Lo que hacemos
Próximos eventos (3)
*Note, the date is TBD... more details coming soon. In this talk, you'll hear about several important practical aspects of training LTR models that one needs to know in order to bring the business impact through changes in LTR algorithms. Anjan Goswami will discuss important lessons of training LTR models for real applications and will also speak about some specific issues that needs to be addressed in e-commerce or employee and job search platforms. Prerequisites: Intermediate What you should bring: No computers needed Meet Your Speaker: Anjan Goswami heads the search science and engineering team at walmart labs. His team is responsible for relevance and ranking of Walmart.com e-commerce engine, its mobile site and other Walmart properties. In past, Anjan held various leadership positions in Elance-odesk, eBay, Amazon and in Microsoft working in the juncture of applied science and business. He is also pursuing a PhD in Computer Science with specialization in machine Learning and information retrieval from UC Davis.
*Note: date is TBD... more details coming soon. Querying large volumes of data at scale in a way optimized for data visualizations is a difficult problem. In this talk, Vadim Ogievetsky presents one solution to this problem. What's this talk about? Database APIs often mirror the architecture and limitations of the database. The developer of an application must bridge the logical chasm between the visualization model and the underlying API. As a result, fragile, one-off, glue code is often written to extract data in the right format for visualization. We will discuss the ideas behind Plywood, an open source library that describes data queries in terms of the UIs and visualizations that will be powered by the queries. Plywood is an ORM-like layer for Druid and other OLAP stores, and is built on top of Hadley Wickham’s split-apply-combine principle. Plywood facilitates the development of highly interactive UIs with rich visualization components. Meet Your Speaker: Vadim Ogievetsky is a co-founder of a stealth startup where he uses Plywood with React to build open source UIs on top of Druid. Prior to founding his own company, he worked as the UI lead at Metamarkets. In 2011, he graduated from Stanford University with a Master’s degree in Computer Science. While at Stanford, he was part of the Data Visualization group where he contributed to Protovis and D3.js. His open-source development is now focused on Plywood and other big data visualization tools.
*Date is currently TBD... coming soon. In this presentation, you'll learn about how remote sensing data can be incorporated for machine learning applications. Data standardization is one of the most critical aspects for accurately analyzing Earth observation data using statistics. This presentation outlines UrtheCast’s framework for delivering multi-source, standardized Earth observation imagery for use in traditional and non-traditional analysis. Prerequisites: This is focused for intermediate users. Meet Your Speaker: William Parkinson is a remote sensing data scientist for Urthecast, an earth observation technology company. His work includes applying and adapting machine learning algorithms for information extraction from earth observation data products. Previously, he developed advanced statistical techniques for monitoring natural resources and food security for the Canadian Government. His key area of interest is in understanding the spatial context of data and how it affects predictive analytics.