Deep learning has been widely successful in solving complex tasks such as image recognition (ImageNet), speech recognition, machine translation, etc. In the area of personalized recommender systems, deep learning has started showing promising advances in recent years. The key to success of deep learning in personalized recommender systems is its ability to learn distributed representations of users’ and items’ attributes in low dimensional dense vector space and combine these to recommend relevant items to users. To address scalability, the implementation of a recommendation system at web scale often leverages components from information retrieval systems, such as inverted indexes where a query is constructed from a user’s attribute and context, learning to rank techniques. Additionally, it relies on machine learning models to predict the relevance of items, such as collaborative filtering. In this tutorial, we present ways to leverage deep learning towards improving recommender system. The tutorial is divided into four parts: (1) In the first part, we will present an overview of concepts in deep learning which are pertinent to recommender systems including sequence modeling, word embedding and named entity recognition. (2) In the second part, we will present how these fundamental building blocks can be used to improve a recommender system at scale. (3) The third part presents a few case studies from large scale recommender systems at LinkedIn and some of the challenges we faced while getting deep learning to work in production.
Nadia Fawaz is a Staff Software Engineer in Machine Learning/Data Mining at LinkedIn, CA working in the Data Relevance Group. She is technical lead for the job recommendation team, where she oversees deep learning projects. Her research and engineering interests include machine learning for personalization, and data privacy. Her work leverages techniques from AI including machine learning and deep learning, information theory, random matrix theory, statistics and privacy theory, and aims at bridging theory and practice. She was a winner of the ACM RecSyS challenge on Context-Aware Movie Recommendations CAMRa2011, and her 2012 UAI paper "Guess Who Rated This Movie: Identifying Users Through Subspace Clustering" was featured in an MIT TechReview article as “The Ultimate Challenge For Recommendation Engines”. From 2011 to 2016, she was a principal research scientist at Technicolor research center in Los Altos, CA. From 2009 to 2011, she was a postdoctoral researcher at the Massachusetts Institute of Technology (MIT) Research Laboratory of Electronics (RLE), Cambridge, MA. She received her Ph.D. degree in 2008 and her Diplome d'ingenieur (M.Sc.) in 2005 both in electrical engineering, from Ecole Nationale Superieure des Telecommunications de Paris and EURECOM, France. She is a Member of IEEE and of ACM. Web: http://nadiafawaz.com
Saurabh Kataria currently works on application of deep learning to personalized job search and recommendations at LinkedIn. He holds a Ph.D. from Penn State University where he developed a citation recommendation system for CiteSeerX digital Library. Prior to LinkedIn, he worked at Palo Alto Research Center (PARC) as a research scientist where he worked on representation learning for medical concepts for various predictive tasks in healthcare domain such as hospital readmission, risk stratification, comorbidity prediction. He has published more than 15 research articles with 300+ citations on various data mining and machine learning venues including SIGKDD, AAAI, IJCAI, and CIKM. He has served as technical program committee of various academic conferences and workshops such as Association for Computational Linguistics (ACL), Empirical Methods in Natural Language Processing (EMNLP), AAAI, IJCAI, International Conference on Computational Linguistics (COLING). Web: http://ktsaurabh.weebly.com/
Benjamin Le is a Senior Engineer at Linkedin working on core relevance in both personalized job search and job recommendations. His contributions on both relevance stacks include candidate selection in job search through multipass ranking, multi-objective optimization of the trade off between member/enterprise customer satisfaction in job search rankings, and model engineering for bad job recommendation elimination. Currently, he is focusing on efforts in the deployment of deep learning models in job recommendations. He graduated from University of California, Berkeley, where he received both his Master of Engineering and Bachelor of Science Degree. For his Master’s Industry Capstone Project, he worked closely with Conviva, a leader in leveraging QoE (Quality of Experience) analytics in online video optimization. During that Capstone Project, he developed anomaly detection algorithms that can be run on real-time data streams to quickly detect, with high confidence, potential service degradation on partner online video platforms.
Ganesh Venkataraman is an Engineering Manager at AirBnB. He used to lead all AI powering jobs relevance at LinkedIn. This includes personalized job search, recommendations and statistical inference and insights from salary. His contributions at LinkedIn include, leading a multi group effort that led to 50% improvement in job applications at LinkedIn, leading end to end re-architecture of job search, machine learned ranking for people search typeahead (system that allows members to search for 400MM+ users via instant results), introducing machine learned ranking towards skills search at LinkedIn (ex: searching for people skilled at ‘information retrieval’). He co-authored a paper on personalized ranking which won the best paper award at the IEEE Big Data Conference 2015. Prior to LinkedIn he was the founding engineer of a payments startup where he developed algorithms to detect/prevent eCommerce fraud. He holds a Ph.D. from Texas A&M in Electrical & Computer Engineering where he was the recipient of the Dean’s graduate merit scholarship. He has several publications with 200+ citations including one fundamental contribution to graph theory.
Liang Zhang is currently a Sr. Staff Engineer at LinkedIn. He obtained his Ph. D. degree at Department of Statistical Science, Duke University in 2008. He worked at Yahoo! Inc. as a Scientist from 2008 to March 2012. Liang has published several papers and patents on applying statistical and machine learning approaches to real world Internet applications involving massive data. His research interests include recommender systems, computational advertising, machine learning, statistical modeling and analysis for large-scale data. Liang has served Program Committee for various data mining and machine learning venues including SIGKDD, NIPS, AISTAT, TKDD, TKDE, JMLR.