Presented By O’Reilly and Intel AI
Put AI to Work
April 29-30, 2018: Training
April 30-May 2, 2018: Tutorials & Conference
New York, NY

Online and active learning for recommender systems

Jorge Silva (SAS)
4:50pm–5:30pm Wednesday, May 2, 2018
Implementing AI, Models and Methods
Location: Concourse A
Average rating: ****.
(4.00, 2 ratings)

Who is this presentation for?

  • Data scientists

Prerequisite knowledge

  • Familiarity with basic linear algebra, probabilities, and Bayes's law
  • A basic understanding of matrix factorization methods in recommendation (useful but not required)

What you'll learn

  • Learn how a probabilistic formulation is naturally suited for both online and active learning, due to the sequential nature of Bayesian updates and to the quantification of uncertainty

Description

Recommender systems require the ability to deal with high velocity of data and strong concept drift, since the users express their preference sequentially, and their preference changes rapidly and often.

Jorge Silva explains how SAS uses a Bayesian variant of the highly successful factorization machine model for recommender systems, in which the company employs a probabilistic model and performs variational Bayesian inference, to tackle both problems by making the learning process online and active. SAS leverages the model to actively select the best subset of observations at each step of the online learning process. This active learning procedure is motivated by the fact that each measurement is often obtained sequentially and requires user interaction, which may be expensive and time consuming. Prioritizing the most informative users and items potentially helps guide the rating acquisition process in a more efficient way.

In general, the problem of finding the exact best ratings to acquire is NP-complete. SAS has avoided this difficulty by developing a greedy algorithm based on the estimated variance of the user and item factors returned by the Bayesian inference. Experiments on benchmark datasets show highly promising results.

Photo of Jorge Silva

Jorge Silva

SAS

Jorge Silva is a principal machine learning developer at SAS. Previously, he was an adjunct professor at Instituto Superior de Engenharia de Lisboa (ISEL) and a senior research scientist at Duke University. His research interests include statistical models applied to large-scale problems, such as manifold learning, computer vision, and recommender systems. He holds multiple US patents and has authored numerous scholarly papers. Jorge holds a PhD in electrical and computer engineering from Instituto Superior T├ęcnico (IST), Lisbon.