Yesterday, tech company Nvidia’s stock jumped 5.3% due to a big opportunity for the company to benefit from a new trend in AI: being able to listen to, understand, speak, and give context to our speech.
With the rise in AI technology and improvements to it, it is a topic worth breaking down. So, what exactly is AI?
Artificial Intelligence (AI) is a branch of computer science that works with the intelligence of computers. This is mostly done through the use of machine learning algorithms and large-data sets.
These algorithms allow AI to take the large amount of information they are given, sort it, identify patterns within it, and use this information to learn how to do other tasks. Yes, you read that right. AI is actively LEARNING new things and applying them on its own.
To better understand AI and how it is used, I’m going to go through a condensed timeline.
1956
The term “Artificial Intelligence” is coined at Dartmouth College at a conference.
1974-1980
The “AI Winter“. Several reports are published criticizing AI which leads to reduced government funding and interest.
1980s
AI experiences a revival of interest as the British government starts refunding efforts in the field to get ahead of Japanese efforts with AI.
1987-1993
People lose interest in AI again and funding decreases once again during a market collapse.
1995
Mercedes creates the first self-driving car with AI technology.
1997
IMB system called Deep Blue beats the grandmaster of chess of the time. Here’s an interesting flip-side to this AI milestone.
2011
IMB system named Watson beats the two Jeopardy champions at the time, and Apple’s Siri is created by SRI International.
2014
Chatbot named Eugene makes judges think he is a human during a Turing test, which was used to evaluate if a machine was intelligent at the time and was developed in the 1950s.
2015
Deep learning systems manage to become better at image recognition than humans, allowing AI to be used to determine which crops are ready for harvest and which are not.
2016-2017
in 2016, Google’s AlphaGo software beats the best player of the board game Go, which is much more complex than chess. The following year, a newer version of the game was released and after three days of playing matches against only itself, the software beat the previous type of match against the best player 100-0.
This is important because sometimes we can’t supply AI with the data needed to solve certain problems, but if it can learn from it’s own experiences and train itself, so to speak, it can open many doors for this technology.
This allowed for more research into Dueling Neural Networks and Generative Adversarial Networks (GAN). GAN makes two systems test each other on something in a sort of game until they become so accurate they can beat the other system at the task.
One way this can be done is with image composition, where the systems have to give a real or fake image to the other system until the other system cannot tell that it is not a real image. Even if you’re only slightly geeky about tech this probably excites you, even if just a little bit.
2019
So what’s happening now? We are seeing a rise in the use of machine learning which builds upon 2018 findings. It allows the machines to learn by themselves, saving scientists all of the time that used to be spent inputing massive amounts of data.
We are also seeing facial recognition technology being used on a wider scale. It is being used in our phones instead of passcodes, on Facebook to automatically tag people in pictures, and it is even being adapted for use the medical field to more easily diagnose patients.
In the computer world, AI chips are allowing regular computers to have some AI capabilities, like facial recognition and machine learning. This most recent way to use facial recognition could soon become an issue for us in the U.S.
Cloud Computing is a rising topic within AI. Cloud Computing delivers things like servers, storage, databases, networking, analytics, software, and intelligence via the internet.
Other applications for AI currently include using it for transportation services like Lyft to minimize wait times; fraud detection, check depositing, and business and personal credit card screenings in banks; personalizing suggestions on streaming services like Netflix; and automating business processes that are otherwise time consuming for employees.
Future
About 15% of businesses use AI today, but an Adobe survey indicates that upwards of 30% of businesses surveyed plan to add AI technologies this year. It seems this complex, somewhat scary, yet very useful tool won’t be going anywhere anytime soon.
The world is also working towards finding a way to use multimodal learning with AI. Currently, each system works independently, but this concept would allow different systems to work together thus increasing the amount of info they have access to and how much they can process at a time.
This technology could be introduced into the auto industry for car computers, for security and payment authentication in businesses, for personalization of many things, as well as into the medical field to help improve imaging technologies.
The race to these breakthroughs is an intense one, with the top contributors being China, the U.S., the UK, Canada, and Russia. The competition is ruthless, as the U.S. just prevented Chinese companies from using our technology to advance their own initiatives.
As AI grows bigger, the need for more emphasis on privacy and security, ethical guidelines, and ways to reduce bias remain ever apparent. So, while we have come a long way, we still have a long way to go. Here’s to hoping it is onward and upward.
