“My personal challenge for 2016 was to build an artificial intelligence simple to run my home — as Jarvis in Iron Man”. The least we can say is that Mark Zuckerberg knows how to instill curiosity from the first sentence of his blog posts.
The server Jarvis is connected to three bricks AI (facial recognition, voice recognition and recognition of language), and peripheral devices. (Source : Facebook)
Has he been able ? And of Jarvis, the personal assistant forged by the boss of Facebook on what must be his free time, seems to be the least clumsy. Her body defending, his master spent only two hours per week on average to make a platform speech recognition, textual and visual. The tool is designed to work exclusively in the house of Palo Alto (California) in which it was designed.
the force of The “Make”
For the time being, Jarvis controls the lighting of the home, the temperature, the lock of the entrance door (via facial recognition, which is widely used by Facebook), and one or two knick-knacks, extra totally unnecessary. The wizard is activated and controlled from a smartphone or a computer. In short, not what to worry about Alexa (Amazon) and Home (Google). Jarvis, however, should continue to evolve in the coming months : Mark Zuckerberg would have really taken a liking to the home automation. “This would be a very good basis for a new product” – t-it.
where now the approach of the head of the company of 32 years is interesting, it is in his dimension, “maker” : hundreds of thousands of people all over the world create pieces of hardware and pieces of software, by way of a hobby or passion, to do personal projects that don’t have the ambition to exceed in popularity the family or the readers of their blog. In France, the standard of the movement is carried by MakerFaire between other actors (FabLabs, Magic Makers).
The boss of Facebook with this project illustrates a strong trend and makes visible a part of the digital economy : the geeky do-it-yourselfers who innovate. It used side code the same tools as those of the employees of his company, either Python, PHP and Objective C.
The question of standards
Then, Mark Zuckerberg gives some feedback to be useful. This is for example realized as a barrier to the progress of his project lies in the lack of standards to connect the various objects needed for the realization of tasks as he waited for Jarvis. “Even before being able to devise any artificial intelligence, I had to immerse myself in the code to succeed in connecting these systems, which all use different languages and protocols” written there.
Another constraint : the interpretation of natural language by machines is not easy : the semantic nuances are often very complex to interpret. He also argues that the system is not self-learner. It is so far of the current work on the deep learning and machine learning.
Finally, Fastcompany ensures that the question of the safety of the device has been taken very seriously by its author. It’s just as well, because it is the main barrier to the current deployment of connected objects in the business.
note that Facebook is working on the issues of artificial intelligence, among other, from his laboratory in Paris, created in 2015.
To go further on this topic
No comments:
Post a Comment