DIANNE: Distributed Artificial Neural Networks for the Internet of Things

摘要: Nowadays artificial neural networks are widely used to accurately classify and recognize patterns. An interesting application area is the Internet of Things (IoT), where physical things connected Internet, generate a huge amount sensor data that can be for myriad new, pervasive applications. Neural networks' ability comprehend unstructured make them useful building block such IoT As require lot processing power, especially during training phase, these most often deployed in cloud environment, or on specialized servers with dedicated GPU hardware. However, applications, sending all raw remote back-end might not feasible, taking into account high variable latency cloud, could lead issues concerning privacy. In this paper DIANNE middleware framework presented optimized single sample feed-forward execution facilitates distributing across multiple devices. The modular approach enables executing network components large number heterogeneous devices, allowing us exploit local compute power at hand, mitigating need server-side infrastructure runtime.