[cs615asa] [CS615asa] Tech Talk Meetup Summary

Steven Nunes snunes at stevens.edu
Fri Apr 28 00:40:59 EDT 2017


Hey all,

For my tech meet-up, I decided to attend a tech talk hosted by engineers
from Bloomberg (https://www.bloomberg.com/) that took place right on
Stevens campus. This talk was an exciting opportunity to learn more
in-depth about some of the real-world responsibilities and endeavors being
pursued involving big data. The talk was certainly pertinent to the field
of System Administration as it discussed techniques on how to analyze large
quantities of data and the hardware architecture required to do so.

In particular, the talk focused on how Bloomberg manages large quantities
of data and evaluates this data by leveraging machine learning techniques.
Specifically in Bloomberg's case, this massive amount of data is being
accumulated from various sources and analyzed for financial information and
reports.

One of the major obstacles that Bloomberg engineers face is effective
detection and parsing of data tables, which can come in a variety of
different forms and appearances. As we've seen in class, this is also an
obstacle that is very much possible for a system administrator to face.

Because Bloomberg is handling the accumulation of data from a wide variety
of sources and these tables can be presented in all different kinds of
forms, they rely on computer vision and natural language processing to
effectively detect and parse these tables. In particular, they utilize a
region proposal technique called "Faster R-CNN" to do just that.

Here is one such implementation of this technique:
https://github.com/rbgirshick/py-faster-rcnn

Ultimately I learned how data is being presented and organized in a growing
variety of different ways and how designing automated systems to get a hold
of this data from various sources and evaluate it is a task that will be
essential for a variety of fields and it is likely a task that systems
admins may be faced with as well.

Thanks for reading,
Steven Nunes
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.stevens.edu/pipermail/cs615asa/attachments/20170428/a03eea92/attachment.html>


More information about the cs615asa mailing list