Advancement: Improving Deep Learning Performance and Privacy Through Distributed Systems

Samaa Gazzaz
Computer Science & Engineering PhD Student
Virtual Event
Katia Obraczka

Join us on Zoom: / Passcode: 111

Description: Emerging edge applications require both a fast response latency and complex processing. This is infeasible without expensive hardware that can process complex operations—such as object detection—within a short time. Many approach this problem by addressing the complexity of the models—via model compression, pruning, and quantization—or compressing the input. In this work, we propose a different perspective when addressing the performance challenges. Croesus is a multistage framework for edge-cloud systems that enables dynamically finding the balance between accuracy and performance. Croesus consists of two stages: an initial and a final stage. The initial stage performs the computation in real-time using approximate/best-effort computation at the edge. The final stage performs the full computation in the cloud and uses the results to correct any errors made at the initial stage. In this work, we demonstrate the implications of such an approach on a video analytics use case and show how multi-stage processing yields a better balance between accuracy and performance. Moreover, we study the safety of multi-stage transactions via two proposals: multi-stage serializability (MS-SR) and multi-stage invariant confluence with Apologies (MS-IA).