1. e-Readers in the Classroom?
This post discusses Princeton’s use of Kindle-DX as the primary reading mechanism in some of its courses. Most of the courses are from non-engineering and so I am curious of the impact on using it for engineering courses. Also now that IPad is out , it will be interesting to try it out.
In another news from Academia Who Really Failed? where a professor was removed from a course for setting a tough course (and exams !) . Where is academic freedom ?
3. Google News
Google was in news for a lot of reasons this week.
A new Google Docs – Google has refreshed Google docs and the new one is pretty impressive and snappy. Nice job !
Drag and drop attachments onto messages – Now you can drag and drop files in to Gmail for attachments. Currently it works in Chrome and Firefox. I can understand why IE is not there , but where is Safari ? Safari also uses WebKit which makes it all the more baffling .
4. ML Stuff
Search with fewer keystrokes and better spelling – City specific Google suggest and a creative spelling correction using context . Its quite awesome to see how machine learning is bringing all these almost magical features.
Google Follow Finder: Find some sweet tweeps : I was pretty impressed with their replay feature. But this one is even more cool. I am not sure how they are doing the recommendation but I am willing to bet that it is better than Twitter’s own suggested users. In the other Twitter news , all the tweets are going to be archived in Library of Congress. Interesting ! See Tweet Preservation.
More data and charts in Top Search Queries – My favorite Google announcement of last week. In the Google Webmaster central , Google already give lot of useful data about your Blog stats and how it fared in Google search. From this week, the data is much more improved. My favorite is that they show much more than top 100 search queries. My blog , for one, has around 30-50% of daily traffic using search and this gives me some more insight into how people search for stuff. I have lot of interesting ideas about using this info and some neat Python data analysis scripts. I will write a post on how to use them soon !
MLcomp: a website for objectively comparing ML algorithms – A real nice idea. I browsed the site and was pretty impressed. I only hope that it becomes better over time.
Drug discovery, Netflix style? : Interesting to note that even a simple ranking algorithm performs better than the traditional AI techniques for selecting drug candidates. Another striking thing was the collaboration between CS/Chemical and Medical schools.
Loose clicks sink ships – A clever idea to use statistical NLP to detect what users are typing using the sound the typing makes. As suggested, turning up the radio is not a solution as even simple techniques like ICA can separate music from typing sound.
5. CHI: Do we really need three reviewers for every paper?
A real neat idea from David Karger. I have to say , I found the savings to be substantial with minimal impact. My understanding is that , he used this years CHI data. I am curious if the argument holds for previous (say 2-3) years submissions too .