"Harnessing Heterogeneous Computing for Cloud- and Mobile-Based Visual Search," A Presentation From Baidu
Dr. Ren Wu, Distinguished Scientist at Baidu, delivers the presentation "Harnessing Heterogeneous Computing for Cloud- and Mobile-Based Visual Search" at the September 2014 Embedded Vision Alliance Member Meeting. As China's leading search engine, Baidu's focus is not only on indexing website text but also multimedia content, such as audio files, still images and video clips. It's in these latter areas where computer vision fits in, enabling automated analysis and identification of the various objects embedded within the visual information.
Dr. Ren Wu is a distinguished scientist at Baidu's Institute of Deep Learning (IDL), well known for his pioneering research in using GPUs, FPGAs and other coprocessors to accelerate "big data" analytics, as well as for his contributions to large-scale clustering algorithms via GPUs. Dr. Wu has recently won recognition for his work on Deep Image, Baidu’s image recognition system based on deep learning techniques using convolutional neural networks. In January 2015, Baidu announced that Deep Image had bested Google on the ImageNet benchmark. Prior to joining Baidu, Dr. Wu served as the chief software architect of Heterogeneous System Architecture (HSA) at AMD. Earlier, he was a senior scientist at HP Labs and the principal researcher at HP Lab's CUDA Research Center.