Tuesday, January 13, 2015

B1 - Future Post

AI:
In the first part of KQED interview, the way artificial intelligence was being portrayed and explained was as a way to help robots function in their environment.  It was being used as a way to help robots learn how to perform tasks that would help out humans with everyday tasks.  This is an important part but I think that the machine learning algorithms and programs that they describe in the second part of the interview are probably the more important and exciting part of artificial intelligence.  Machine learning is almost the same as how a human child learns.  The computer is presented with many examples, for example a giraffe in the video, and from the examples it learns how to recognize that thing in a larger set of data, such as a giraffe in a picture.  This interview also interviews an engineer from IBM who talks about the importance of machine learning, which was critical in IBM’s supercomputer Watson, who competed and won on Jeopardy against the best human players in the world.  The real benefit of machine learning is the way that a computer can sift through immense amounts of data looking for patterns that it can recognize and act on that data much faster than any human would be able to.

Computer:
As important and integral to modern life as computers are, it’s amazing how lax the security is for many of them.  According to the report prepared by Fire Eye, a cyber-security consulting firm, 97% of the companies which they monitored during the course of their testing had their security software breached, and the typical organization was attacked by some sort of threat more than once per week.  Since this study only covered organizations, who are supposed to have the best security and professionals who help minimize their risk, what does that say about the average consumer’s computer?  More importantly, as more and more legal and financial transactions are brought online, how safe can anyone’s data be?  It is important to recognize that in order to protect ourselves, we need to invest more in security to keep our data private, and keep strangers out of our computers.

Software:
Going along with what I talked about earlier in the computer section, some of the most dangerous threats to computer security come from software run by APT (or advanced persistent threat) actors.  These organizations are typically backed by a national government, and use some of the most advanced software available to attempt to disrupt operations, steal data or destroy infrastructure.  One of the best examples of this type of software may be the Stuxnet malware, which reportedly destroyed 20% of Iran’s nuclear centrifuges by infecting the computers controlling them.  According to Ralph Langner, who first discovered the malware, the malware most likely was developed with both United States and Israeli backing, to halt Iran’s nuclear capability.  Regardless of who was behind Stuxnet, the reality is that in the 21st century it is possible for large organizations and even nations to use malware to steal data and cripple critical infrastructure, and it is very important that the defense of our computer infrastructure be considered.

Future:

I think that in the future, the combination of artificial intelligence’s knack for seeing patterns and responding in an appropriate way will give us a huge advantage over malware when it comes to defending our computers.  If a computer can use machine learning to learn the signs of specific and commonly used types of malware, and how to respond to remove them, it would be a huge advance in defending computers.  It may still be a ways into the future, but I do believe that AI could successfully be used to detect and prevent malware much more successfully than the current methods such as commercial anti-viruses.

No comments:

Post a Comment