Technological Innovation Pushes traditional companies out of the market (Source: Netease Technology Channel)
This is a clear example of how Apple has jumped from being on the verge of bankruptcy to being the most popular brand in the world. The same applies to companies that have less than 20 years to go, such as Google, Amazon and Facebook.
Future Outlook: If your company doesnt invest a lot of money in data-driven AI, it wont last for the next 10 years.
Video 2: Learning Gene Programming Cells to Hunt Cancer Cells
Learning to Drive by Evolutionary Algorithms (Source: Netease Technology Channel)
This example is in the field of intensive learning, using neuroevolutionary technology to teach cars to learn racing.
Future Outlook: In the 1960s, driving was a pleasure and privilege. Nowadays, driving is a big problem because of the close connection between cities and traffic. Algorithms like this will help us develop driverless cars, which can do better than we do without pressure.
High Performance Computing (HPC) will become the pillar of AI innovation
Technological innovation is happening at a very fast speed, and AI and its related architecture will make it develop faster.
Fast broadband Internet, inexpensive high-performance computing hardware will emerge, and in the future, engineers and managers will also have high-speed computers on their desks. High-performance computing is a back-end technology, which makes cities more intelligent, makes organizational data-driven and decision-making a simplified process, and can easily screen large amounts of data.
Artificial intelligence will further accelerate the pace of technological innovation
It all started with AI, Googles open source machine learning library TensorFlow, AI library and tensor processing unit, and Facebooks open source way of publishing PyTorch. Today, we hear that Uber, Netflix, Tesla and almost all fast-growing companies are using some form of machine learning or in-depth learning architecture. Invida is clearly ahead of the GPU, but there are two things that will determine the next wave of the revolution. One is about tensor architecture; the other is that we will see the decline and slow decline of 32-bit and 64-bit IEEE754 floating-point architecture.
Artificial intelligence will completely obsolete many computing paradigms in the 1990s and the first decade of the 21st century, as new models, architectures and hardware solutions will flood the market in the next five to seven years.
Technically, we will see more advanced linear algebraic solutions embedded in the hardware for parallel computing, which is the best of the in-depth learning system, using multi-layer sub-matrix BLAS (basic linear algebraic subroutine) to further accelerate matrix multiplication.
One day, when we look back on the past, we will say that artificial intelligence is a turning point for mankind. Ten to twenty years from now, when we look back at the current computing infrastructure, data centers, desktop computers and devices, we smile as we look back at the outdated computing methods of the 1950s.
Thats why the AI economy will be huge, because it will lead to massive industrial-scale reforms, as the Internet did 20 years ago.