Artificial Intelligence
Artificial intelligence is a thing of the present, there are numerous tech companies that have their own AI and actively use their AI’s to perform services for the public or commercial users. So it should not be surprising that Google is continually developing and producing its own form of artificial intelligence. But unlike other tech companies, Google has made a version of it’s artificial intelligence software public so that researchers, students, and companies can modify it for new tasks, upgrade it to make it more efficient and intelligent and learn from it. Google’s decision to crowd source the development of its artificial intelligence has given the company some valuable feedback that it would not have received or been able to develop otherwise. Google did not expressedly require feedback from those that chose use its AI software, but does expect some feedback for enabling the industry with its software. This also shows that AI development is no small task and even a company as big as Google cannot hire all the specialists that exist.
The artificial intelligence technology that Google creates is being worked on and implemented to provide more intelligent, improved services like Google search, voice search, and numerous android applications. The software is constantly tested with the data received from Google users that permit the use of their data to improve Google’s AI software. The AI software learns and optimizes itself based on the data and Google programmers use the data to teach, test and improve the software. The results of this can be seen in how quickly Google’s voice search has improved to understand Google users speaking in different languages and in with different accents and / or dialects. In two years the error rate for voice searches fell from 25% to 8%. [1] Considering the difficulty of this task, that is an amazing improvement driven by data analytics and efficient program architecture.
[1] Gershgorn, Dave. "How Google Aims To Dominate Artificial Intelligence." Popular Science. N.p., n.d. Web. 12 Jan. 2016. < http://www.popsci.com/google-ai >
Robotics
The advance in robotics has come relatively quickly but has not really completely caught up with science fiction and implement in many industries. One of the few industries that heavily relies on robotics is space travel and exploration. There are many probes and rovers that are often too far away from earth to be remotely controlled and have to autonomous to a certain extent. This largely depends on how far away the science vehicle is away from earth. The mars science laboratory rover named Curiosity is only autonomous at navigation and relies on human input to decide whether or not to perform some tests and experiments. [2] On the other hand the New Horizons probe that flew by pluto last year was completely autonomous during its flyby. The reason for this difference is because a round trip communication with the New Horizons probe would take about 9 hours while a similar communication with mars would take 45 minutes. Because of the extreme communication lag, the need for communication between the probe and earth had to be reduced and the nature of how fast the probe would pass pluto also made it imperative to make use of every second by programming the actions of the probe. This autonomy lead to the probe being designed to perform the entire flyby, full system checks and error corrections autonomously. [3] These are huge accomplishments and show how the automation of vehicles, at least in space, can improve the performance of a vehicle and operate in relatively hazardous environments. They prove the validity of developing future autonomous vehicles to perform mining and construction tasks. [4, 5]
[2] "Spacecraft Systems and Components." New Horizons. NASA, n.d. Web. 12 Jan. 2016.
[3] "NASA'S Mars Curiosity Debuts Autonomous Navigation." NASA'S Mars Curiosity Debuts Autonomous Navigation. NASA, 27 Aug. 2013. Web. 12 Jan. 2016.
< http://www.jpl.nasa.gov/news/news.php?release=2013-259 >
[4] "Building a Lunar Base with 3D Printing." Solar System Exploration Research Virtual Institute. NASA, n.d. Web. 12 Jan. 2016.
[5] "RMC - About the Competition." NASA’s Robotic Mining Competition. NASA, 17 Apr. 2015. Web. 12 Jan. 2016.
< http://www.nasa.gov/offices/education/centers/kennedy/technology/nasarmc/about >
Comments:
Bryan,
I found all three of your articles very interesting and inspiring, but I wonder how long the industry will take to advance to making these processes feasible on an everyday basis. There are many roadblocks, such as capability limitations of the 3D printers that you have mentioned and those that have been mentioned in class, that will need to overcome before that will happen. I also see it being hard for the 3D printed building industry to grow without the development of 3D printers that are reusable. I do not see the industry developing large, unique 3D printers that can only be used for 1 building design. This is approach would be economically infeasible and limit the construction speed. This would favor the use of small 3D printers like those used to make the steel bridge in Amsterdam. Because of their smaller size, the printers could be quickly produced, easily replaced, more capable of constructing different building designs, and readily collaborate with other printer types to produce composite structures. If this were to happen in the future, it would all but revolutionize how buildings are constructed.
Redus,
I also researched artificial intelligence and found an example of how Google has incorporated artificial intelligence in its applications to the application created by Dulight to assist the blind. Google actually uses artificial intelligence in a lot of its applications to perform tasks like translating text in a picture from one language to another, understanding speech, and recognizing the people in pictures taken and upload to the cloud. There are numerous other examples and it is amazing and somewhat unsettling what can be accomplished with the use of artificial intelligence.
Updated to include tags, originally submitted at 3:45pm on January 12th
ReplyDeleteUpdated to include tags, originally submitted at 3:45pm on January 12th
ReplyDelete