LinkedIn Knowledge Graph
LinkedIn’s knowledge graph is a large knowledge base built upon “entities” of LinkedIn, such as members, jobs, titles, skills, companies, geographical locations, schools, etc. These entities (~1B) and the relationships among them (~50B) form the ontology of the professional world and are used by LinkedIn to enhance its recommender systems, search, monetization/consumer products, and business/consumer analytics.
In this talk, I will give a data overview of the knowledge graph, how we embed the knowledge graph for relevance using deep neural network, how we serve the data to 200+ LinkedIn clients through a near-time content processing system, and how we bring values to our 500M members through LinkedIn products like ads targeting, people search, job recommendation, and inferred profile etc.
Dr. Qi He is the Head of LinkedIn Data Standardization. He is leading a team of 20 machine learning scientists and software engineers to help LinkedIn realize its vision of creating economic opportunities through building the LinkedIn Knowledge Graph. Before that, he managed LinkedIn Feed Relevance team, with a focus on developing and deploying personalized machine learning and statistical methods for improving the relevance of LinkedIn Feed. Prior to LinkedIn, he was a Research Staff Member at IBM Almaden Research until 2013. He completed two years of postdoctoral work on citation recommendations in CiteSeer at PSU from 2008 to 2010 and completed his PhD in information retrieval at NTU with a Microsoft Research Fellowship from 2005 to 2008. He was the General Chair of CIKM 2013, serves as Associate Editor of TKDE and Neurocomputing, and regularly serves on the program committee of SIGKDD, WWW, SIGIR, CIKM and WSDM. He serves the Program Chair of CIKM 2019. He received the 2008 SIGKDD Best Application Paper Award and the 2014 SDM Best Research Paper Runner Up award. He published about 50 papers with 3500+ citations in top-tier conferences and journals.
How hires are made: the economist’s view
More and better hiring means greater employment, greater productivity, and stronger economic growth. From the economist’s perspective, hires are the result of a long chain. It starts with education and the building of human capital. Then, people engage in job search using online platforms, and policies like unemployment insurance can influence this process. At the same time, companies use an array of strategies to attract the right workers. This matching process between people and firms can be hindered by geographic or skill mismatch. What matters for efficient outcomes at the different stages in hiring? How can online tools, algorithms and big data facilitate the matching process, and increase employment and productivity in the economy? We will learn about the crucial role of job titles in the matching process. While geographic mismatch is not a big problem, skills mismatch may be more important and could be addressed through better search & recommendation algorithms for students, job seekers, and employers.
Ioana Marinescu is a labor economist, an assistant professor at the University of
Pennsylvania School of Social Policy & Practice, and a member of the National Bureau of Economic Research. She has trained at Harvard University and the London School of Economics. Ioana’s current research focuses on the dynamics of matching and search in the labor market, and how matching mechanisms determine unemployment and productivity. She has been working with “big data” from CareerBuilder.com to better understand employers’ and job seekers’ online search behavior. She has also worked on the education to work pathway by documenting the relationship between salary and major choice among community college students.
Applying Data Science for HCM in the modern enterprise: a journey, lessons learnt, and pitfalls to avoid
Organizations are increasingly turning to data to help acquire, maintain, and retain critical talent to stay ahead of the competition in the rapidly changing economic landscape. This is especially true for large enterprises that often have complex, geographically diverse workforces. I briefly describe a decade long journey in applying data science to problems in Human Capital Management and Workforce Analytics, including employee retention, talent management, re-skilling, expertise location, forecasting and skill planning, and workforce evolution. I then delve deeper into some of these initiatives to highlight some key lessons learnt and main pitfalls to avoid when applying data science to address various problems in the HCM domain in an enterprise.
Moninder Singh joined IBM Research in 1998 after receiving his Ph.D. in Computer and Information Science from the University of Pennsylvania. He is a member of the Data Science Group and is primarily interested in developing and deploying solutions for interesting problems in business analytics and decision support. His main areas of interest are machine learning and data mining, data privacy, information retrieval, probabilistic modeling and reasoning, and text mining. Over the past several years, he has led several highly successful workforce analytics and human capital management initiatives for IBM and its clients.