Data platforms and analytics
Project
Algorithmic Crowdsourcing
To build a machine learning based intelligent system, we often need to collect training labels and feed them into the system. A useful lesson in machine learning is that “more data beats a clever algorithm”.…
Project
Trill
Trill is a high-performance open-source in-memory incremental analytics library. It can handle both real-time and offline data, and is based on a temporal data and query model. Trill can be used as a streaming engine,…
Publication
Placer: Semantic Place Labels from Diary Data
Publication