Data analysis techniques are the methods used to collect and interpret large amounts of data. Developments in technology have been leading to a larger emphasis on data acquisition and analysis. One new frontier in data analysis that will be explored in the future is in the development of computers which can learn from experience.
As written about by John Markoff for the New York Times, computers are entering a new age where they are able to learn from mistakes and begin to automate tasks that previously relied heavily on human programming. This system is based on a biological nervous system, and how neurons connect with each other to interpret information. Computers systems built using these principles have been able to scan huge databases and train themselves to recognize things, such as cats, without supervision. This new approach also allows a tolerance for glitches. While a failure to complete a task could halt a traditional system, this new design is continuously adapting and working around failures.
I see this new computer system having a major effect on the data analysis field, and on intelligent building design in general. Firstly, as described in the article, this type of system opens the door for robots being able to walk and navigate safely in the real world. With further work, this could lead to robots which would be able to build autonomously, or with minimal human supervision, which is practically the definition of intelligent building design. In terms of simply analysis techniques, this allows for data analysis not previously possible. It would allow computers to look for and recognize trends independently without human supervision. This is important because the amount of data being collected is rapidly becoming too overwhelming for humans to deal with. However, humans have an ability to recognize trends and interpret data that a computer which purely sorts numbers cannot. If a system could be trained to identify cats, eventually it is likely that it could be trained to identify possible faults or instabilities based on images. If nothing else, computers could be fed construction images and be used to double check humans to ensure that no major construction errors exist. If one of these systems is provided access to sensors placed throughout a building, along with seismic data, it would be possible for it to interpret data in real time and detect possible collapses that humans may not be able to notice. Overall, the development of a computer system more based on the human brain brings up exciting possibilities for how machines can autonomously perform data analysis in the future.
Sources:
Markoff, John. "Brainlike Computers, Learning From Experience." New York Times. New York Times, 28 Dec. 2013. Web. 12 Jan. 2014. <http://www.nytimes.com/2013/12/29/science/brainlike-computers-learning-from-experience.html?partner=rss&emc=rss&_r=1&>.
This is the first I'm hearing about robots that are able to adapt and "learn" on their own. So many improvements to computing could be made if programs can make improvements without any human intervention. Its really cool to see that this is possible, but then again, this is how every plot to a robot revolution begins. Scientists better keep a close eye on these advancements.
ReplyDeleteYea that was my thought when I read this article haha. It's awesome what we can do with technology now...but do we really need robots that can walk around and teach themselves things? Seems like a recipe for disaster.
DeleteIt's interesting because in my sensor article, I read about a sensor that also adapts and learns similar to how you describe here - it recognizes routine motions and does not alarm the user as a result of them. It seems a lot of tech these days is becoming self-aware.
ReplyDeleteThis is a really interesting post, Steve. One thing I read about is how people believe that technology (more specifically AI) will one day be at an equivalent level to that of human intelligence and may be able to essentially take over and remake the way we live our daily lives. Kind of creepy when you think about it.
ReplyDelete