Matching Users and Items Across Domains to Improve the Recommendation Quality Created by: Chung-Yi Li, Shou-De Lin Presented by: I Gde Dharma Nugraha 1.

Slides:



Advertisements
Similar presentations
A Robust Super Resolution Method for Images of 3D Scenes Pablo L. Sala Department of Computer Science University of Toronto.
Advertisements

Elementary Linear Algebra Anton & Rorres, 9th Edition
C ONTEXT - AWARE SIMILARITIES WITHIN THE FACTORIZATION FRAMEWORK Balázs Hidasi Domonkos Tikk C A RR WORKSHOP, 5 TH F EBRUARY 2013, R OME.
Tensors and Component Analysis Musawir Ali. Tensor: Generalization of an n-dimensional array Vector: order-1 tensor Matrix: order-2 tensor Order-3 tensor.
Efficient Retrieval of Recommendations in a Matrix Factorization Framework Noam KoenigsteinParikshit RamYuval Shavitt School of Electrical Engineering Tel.
Algebraic, transcendental (i.e., involving trigonometric and exponential functions), ordinary differential equations, or partial differential equations...
CmpE 104 SOFTWARE STATISTICAL TOOLS & METHODS MEASURING & ESTIMATING SOFTWARE SIZE AND RESOURCE & SCHEDULE ESTIMATING.
Principal Component Analysis (PCA) for Clustering Gene Expression Data K. Y. Yeung and W. L. Ruzzo.
MATH 685/ CSI 700/ OR 682 Lecture Notes
Time-domain Crosstalk The equations that describe crosstalk in time-domain are derived from those obtained in the frequency-domain under the following.
What is missing? Reasons that ideal effectiveness hard to achieve: 1. Users’ inability to describe queries precisely. 2. Document representation loses.
I NCREMENTAL S INGULAR V ALUE D ECOMPOSITION A LGORITHMS FOR H IGHLY S CALABLE R ECOMMENDER S YSTEMS (S ARWAR ET AL ) Presented by Sameer Saproo.
Simple Neural Nets For Pattern Classification
Some useful linear algebra. Linearly independent vectors span(V): span of vector space V is all linear combinations of vectors v i, i.e.
Presented by Li-Tal Mashiach Learning to Rank: A Machine Learning Approach to Static Ranking Algorithms for Large Data Sets Student Symposium.
Epipolar geometry. (i)Correspondence geometry: Given an image point x in the first view, how does this constrain the position of the corresponding point.
Information Retrieval in Text Part III Reference: Michael W. Berry and Murray Browne. Understanding Search Engines: Mathematical Modeling and Text Retrieval.
Malicious parties may employ (a) structure-based or (b) label-based attacks to re-identify users and thus learn sensitive information about their rating.
A Constraint Generation Approach to Learning Stable Linear Dynamical Systems Sajid M. Siddiqi Byron Boots Geoffrey J. Gordon Carnegie Mellon University.
Camera Calibration CS485/685 Computer Vision Prof. Bebis.
Do Supervised Distributional Methods Really Learn Lexical Inference Relations? Omer Levy Ido Dagan Bar-Ilan University Israel Steffen Remus Chris Biemann.
On Comparing Classifiers: Pitfalls to Avoid and Recommended Approach Published by Steven L. Salzberg Presented by Prakash Tilwani MACS 598 April 25 th.
Performance of Recommender Algorithms on Top-N Recommendation Tasks
Advanced Computer Graphics Spring 2014 K. H. Ko School of Mechatronics Gwangju Institute of Science and Technology.
Square n-by-n Matrix.
MATH 250 Linear Equations and Matrices
CALCULUS – II Gauss Elimination
Performance of Recommender Algorithms on Top-N Recommendation Tasks RecSys 2010 Intelligent Database Systems Lab. School of Computer Science & Engineering.
WEMAREC: Accurate and Scalable Recommendation through Weighted and Ensemble Matrix Approximation Chao Chen ⨳ , Dongsheng Li
Yan Yan, Mingkui Tan, Ivor W. Tsang, Yi Yang,
Algorithms for a large sparse nonlinear eigenvalue problem Yusaku Yamamoto Dept. of Computational Science & Engineering Nagoya University.
Focused Matrix Factorization for Audience Selection in Display Advertising BHARGAV KANAGAL, AMR AHMED, SANDEEP PANDEY, VANJA JOSIFOVSKI, LLUIS GARCIA-PUEYO,
A Simple Unsupervised Query Categorizer for Web Search Engines Prashant Ullegaddi and Vasudeva Varma Search and Information Extraction Lab Language Technologies.
EMIS 8381 – Spring Netflix and Your Next Movie Night Nonlinear Programming Ron Andrews EMIS 8381.
Classifier Evaluation Vasileios Hatzivassiloglou University of Texas at Dallas.
WEEK 8 SYSTEMS OF EQUATIONS DETERMINANTS AND CRAMER’S RULE.
Online Learning for Collaborative Filtering
June 5, 2006University of Trento1 Latent Semantic Indexing for the Routing Problem Doctorate course “Web Information Retrieval” PhD Student Irina Veredina.
Badrul M. Sarwar, George Karypis, Joseph A. Konstan, and John T. Riedl
SINGULAR VALUE DECOMPOSITION (SVD)
Elementary Linear Algebra Anton & Rorres, 9 th Edition Lecture Set – 02 Chapter 2: Determinants.
MATH 685/ CSI 700/ OR 682 Lecture Notes Lecture 4. Least squares.
CROSS-VALIDATION AND MODEL SELECTION Many Slides are from: Dr. Thomas Jensen -Expedia.com and Prof. Olga Veksler - CS Learning and Computer Vision.
Geodesic Flow Kernel for Unsupervised Domain Adaptation Boqing Gong University of Southern California Joint work with Yuan Shi, Fei Sha, and Kristen Grauman.
Section 5.1 First-Order Systems & Applications
Diagonalization and Similar Matrices In Section 4.2 we showed how to compute eigenpairs (,p) of a matrix A by determining the roots of the characteristic.
Temporal Diversity in Recommender Systems Neal Lathia, Stephen Hailes, Licia Capra, and Xavier Amatriain SIGIR 2010 April 6, 2011 Hyunwoo Kim.
Evaluation of Recommender Systems Joonseok Lee Georgia Institute of Technology 2011/04/12 1.
Pairwise Preference Regression for Cold-start Recommendation Speaker: Yuanshuai Sun
Amanda Lambert Jimmy Bobowski Shi Hui Lim Mentors: Brent Castle, Huijun Wang.
Unsupervised Auxiliary Visual Words Discovery for Large-Scale Image Object Retrieval Yin-Hsi Kuo1,2, Hsuan-Tien Lin 1, Wen-Huang Cheng 2, Yi-Hsuan Yang.
Model Fusion and its Use in Earth Sciences R. Romero, O. Ochoa, A. A. Velasco, and V. Kreinovich Joint Annual Meeting NSF Division of Human Resource Development.
Matrix Factorization and its applications By Zachary 16 th Nov, 2010.
Collaborative Filtering via Euclidean Embedding M. Khoshneshin and W. Street Proc. of ACM RecSys, pp , 2010.
Optimal Reverse Prediction: Linli Xu, Martha White and Dale Schuurmans ICML 2009, Best Overall Paper Honorable Mention A Unified Perspective on Supervised,
Identifying “Best Bet” Web Search Results by Mining Past User Behavior Author: Eugene Agichtein, Zijian Zheng (Microsoft Research) Source: KDD2006 Reporter:
Online Evolutionary Collaborative Filtering RECSYS 2010 Intelligent Database Systems Lab. School of Computer Science & Engineering Seoul National University.
Copyright © Cengage Learning. All rights reserved. 7.7 The Determinant of a Square Matrix.
Matrix Factorization & Singular Value Decomposition Bamshad Mobasher DePaul University Bamshad Mobasher DePaul University.
Section 2.1 Determinants by Cofactor Expansion. THE DETERMINANT Recall from algebra, that the function f (x) = x 2 is a function from the real numbers.
ECE 530 – Analysis Techniques for Large-Scale Electrical Systems Prof. Hao Zhu Dept. of Electrical and Computer Engineering University of Illinois at Urbana-Champaign.
Collaborative Filtering - Pooja Hegde. The Problem : OVERLOAD Too much stuff!!!! Too many books! Too many journals! Too many movies! Too much content!
Lecture XXVII. Orthonormal Bases and Projections Suppose that a set of vectors {x 1,…,x r } for a basis for some space S in R m space such that r  m.
Matrix Factorization and Collaborative Filtering
Linear Equations Gauss & Cramer’s
Singular Value Decomposition
The European Conference on e-learing ,2017/10
The use of Neural Networks to schedule flow-shop with dynamic job arrival ‘A Multi-Neural Network Learning for lot Sizing and Sequencing on a Flow-Shop’
Probabilistic Latent Preference Analysis
A Data Partitioning Scheme for Spatial Regression
Presentation transcript:

Matching Users and Items Across Domains to Improve the Recommendation Quality Created by: Chung-Yi Li, Shou-De Lin Presented by: I Gde Dharma Nugraha 1

Motivation  Lack of data is a serious concern in building a recommender system, in particular for newly established services.  Can we leverage the information from other domains to improve the quality of a recommender system? 2 2

Problem Definition Given: Two homogeneous rating matrices  They model the same type of preference.  Decent portion of overlap in users and in items. Target Rating Matrix ♫ ♫ ♫ Source Rating Matrix ♫ ♫ ♫ Challenge: The mapping of users is unknown, and so is the mapping of items. Goals: 1.Identify the user mapping and item mapping. 2.Use the identified mappings to boost the recommendation performance. 3 3

Why This Problem Is Challenging  When item correspondence is known, the problem is much easier  Define user similarity. If the similarity is large, they are likely to be the same users. [Narayanan 2008]  In our case, both sides are unknown  no clear solution yet ♫ ♫ ♫ 4 4

Basic Idea  low rank assumption and factorization models 5 R1R1 R2R2 n1n1 n2n2 n3n3 n4n4 m1m1 m2m2 m3m3 m4m4 m5m5 n4n4 n3n3 n2n2 n1n1 m5m5 m4m4 m3m3 m2m2 m1m1 = = n1n1 n2n2 n3n3 n4n4 m1m1 m2m2 m3m3 m4m4 m5m5 n4n4 n3n3 n2n2 n1n1 m5m5 m4m4 m3m3 m2m2 m1m1 ? ? 5

≈ M1×N1M1×N1 M2×N2M2×N2 M1×M2M1×M2 N2×N1N2×N1 A Two-Stage Model to Find the Matching ? O O ? ? O Rough Matching Result Final Matching Result 6 6

Stage 1: Latent Space Matching 1. Latent Space Matching 7 7

How can we perform SVD on a Partially Observed Matrix? 1. Latent Space Matching = = = 8 8

We want to solve G from Now we know how to get Thus Since SVD is unique, we can separate user and item sides: Matching in Latent Space Same subproblem S: sign matrix (K by K, diagonal, -1 or 1) 1. Latent Space Matching 9 9

Solving ≈ (M 1 × K) (M 1 × M 2 ) (M 2 × K) 1. Latent Space Matching

 More accurate but harder to solve.  Obtain good initialization and reduced search space from latent space matching.  Solve G user and G item alternatingly.  The objective value always decreases & converges. Rough Matching Result Final Matching Result 11

Goals 1.Identify the user mapping and item mapping 2.Then, use the identified mappings to boost recommendation performance Rough Matching Result Final Matching Result 12

 Matched latent factors are constrained to be similar Transferring Imperfect Matching to Predict Ratings 13

Experiment Setup Disjoint Split Overlap SplitContained Split Subset Split training set of R 1 training set of R 2 Partial Split users items Yahoo! Music Dataset 14

Accuracy and Mean Average Precision: The higher the better 15

Rating Prediction (Root Mean Square Error) RMSE: the lower the better 16

(root mean square error)

Conclusion  It is possible to identify user or item correspondence unsupervisedly based on homogeneous rating data  Even with imperfect matching, out model can still improve the recommendation accuracy.  Questions? 18 17