Contrastive Estimation for One-Class Collaborative Filtering
    Ga Wu, Maksims Volkovs, Chee Loong Soon, Scott Sanner and Himanshu Rai
    SIGIR-2019: Special Interest Group on Information Retrieval
    [pdf]
  • Two-stage Model for Automatic Playlist Continuation at Scale
    (1'st Place 2018 ACM RecSys Challenge)
    Maksims Volkovs, Himanshu Rai, Zhaoyue Cheng, Ga Wu, Yichao Lu and Scott Sanner
    RecSys-2018: ACM Recommender Systems
    [pdf][code]
  • DropoutNet: Addressing Cold Start in Recommender Systems
    Maksims Volkovs, Guang Wei Yu and Tomi Poutanen
    NeurIPS-2017: Neural Information Processing Systems
    [pdf][code]
  • Content-based Neighbor Models for Cold Start in Recommender Systems
    (1'st Place 2017 ACM RecSys Challenge)
    Maksims Volkovs, Guang Wei Yu and Tomi Poutanen
    RecSys-2017: ACM Recommender Systems
    [pdf]
  • Two-Stage Approach to Item Recommendation from User Sessions
    Maksims Volkovs
    RecSys-2015: ACM Recommender Systems
    [pdf]
  • Effective Latent Models for Binary Feedback in Recommender Systems
    Maksims Volkovs and Guang Wei Yu
    SIGIR-2015: Special Interest Group on Information Retrieval
    [pdf][code]
  • Context Models For Web Search Personalization
    Maksims Volkovs
    WSDM-2014: WSCD Workshop on Log-based Personalization
    [pdf]
  • New Learning Methods for Supervised and Unsupervised Preference Aggregation
    Maksims Volkovs and Richard Zemel
    JMLR-2014: Journal of Machine Learning Research
    [pdf]
  • Continuous Data Cleaning
    Maksims Volkovs, Fei Chiang, Jaroslaw Szlichta and Renée J. Miller
    ICDE-2014: IEEE International Conference on Data Engineering
    [pdf]
  • Supervised CRF Framework for Preference Aggregation
    Maksims Volkovs and Richard Zemel
    CIKM-2013: International Conference on Information and Knowledge Management
    [pdf][code]
  • Collaborative Ranking with 17 Parameters
    Maksims Volkovs and Richard Zemel
    NeurIPS-2012: Neural Information Processing Systems
    [pdf]
  • Efficient Sampling for Bipartite Matching Problems
    Maksims Volkovs and Richard Zemel
    NeurIPS-2012: Neural Information Processing Systems
    [pdf][supp]
  • Learning to Rank By Aggregating Expert Preferences
    Maksims Volkovs, Hugo Larochelle and Richard Zemel
    CIKM-2012: International Conference on Information and Knowledge Management
    [pdf][code]
  • A Flexible Generative Model for Preference Aggregation
    Maksims Volkovs and Richard Zemel
    WWW-2012: International World Wide Web Conference
    [pdf][code]
  • Learning to Rank with Multiple Objective Functions
    Krysta M. Svore, Maksims Volkovs and Christopher J. C. Burges
    WWW-2011: International World Wide Web Conference
    [pdf]
  • BoltzRank: Learning to Maximize Expected Ranking Gain
    (Best student paper)
    Maksims Volkovs and Richard Zemel
    ICML-2009: International Conference on Machine Learning
    [pdf]
  • ConEx: A System for Monitoring Queries
    Chaitanya Mishra and Maksims Volkovs
    SIGMOD-2007: International Conference on Management of Data
    [pdf]