Pages

Monday, 28 December 2015

10 AMAZING ALGORITHMS


wps9D84.tmp
10 amazing algorithms
By Tim Greene,
Network World, 22 December 2015.

Cyber technology couldn’t get by without algorithms to encrypt, analyze metadata and find traffic anomalies, but they are used more and more widely in other fields. Here are 10 algorithms that perform functions as varied as scanning for disease genes, catching classroom cheats and figuring out murder mysteries as well as Agatha Christie’s heroine Miss Marple.

1. Scan genes for disease

wpsD067.tmp

An algorithm developed by doctors at Massachusetts General Hospital and Hebrew University could scan gene maps for genes that are associated with diseases. By doing so across organisms they could reveal how these genes evolved and what useful function they might have originally served.

2. Find patient zero

wps3778.tmp

Researchers in Croatia have an algorithm that performs statistical analysis to discover the first person infected by a disease in an epidemic or at least come close. It works better when the disease spreads quickly and the algorithm is applied soon after the epidemic is discovered. The algorithm could be used to track down the first machine infected in a malware outbreak as well, they say.

3. Perform facial ID without a face

wpsB827.tmp

Facebook has an experimental algorithm that can identify people in photos 83% of the time even if their face is obscured. It uses other cues such as attire, hair style, posture and body type to figure out who’s who.

4. Figure out whodunnit

wps7B1.tmp

Agatha Christie murder mysteries can be solved by an algorithm that takes into account the relationship between the victim and suspects, modes of transportation used in the crime, when suspects are introduced and how they are described, among other factors. Authors of the algorithm came up with it after analyzing 27 of Christie’s 83 novels.

5. Capitalize on tele-boredom

wpsC91F.tmp

Boredom of phone users is detectable up to 82.9% of the time using an algorithm that looks at usage logs and self-reporting of how bored users are. One useful result for marketers: users are more likely to investigate suggested content when they are bored.

6. Eradicate photo-bombers

wps94ED.tmp

An experimental application from Adobe called Monument Mode can improve vacation photos by digitally removing people who wander into and obstruct the view while someone is shooting a picture of a tourist attraction like the Grand Canyon or Statue of Liberty. Users shoot several frames and an algorithm in the app analyzes them, discerns the difference between the attraction and the people, and excises them.

7. Catch exam cheats

wps193.tmp

Freakonomics author Steven Levitt and economist Ming-Jen Lin of National Taiwan University wrote an algorithm to figure out who cheated on college exams. They considered where students sat and what answers they got wrong to discover that about 10% of the students cheated. After assigning random seats and beefing up monitoring, cheating virtually disappeared.

8. Create memories in damaged brains

wps1836.tmp

An algorithm in a prosthetic device figures out how to take the electrical signature of a short-term memory and convert it to the signature of a long-term memory, bypassing the damaged part of the brain that would otherwise perform the translation. “It’s like being able to translate from Spanish to French without being able to understand either language,” said researcher Ted Berger of University of Southern California.

9. Control blood sugar with diet recommendations

wps400F.tmp

Monitoring what people eat and how their blood sugar levels respond led to an algorithm that tells them what to consume in order to prevent spiking that diabetics experience. Using 137 data points such as age, body mass index and even gut bacteria, the algorithm could predict seven out of 10 times how subjects would react to the foods they ate.

10. Predict success of couples therapy

wpsB5CD.tmp

Voice qualities - pitch, intensity, jitter, warbles and shimmer among many others - when run through an algorithm created at the University of Southern California and the University of Utah was better at predicting whether or not couples would succeed in marriage counseling than the words they spoke. The algorithm broke down speech recordings into acoustic features that the algorithm analyzed. The predictions were compared to marital status after five years and were 79% accurate.

Top image credit: Kim Manley Ort/Flickr, CC BY-ND 2.0.

[Source: Network World. Edited.]

No comments:

Post a Comment

Please adhere to proper blog etiquette when posting your comments. This blog owner will exercise his absolution discretion in allowing or rejecting any comments that are deemed seditious, defamatory, libelous, racist, vulgar, insulting, and other remarks that exhibit similar characteristics. If you insist on using anonymous comments, please write your name or other IDs at the end of your message.