Simon Chan - PredictionIO
eBay Whitman Campus
2065 Hamilton Ave
San Jose, CA
6:30 Food & Networking
With the advancement of Machine Learning, software applications can now learn from Big Data efficiently and perform all kinds of intelligent tasks. One of the biggest challenges for engineers in building real world predictive applications is the steep learning curve of mastering multiple data-processing frameworks, algorithms and scalable systems
This talk will begin with an overview of how Machine Learning, and user preference prediction in particular, is empowering our world in many aspects, and why it will play a key role in our future daily life. I will then discuss some core Machine Learning concepts such as Collaborative Filtering techniques, hyper-parameter optimization and scientific evaluation metrics. I will demonstrate, with live sample codes, the use of open-source technologies such as Hadoop, Cascading, Scalding and PredictionIO in adding predictive features such as personalization, recommendation and content discovery to an application. You will also see how PredictionIO's user-friendly control interfaces can evaluate, compare, select and deploy learning algorithms; tune hyperparameters of algorithms manually or automatically; and review the predictive model training status. By the end of the talk, you will have mastered the whole process of predictive modeling, and will be able to apply scalable algorithms to a real software production environment.
Simon Chan is a co-founder of PredictionIO. After a brief spell as a software engineer in the Bay Area after graduating from university, Simon became an entrepreneur and founded three startups in the past 10 years. He specializes in machine learning and recommendation technology, and has a strong interest in social applications. He graduated from the University of Michigan, Ann Arbor with a degree in Computer Science. While being an entrepreneur, he is also a PhD candidate in Machine Learning at University College London.
Event page provided by ACM