Artificial intelligence (AI) is used in many different sectors nowadays, for example robotics, health care, gaming, motion and manipulation and music but its basis is still the analysis of big data.
Today social media is one of the main stream and source of digital information (images, sounds and texts) and AI is used to analyses the massive amount of information and predict events based on the behavioral traits, events history and such.
For example, one of the most important social networks, Facebook, uses own algorithms and it created an automated system, which is capable to promote the right ads to the right users and furthermore the system knows the right time and right place to propose those ads.
In the example above, the raw data come from the “like” buttons people use for approval, from cookies, from the texts or from image recognition.
Overall IBM is the leading company for the number of patents related to AI (source USPTO) but Google is the company that more invested in image recognition, natural language processing and neural networks buying 5 different companies in the past two years. Other companies that are also investing in this field are: Yahoo, Twitter and Facebook. .
The main interest of those companies is to develop an application able to understand how people take images, what they take pictures of and why.
An example of software algorithm is the application of Microsoft “How old do I look?”, which is capable to tell the person age from a picture. Pinterest’s search engine “Guided Search” is based on images search and it is very inspiration for idea or narrowing down a research based on general need/idea.
The company which made a big progress in image search is Facebook. Indeed, according to Mike Schroepfer, Facebook Chief Technology Officer, Facebook’s AI Research team has developed a system that reached unprecedented results, the main ones are:
-
It detects objects in pictures 30% faster and using 10x less data training compared to existing systems.
-
The system is able to understand human speaking, to comprehend questions and to provide with answers.
-
It can predict the outcome (or consequences) by learning from observation.
The info above can be found at the web address: http://tinyurl.com/nz68bmy.
As seen above there is allot of interest on AI and, furthermore, the technology is available today for research and these two components can bring to sensational findings really soon.
.
