Track: Internet of Things |
| Artificial Intelligence for Internet of Things: End-to-end software stack to consume and capitalize on IoT enabling state-of-the-art AI |
| Applications for Internet of Things (IoT) ecosystems are everywhere—agriculture, manufacturing, healthcare, retail, supply chain, transportation and many more. Speed and efficiency are paramount in such a context, not only to process but also to act on IoT data. Traditionally, IoT-like systems leverage propositional logic from various components to act on the data received. While this approach is effective, it can take a great deal of time and engineering effort to develop systems that can act on such data. We will discuss how to leverage state-of-the-art data science components using REFIT’s architecture to consume streaming data from ingestion, to training and inference, and visualization of AI models. REFIT enables seamless integration of the data science darling python and computationally efficient Scala. AI processes such as data engineering need to be written only once in python and then they are propagated to training and inference. REFIT allows data injection through low latency streaming; the streamed data is stored in a high throughput persistent storage database which is augmented with other static data sources to train predictive AI models. The current trained model is pushed to the main processing unit that conducts predictions on real time data. The various data streams and predictions are stored in a persistent storage and can be visualized in real time in a dashboard. REFIT’s data streaming architecture ensures the seamless deployment of new predictive models to the stream processing pipeline. We present in detail the use cases of monitoring and predicting traffic congestion and estimating fine-granular demand for bikes of a bike-sharing company. |
|
| Presentation Video |
