Join us for an evening of talks, discussion, and networking on the topic of industrial-scale data and machine learning. Learn how teams at companies such as LinkedIn and Google leverage large, rich datasets, and apply the latest in machine learning techniques to build useful applications - from improving UX of engaging in online discussions to inferring new relationships in LinkedIn’s economic graph and more.
The evening will be structured around 3 talks, with opportunities for Q&A and networking. Food and drink provided!
1. "Inferring Enterprise Relationships from Professional Social Networking Data and Beyond"
Abstract: Events and relationships among enterprises can provide valuable insights for understanding a company’s need in marketing, sales and recruiting. However, collecting and compiling this type of data at scale are often hard: there is no centralized authority (except publicly-traded companies which comprises a small fraction of enterprise entities) to mandate enterprises to report an event resulting in such changes. On the other hand, we observe that professional social network data contains rich information about such events and relationships, e.g., statements about one company acquiring another, or one company being a parent or subsidiary of another company. We start off with mining the titles of work experience text on LinkedIn and are able to harvest millions of M&A instances. We then align these instances with natural language text, through which a large labeled data set is curated. We then build a machine-learning pipeline to extract fresh M&A instances from news articles.
Speaker Bio: Xiaoqiang Luo
Xiaoqiang Luo is a Senior Engineer Manager and leads the data mining and machine learning efforts in the LinkedIn's NYC office. The team's mission is to utilize structured and unstructured data towards constructing LinkedIn's Economic Graph (https://www.linkedin.com/company/linkedin-economic-graph), and to help deliver the right information to the right individuals and enterprise at the right time. Before joining LinkedIn in 2015, he led a team of engineers and computational linguists to build the multi-lingual semantic understanding components at Google NYC that power many products with billions of users, such as Google Assistant, Photo search, and Local search. He received his Ph.D. in Electrical Engineering from The Johns Hopkins University.
LinkedIn profile: https://www.linkedin.com/in/xqluo/.
2. "ML for good conversations at scale" https://jigsaw.google.com/
Abstract: Leveraging machine learning as a tool for better online discussions at scale. Using the Perspective API as a case study, this talk explore challenges and opportunities related to building scaled ML systems that help humans with subjective classification tasks (e.g. identifying toxic language). It will cover learnings on ML transparency and fairness, and examples of new human-centered product experiences using the outputs of a ML system (e.g. improving the UX of writing, moderating, and reading online discussions).
CJ Adams is a Product Manager at Jigsaw, a team at Alphabet (Google) focused on building technology to help people facing emerging digital threats around the world. Currently he is the PM on Perspective, an API that aims to use ML as a tool to increase the participation, quality and empathy of online discussions. Previously he was the PM on Project Shield, a free DDoS mitigation service for news organizations.
LinkedIn Profile: https://www.linkedin.com/in/adamscj/
3. "Trajectories of Health: Insights from Electronic Health Records and Beyond"
Bio: Yiye Zhang, Ph.D., MS is Assistant Professor at Weill Cornell Medicine, Cornell University. She received her Ph.D. in Information Systems Management at Carnegie Mellon University. Her research is concerned with developing health informatics methods and tools to assist clinical decision making for patients and caregivers.