Baidu has a computer vision system that is better than Google and close to human accuracy
Baidu has a state -of-the-art image recognition system, Deep Image, developed using end-to-end deep learning. The key components are a custom-built supercomputer dedicated to deep learning, a highly optimized parallel algorithm using new strategies for data partitioning and communication, larger deep neural network models, novel data augmentation approaches, and usage of multiscale high-resolution images. On one of the most challenging computer vision benchmarks, the ImageNet classification challenge, our system has achieved the best result to date, with a top – 5 error rate of 5.98% -a relative 10.2% improvement over the previous best result.
Chinese search engine company Baidu says it has built the world’s most-accurate computer vision system, dubbed Deep Image, which runs on a supercomputer optimized for deep learning algorithms. Baidu’s 5.98 percent error rate on the ImageNet object classification benchmark is better than Google’s 2014 ImageNet competition winning 6.66 percent error rate.
In experiments, humans achieved an estimated error rate of 5.1 percent on the ImageNet dataset.
A sample of the effects Baidu used to augment images. Credit Baidu
SOURCES – Arxiv, Gigaom
Wang is a prolific business-oriented writer of emerging and disruptive technologies. He is known for insightful articles that combine business and technical analysis that catches the attention of the general public and is also useful for those in the industries. He is the sole author and writer of nextbigfuture.com
, the top online science blog. He is also involved in angel investing and raising funds for breakthrough technology startup companies.
He gave the recent keynote presentation at Monte Jade event with a talk entitled the Future for You. He gave an annual update on molecular nanotechnology at Singularity University on nanotechnology, gave a TEDX talk on energy, and advises USC ASTE 527 (advanced space projects program). He has been interviewed for radio, professional organizations. podcasts and corporate events. He was recently interviewed by the radio program Steel on Steel on satellites and high altitude balloons that will track all movement in many parts of the USA.
He fundraises for various high impact technology companies and has worked in computer technology, insurance, healthcare and with corporate finance.
He has substantial familiarity with a broad range of breakthrough technologies like age reversal and antiaging, quantum computers, artificial intelligence, ocean tech, agtech, nuclear fission, advanced nuclear fission, space propulsion, satellites, imaging, molecular nanotechnology, biotechnology, medicine, blockchain, crypto and many other areas.