1 / 2100%
IBM Watson is an AI that can understand written and spoken language
and can make accurate predictions on what a text is about and find the
answer to a question posed. This is a revolutionary technology that
continues to improve, even beyond it's show of dominance at
Jeopardy several years ago. Watson has the capability to use
numerous methods and algorithms on the backend, some of which
have been taught in this program, such as logistic regressions, support
vector machines (SVM), decision trees, and neural networks and can
use any or all of these methods in tandem, depending on the situation.
For example, since Watson had a small pool of training data to work
with while preparing for Jeopardy, the team at IBM elected to not use
a neural network, since it works much better with large training
datasets.
In any Watson application, the ability to accurately decipher natural
text is a great aid to any business. Whether it's analyzing business
contracts or user reviews, having an AI know the sentiments of
phrases or a document as a whole can aid in making better products or
pivoting a business model due to outside concerns, or when trying to
eliminate a weakness from the business that's hindering growth.
My current career has little need of the full power of IBM Watson or
any of it's subordinate functions, but one tool that could be useful in
just about any business is the Speech-to-Text service. Working for a
government contractor, my day is full of meetings talking about ships
and airplanes with Naval jargon and acronyms thrown about all the
time, which can sometimes make it difficult to take meeting minutes or
even write down any relevant information, especially if you're
someone who isn't from this world and familiar with the terminology.
The Watson Speech-to-text service is like any other speech-to-text
service, but it seems to perform very well in determining proper noun
usage with a high degree of confidence as well as identifying which
words belong to which speaker. The ability to determine what
acronyms are being used while 4 or more people are talking over each
other (as is often the case) would be very helpful in keeping useful
notes.
Another of Watson's service that could be beneficial would be the
visual recognition service. As a software developer, we will sometimes
have bugs in the code and users will send us a screenshot of where the
bug is occurring, without explicitly saying where we would need to
navigate to in order to find the bug and fix it. With enough training
data, Watson could aid in finding where in the software these
screenshots were taken, reducing the time it takes to find the bug and
ultimately the time to fix it.
Students also viewed