ITU AI/ML in 5G Challenge: DNN Inference Optimization Challenge
The talk will describe the problem ITU-ML5G-PS-018 DNN Inference Optimization. This problem is about how to optimize inference efficiency of deep learning models since computing efficiency, memory footprint and inference latency tends to be the bottleneck when deploying large deep learning models. The problem will be described in details in the presentation, including why we propose this problem, status quo of related research, what are expected from the competitors/participants and some suggestions will be given. Hopefully more competitors will get intested in this problem and propose their innovative solutions.
Speakers, Panelists and Moderators
LIYA YUANOpen Source & Standardization EngineerZTELiya Yuan is an open source & standardization engineer in ZTE, focusing on the field of AI for telecommunication network. She is currently the Technical Steering Committee (TSC) chair of AI project Adlik, which is incubated in Linux Foundation AI Foundation. She also takes an active part in ITU-T standardization activities and is the main editor of ITU-T FG ML5G specifications on "Serving framework for ML models in future networks including IMT-2020".