Srikanth
We have seen many impossible things happen in 2017, including electric cars taking over fuel cars, AI powered robots like Sophia, capable of mimicking thousands of human facial expressions, and not to forget the craze of bezel less phones. Now 2018, will mark a very important point in the future. With 5G, Machine to machine communication, AI and other technologies on the rise, there is nothing definitive to be said. The course of technological evolution has taken many bumps and hiccups over the last decade, but it always found its way to be more creative. Extending the trends followed in the previous years, 2018 should be attracting most of the tech companies to release their AI supported gear very soon.
When we think of future of AI: it all depends on what kind of hardware is used to process the algorithms so meticulously crafted. The computing power required by neural networks is just mind boggling. AI giants like, Google, Microsoft and other companies alike are using different types of processors to power their systems. For example, Google’s Tensor Processing Unit (TPU) can compute the algorithms much faster than a regular GPU (Graphic Processing Unit). This will definitely affect the balance in chip making industry. Popular chip makers like Intel, nVidia, AMD has to up their game to compete with growing AI needs.
The AI systems work in two stages, one is the training stage where you give the system to learn the concepts and in the second stage the end users will take benefit of it. Majority of times the power needed to train the system is more then execute it. That is why Google and other AI based companies custom manufacture the chips that can train their AI systems accurately in a very short period it time. However, the ones we use on our smartphones or laptops are not that powerful.
Manufacturing chips that can enable the use of machine learning and AI in user’s mobile phones or gadgets can greatly influence the way we see AI today.
We can only say that Machine Learning is ready for prime time only when we can let the users take advantage of everything that ML has to offer. FPGA chips and other very powerful AI chips that are coming out in our smartphones and computers are reliable. These chips are likely to see major updates in 2018 as the prime focus would be on enabling Machine Learning in users’ lives.
AI can’t fail like VR (Virtual Reality) and AR (Augmented Reality) did. May be industries and companies do use VR and AR, but they are not a part of daily life of people. However, Machine Learning has higher chances of embedding people’s life with much ease and 2018’ outlook shows that ML is ready engage the users. This can be a paradigm shift in tech world.
