Samsung

2015-09 to 2018-06

Research Manager & Senior Staff Research Scientist

Lead deep learning & computer vision research and engineering team at AI R&D center of SRA and built the whole deep learning and computer vision team from the beginning in limited time. My team has developed a set of advanced prototypes based on real-time deep learning inference on device technology with focusing on CV domain. 18 industrial patents and 12 academic papers have been submitted. These papers and patents cover deep model compression (CNN, LSTM and RNN) and acceleration on mobile devices, action classification, real-time facial image analysis (detection, verification and liveness detection), automatic speech recognition(ASR), natural language processing (NLP), real time lipreading on device, speak activity detection on device, dynamic video frames generation, semantic segmentation, deep learning mobile system design and image captioning. One recent work: DeepRebirth: Accelerating Deep Neural Network Execution on Mobile Devices has been reported by several popular Medias including esestech.com, sohu.com, sina.com, ifeng.com, etc. (The only public reported research work of SRA since 2015.)