The Institute of Technology is holding its end of the quarter colloquium today. Students from the undergraduate and graduate programs in Computer Science, Computer Engineering and Information Technology programs report on their directed readings, directed research and internships. The number of students who have projects and readings to report on are a record number this quarter. Forty students, presenting individually and in teams are filling to large classrooms in the Cherry Parkes building on the campus for the majority of the day.Here is just a sampling of presentation topics
Improving Accuracy in Image Spam Detection
Last quarter's research into detecting image spammers ended with significantly sub-optimal accuracy rates for image spam detection in online photo-sharing site Flickr. This quarter, the project was centered around developing a more accurate pattern for use in identifying spam. The number and variety of keywords recognized by the system were expanded, and additional information gathered from the image itself. A data set was then constructed such that a bio-informatics team could run scans and develop an accurate pattern, program the pattern into last quarter's spam detector, and thus improve the accuracy of the detector.
A Study of Information Security and the Tor Internet Anonymity Tool
This paper represents a comprehensive review of the knowledge and information gained through a 10-week study and research course into the field of information security. Topics covered include e-health security, honeynets and botnets, access control, web security and IPSec, Firewalls and VPNs, and online anonymity, including data from a project consisting of a performance analysis of the Tor Internet Anonymity tool.
Exploring Computer Forensics: Is Data Ever Really Gone?
As technology changes so does the way data is stored. Is data really gone when you delete it? Various reasons exist on why data is deleted or damaged but the most prevalent reason is when someone is hiding or deleting illegal information. A computer forensics specialist uses tools to find data that has been deleted or destroyed. The government uses computer forensics to find information on a system that was tampered with by a criminal or to find information about a criminal. Computer forensic specialists aren’t simply handed the data, they must find the data in the machine using different tools and then deduce what happened. There is also a major problem around company employees that have become irate; when this happens a computer forensics professional comes in to get evidence from the irate employee’s computer to find information about what data the employee tapered with. There will be a presentation given by two students detailing various signs and methods that computer forensic specialists look for. There will also be a Demo shown on how the ProDiscover software will recover data from a thumb drive.
Reinforcement Learning for Sponsored Search Auctions
The use of sponsored search is one of the key mechanisms by which search engines like Yahoo!, Google, Bing, etc. generate profit. Due to increasing demands for advertising space, most search engines are currently using auction mechanisms. Because of the limited budget of each advertiser, strategic bidding behavior plays a crucial role in sponsored search auctions. From an advertiser's perspective, an ideal strategy has to prevent the advertiser from overbidding, while at the same time bidding enough to obtain his favorite position, thereby beating competing advertisers. The goal of our independent study was to devise an optimal bidding strategy, based on reinforcement learning, that automatically decides how much to bid for a given keyword query on a given day in order to maximize profits by keeping costs per click down while achieving high conversion rates from potential customers.
Fuzzy Logic and Web Intelligence
Fuzzy logic is a computing paradigm that allows an object to satisfy a property only to an extent, thereby providing a way to deal with vague, imprecise, and uncertain information. During the last decade, many proposals have been made for the use of fuzzy logic to power intelligent web applications, such as information retrieval, sponsored search, web service selection, recommender systems, semantic web search, etc. The goal of our independent study was to compile a critical state-of-the-art overview of the use of fuzzy logic in this application domain, with a particular emphasis on the information retrieval area. As will become clear in our presentation, we have not only distilled an overview of existing work, but we have also dug deeper to find out what makes fuzzy logic suitable (or perhaps unsuitable) to be of use for the future generation of intelligent web applications.
This quarter’s colloquium also features student demonstrations from the Computer Engineering and Systems program of their Senior Projects. Just a few samples...
Fire Alarm System
The project shall be designing a fire alarm system with a Lego NXT robot and a cellular phone which can run Android Operating System and has a Bluetooth. The Robot will detect fire and send a signal to a cellular phone through a Bluetooth. The cellular phone will ring a bell and show the image from the robot to warn the user about a fire. Because user receives alarm signal with the phone, user can move in a distance that a phone can receive Bluetooth signal from a Lego Nxt.
Intel Car
The project shall be designing a sensing control system and an unlocking system for a vehicle. The sensing control system will receive signals from front and rear end of the vehicle to which the sensor system is installed on and display alert message and distance on the LCD display. The unlocking system will be sensing specific knocks, “secret knocks.” The system will receive and count knocks from the user. If the knocks are in the sequence, the system triggers the motor to turn the lock and unlock the door.
For a full program with student names and times see the Institutes we site.
No comments:
Post a Comment