Take smart people, add the most disruptive new technology of our time, and inject money. Welcome to NVIDIA's AI startup challenge. Of the 12 semi-finalists who gave their pitches -- live before an audience of investors and press at NVIDIA's Silicon Valley campus -- six were selected to go on to the finals, to be held on March 27 at GTC, our annual GPU developers conference. There, the two finalists in three categories -- autonomous systems, enterprise and healthcare -- will get winnowed down to a single winner, who will get $330,000. Kicking off the event NVIDIA founder and CEO Jensen Huang said: ""AI is enjoying a revolution: software that writes software, machines that learn by themselves, solving problems that human software engineers had no possibility of addressing until now have finally hit the scene.
By happy circumstance, Santa Clara-based chip maker NVIDIA finds itself in the position of being an artificial intelligence (AI) startup king maker. The company designs and manufactures the entire computing platform for deep learning, the fastest growing field in AI, building everything from graphics processing units (GPUs) to software to systems purpose-built for deep learning. As the name might suggest, GPUs were developed to improve the computer graphics experience by offloading certain computationally intense image processing tasks from the standard central processing unit (CPU). The particular strength of a GPU is performing large numbers of parallel floating point calculations. This helps computer screens to increase in detail and complexity without sacrificing system performance, and modern gaming would not be possible without it.
NVIDIA's Deep Learning Institute (DLI) offers online and in-person courses on the latest techniques for designing, training, and deploying neural-network powered machine learning in a range of applications. As an Inception member, you'll have access to a wide variety of perks, including credits and exclusive courses.
In this article, we mathematically study several GAN related topics, including Inception score, label smoothing, gradient vanishing and the -log(D(x)) alternative. --- An advanced version is included in arXiv:1703.02000 "Activation Maximization Generative Adversarial Nets". Please refer Section 6 in 1703.02000 for detailed analysis on Inception Score, and refer its appendix for the discussions on Label Smoothing, Gradient Vanishing and -log(D(x)) Alternative.