Yehuda Koren’s research while affiliated with Google Inc. and other places

A common task of recommender systems is to improve customer experience through personalized recommenda- tions based on prior implicit feedback. These systems pas- sively track different sorts of user behavior, such as pur- chase history, watching habits and browsing activity, in or- der to model user preferences. Unlike the much more ex- tensively researched explicit feedback, we do not have any direct input from the users regarding their preferences. In particular, we lack substantial evidence on which products consumer dislike. In this work we identify unique proper- ties of implicit feedback datasets. We propose treating the data as indication of positive and negative preference asso- ciated with vastly varying confidence levels. This leads to a factor model which is especially tailored for implicit feed- back recommenders. We also suggest a scalable optimiza- tion procedure, which scales linearly with the data size. The algorithmis used successfully within a recommender system for television shows. It compares favorably with well tuned implementations of other known methods. In addition, we offer a novel way to give explanations to recommendations given by this factor model.

We introduce a new force-directed model for computing graph layout. The model bridges the two more popular force directed approaches – the stress and the electrical-spring models – through the binary stress cost function, which is a carefully defined energy function with low descriptive complexity allowing fast computation via a Barnes-Hut scheme. This allows us to overcome optimization pitfalls from which previous methods suffer. In addition, the binary stress model often offers a unique viewpoint to the graph, which can occasionally add useful insight to its topology. The model uniformly spreads the nodes within a circle. This helps in achieving an efficient utilization of the drawing area. Moreover, the ability to uniformly spread nodes regardless of topology, becomes particularly helpful for graphs with low connectivity, or even with multiple connected components, where there is not enough structure for defining a readable layout.

Recommender systems provide users with personalized suggestions for products or services. These systems often rely on Collaborating Filtering (CF), where past transactions are analyzed in order to establish connections between users and products. The two more successful approaches to CF are latent factor models, which directly profile both users and products, and neighborhood models, which analyze similarities between products or users. In this work we introduce some innovations to both approaches. The factor and neighborhood models can now be smoothly merged, thereby building a more accurate combined model. Further accuracy improvements are achieved by extending the models to exploit both explicit and implicit feedback by the users. The methods are tested on the Netflix data. Results are better than those previously published on that dataset. In addition, we suggest a new evaluation metric, which highlights the differences among methods, based on their performance at a top-K recommendation task.

In many graph mining settings, measuring node proximity is a fundamental problem. While most of existing measurements are (implicitly or explicitly) designed for undirected graphs; edge directions in the graph provide a new perspective to proximity measurement: measuring the proximity from A to B; rather than between A and B. (See Figure 1 as an example). In this chapter, we study the role of edge direction in measuring proximity on graphs. To be specific, we will address the following fundamental research questions in the context of direction-aware proximity: 1. Problem definitions: How to define a directionaware proximity? 2. Computational issues: How to compute the proximity score efficiently? 3. Applications: How can direction-aware proximity benefit graph mining?

This article outlines the overall strategy and summarizes a few key innovations of the team that won the first Netflix progress prize.

Measuring distance or some other form of proximity between objects is a standard data mining tool. Connection subgraphs were recently proposed as a way to demonstrate proximity between nodes in networks. We propose a new way of measuring and extracting proximity in networks called "cycle- free effective conductance" (CFEC). Importantly, the measured proximity is accompanied with a proximity subgraph which allows assessing and understanding measured values. Our proximity calculation can handle more than two endpoints, directed edges, is statistically well behaved, and produces an effectiveness score for the computed subgraphs. We provide an efficient algorithm to measure and extract proximity. Also, we report experimental results and show examples for four large network datasets: a telecommunications calling graph, the IMDB actors graph, an academic coauthorship network, and a movie recommendation system.

Recommender systems based on collaborative filtering predict user preferences for products or services by learning past user-item relationships. A predominant approach to collaborative filtering is neighborhood based ("k-nearest neighbors"), where a user-item preference rating is interpolated from ratings of similar items and/or users. We enhance the neighborhood-based approach leading to substantial improvement of prediction accuracy, without a meaningful increase in running time. First, we remove certain so-called "global effects" from the data to make the ratings more comparable, thereby improving interpolation accuracy. Second, we show how to simultaneously derive interpolation weights for all nearest neighbors, unlike previous approaches where each weight is computed separately. By globally solving a suitable optimization problem, this simultaneous interpolation accounts for the many interactions between neighbors leading to improved accuracy. Our method is very fast in practice, generating a prediction in about 0.2 milliseconds. Importantly, it does not require training many parameters or a lengthy preprocessing, making it very practical for large scale applications. Finally, we show how to apply these methods to the perceivably much slower user-oriented approach. To this end, we suggest a novel scheme for low dimensional embedding of the users. We evaluate these methods on the netflix dataset, where they deliver significantly better results than the commercial netflix cinematch recommender system.

Recommender systems based on collaborative filtering predict user preferences for products or services by learning past user-item re-lationships. A predominant approach to collaborative filtering is neighborhood based ("k-nearest neighbors"), where a user-item pref-erence rating is interpolated from ratings of similar items and/or users. In this work, we enhance the neighborhood-based approach leading to a substantial improvement of prediction accuracy, with-out a meaningful increase in running time. First, we remove certain so-called "global effects" from the data to make the different ratings more comparable, thereby improving interpolation accuracy. Sec-ond, we show how to simultaneously derive interpolation weights for all nearest neighbors. Unlike previous approaches where each interpolation weight is computed separately, simultaneous interpo-lation accounts for the many interactions between neighbors by globally solving a suitable optimization problem, also leading to improved accuracy. Our method is very fast in practice, generat-ing a prediction in about 0.2 milliseconds. Importantly, it does not require training many parameters or a lengthy preprocessing, mak-ing it very practical for large scale applications. The method was evaluated on the Netflix dataset. We could process the 2.8 million queries of the Qualifying set in 10 minutes yielding a RMSE of 0.9086. Moreover, when an extensive training is allowed, such as SVD-factorization at the preprocessing stage, our method can pro-duce results with a RMSE of 0.8982.

In this paper we study asymmetric proximity measures on directed graphs, which quantify the relationships between two nodes or two groups of nodes. The measures are useful in several graph min- ing tasks, including clustering, link prediction and connection sub- graph discovery. Our proximity measure is based on the concept of escape probability. This way, we strive to summarize the mul- tiple facets of nodes-proximity, while avoiding some of the pit- falls to which alternative proximity measures are susceptible. A unique feature of the measures is accounting for the underlying directional information. We put a special emphasis on computa- tional efficiency, and develop fast solutions that are applicable in several settings. Our experimental study shows the usefulness of our proposed direction-aware proximity method for several appli- cations, and that our algorithms achieve a significant speedup (up to 50,000x) over straightforward implementations.