Thursday, July 24, 2008

Music Technology

I have recently developed an interest in applications of computing technology to music. Not surprising since most engineering is done computationally these days and I have enjoyed music (performance, composition, theory, history, and OH YEAH listening) since I was a kid. So I perked up today when I saw two articles on this topic. One described a researcher in Japan who creates humanoid robots capable of playing instruments. This is quite a task since the robot has to both mimic human biomechanics and understand how to adjust so that the result is "good." The other article described a piece of software that Microsoft is developing which fits chord progressions to tunes. This is really cool since it does this probabilistically by mining a database of lead sheets. There is some really neat potential in this software, and the developers thankfully appear to have a fairly open mindset so that all user creativity doesn't end up squashed.
http://spectrum.ieee.org/jul08/6454
http://spectrum.ieee.org/jul08/6442

Thursday, July 3, 2008

Commonality Between Tennis and Hospitals - Trusting in the Status Quo

I just read about a controversial call made by the HawkEye camera system at last year's Wimbledon. As a tennis fan, I immediately found this interesting. Apparently the system is accurate to within 3.6 mm on average. In the 2007 final between Nadal and Federer, a critical shot by Nadal was called out, and looked out on TV. However, Nadal requested HawkEye's opinion, and it turned out that the computer called in it by 1 mm. He was given the point. This seems ridiculous to me (and others) since the judgment was taken at face value (100% accurate) even though by its own specifications the computer's ruling was statistically inconclusive. I can't wait to watch this year's final! It seems like the computer should either say "it's in," "it's out," or "I don't know." Nobody questioned this until now?
http://www.spectrum.ieee.org/jun08/6400

It turns out that medical professionals have also been going with the flow. Only recently did someone finally do a study indicating that RFID tags used to track medical devices can sometimes interfere with other life-critical equipment. Previously people assumed that they might but really didn't know if or how severely. Maybe some standardization of a "medical-class" RFID tag would help prevent this interference.
http://www.spectrum.ieee.org/jun08/6405
On a related note, the Supreme Court has apparently indemnified class-III medical device manufacturers from liability for device failures once the FDA has approved them. Limits on liability ok; no liability not ok! Danger Will Robinson!
http://blogs.spectrum.ieee.org/riskfactor/2008/06/faulty_software_in_medical_dev.html