Thursday, February 6, 2014

Hello Feedly!

Hi there again,

I didn't think I'd be blogging about my RSS reader this fast :x Since my last post I thought I'd be fairly happy with Inoreader. The easy google login and android app made it very attractive to me. The app is more usable than that of feedly, but the whole polish is a little less nice. And I'm a pretty bad sucker for eye-candy. Also in the end it turned out that the android app doesn't show embedded videos. :-/ Today we discussed it at work and theoldreader came up. I already checked this out on Monday, however there was no android app so I dismissed it quickly. Today we looked at it again, also nice google login (and facebook for both) and quite a lot of 3rd party nice looking android apps. (Especially gReader!) Sadly none of them seem to give the combination of looking good, being usable, not having adds and showing videos. So I decided to try and see if I could get the cvpapers.com RSS feed to work in feedly. The options was a lot easier than expected. Instead of using the provided rss link, just use this.  Same thing but compatible with feedly! The only bad thing about feedly now: it doesn't have a search, which I honestly never ever missed or looked for and the app has a crappy way it reacts to the back button. :-/ Apart from that though, it is the best one I have seen so far... so here it is: "Hello Feedly!",... welcome back.

Apart from that work has been busy finalizing my paper. Shit loads of things I didn't expect to come up and soon I'm leaving for Birmingham :x I hope I can finish it soon, so that I can finally get going on some new ideas again.

So see you again soon, not sure when I will have time to blog though :x Hopefully sooner than later!

Here is a pretty interesting video. I haven't watched it completely yet, but if you like creationists vs atheists discussions then this is cut out for you. I guess you know which side I am on ;) 

Monday, February 3, 2014

Goodbye Feedly!

Hello there everyone!

Nothing big happened since last Thursday, so I will tell you about the smaller things that changed ^^.

Ever since the guys and gals over at google decided to put the final nail in greader's coffin, I've been hanging with the cool kids over at feedly. I really liked it, although their android app had some usability issues in my opinion. The one big problem though: some RSS feeds simply didn't work :( I couldn't get the cvpapers feed to show up properly and this is the one feed I really need for work. It is a super easy way to get a nice overview of some papers you might otherwise miss. Because of that I really stopped looking at my feeds and large amounts of posts were just ignored. This also included all worldly news and other interesting things. So yesterday I decided to finally look for some other reader. There wasn't an obvious choice, but in the end I found inoreader. I'm still not sure if I like it, but it looks slick, you can log in with facebook and google, it has a useable android app and so far it worked out reliable! I hope I can recreate the good old greader feeling with it :) But I'll post again if I need to switch after all. :x

Also, we finally got to play our new game last weekend. I bought "Legends of Andor" (sadly no english wiki, sorry) before last christmas, but we only played it with two or three people. This time we played with four and also figured that we misunderstood some rules last time x). While the first few times we played it seemed fairly easy, this time we understood that it was actually pretty hard and it was very exciting. We even ended up, well...  "cheating" a little, although it turned out that it wasn't really necessary as I just miscounted something :p. The game contains 6 "stories" you can play and the nice part is that it is a cooperative game, so you play together against the game. So far we have only played the first story, but I'm really looking forward to playing the next one! :)

Apart from that work has been going rather slow. I've been reading a lot of papers, looking for some ideas and making a collection of features used to classify scenes. I'm not looking to use all of them, but I want to have some idea what has been used and I think some of them are really interesting :).

Also two people have applied for a Phd thesis. Both of them seem pretty good, but I'm a bit skeptical considering every PhD student we hire, means a Postdoc we can hire less. And I think some Postdocs would really do our lab some good ^^! Sadly they are a sparse good these days, so let's see if we can find one that wants to come to Aachen :D.

Okay, enough for today. As always after blogging, I want to do a bit of coding. After my last post I took a long time thinking about what to code and how to do it, but I decided that without actually seeing what I need in order to achieve some things, there is no point in over thinking it. So I'm just gonna start implementing and adjusting once I see it's not working out :).

See you soon :D Also, a classic from the "old days":

Thursday, January 30, 2014

The most terrifying creatures: students! ;)

Hey there everyone!

If you read the title you might think I am a whiny little guy that shouldn't complain about students, as I was a student until about a year ago and in fact I am still a kind of student right now. But... let me tell you about the people I deal with currently, of course only the superficial stories I can tell here without identifying people :D

This semester I was the teaching assistant for the seminar of the chair. It was only for bachelor students, so our expectations are typically not that high. They haven't done a lot of scientific reading and writing, so the main point is that they get some practice in finding and reading some papers and then writing about it. In order to get people to actually do all this, we give them clear rules about how many pages to write, about what content should be in their reports and how to do their presentations. Back when I did all this stuff, I would see the rules, fear the supervisors and do the freaking stuff. These day, people apparently can't read anymore, are just not interested in rules, or...? I don't know. So many questions arise about if they can do something which the rules clearly forbid :D ... But well, let's see the final grades :) Luckily I get to make the recommendations for those x) Which doesn't imply I am evil, I just think people should get what they worked for and I'm glad that there are also students that actually do a great job! :)

Then today I saw another set of students, those came looking for master thesis's and student assistant jobs though. So no rules, but just talking to people, checking out their knowledge and interests. My colleague had the awesome idea to do the "FizzBuzz"-test. I had never heard of it so far and actually took the test as well :) The idea of the test is that you give someone a sheet paper and ask them to code a little program that does the following:
  • Print the numbers 1 to 100 below each other. 
  • If the number can be divided by 3, print "Fizz" instead of the number.
  • If it can be divided by 5, print "Buzz" instead.
  • If it can be divided by both 3 and 5, print "FizzBuzz" instead.
It might sound like an easy problem to solve, but if you think about it a little, there is quite a lot that can go wrong. There are no real constraints, as to which language you can use, if it should be efficient or memory conservative. There is a lot that can be learned from this about someone. While I thought it would be kind of trivial at first, in the end you could really see who understood the program, who payed attention to the detail, who coded in which language most and who understood some math. Some of the results were super clean, others were short, some were just wrong, one even wrote down the C includes. :o Sadly my solution was wrong as well xD But the mistake was kind of minor and rather obvious. Very cool idea! Here is some further reading about it, there is even a Brainfuck implementation.

Okay,  time for bed now :) I just saw a trailer for 47 Ronin on tv.  IMDB says it sucks, but I know that I am a sucker for eye-candy so I think it will cater me well! Also here a cool song from one of the tv commercial, very.. interesting clip :D(Not completely sfw :o)

Monday, January 27, 2014

So much traveling!

Heyho everyone,

It's been a week and all I did was create a cmake file and set up my global git ignore files. Awesome right? (For those of you who don't know that much about computer science, it ain't much :p) Everything else has been keeping me rather busy lately. Little things like shopping, cooking, ordering important paperwork, watching series and all stuff.  ;)

So as I wasn't too successful at coding last week, nothing really new came up for my publication plans. But instead, we noticed that there are a shit load of project meetings coming up :-/ In the end this means that I will have to travel at least a few more times this year... I'm not really looking forward that much to these short little trips, but I guess some of them will be fun as my best colleague is coming too :D So maybe I can also see the fun part in it! Sadly not only the meetings came up, but also nasty paperwork we need to do, seminar work and master thesis supervision. All nagging away at research time. :x

But yea that is pretty much it. Tomorrow I need to get up early, so I guess I shouldn't go to bed too late. I don't think I will do too much coding, but I want to do some reading on semi-supervised random forests tonight :D I wonder why I never read this stuff earlier... :-/

Here is a cool video! Really inspiring! From a really cool channel.

Monday, January 20, 2014

Coding! Yea!

Hey ho,

The remainder of last week was pretty much spent on writing about what I did in the weeks before. It wasn't very fruitful, as I didn't really know what to write about, but hell, it was a really nice time learning about all this stuff! In fact we discovered that a large part of what we came up with, was exactly replicating a well know paper (Spin images) and also my wall detection algorithm was actually based on a known theory of relaxation labeling. Sadly we didn't get the whole pipeline running as we had hoped to, but I hope it will suffice for the project.

After reading the first few papers some days back, I actually read a few more and I really have an idea for ECCV now. Sadly I have no idea if I can get everything to work in time and if it will suffice for such a big conference :x But well we'll see soon! I will need to code a bit at home for it to work though, but I'm really looking forward to it :)

Today we also finally let our NIPS idea go. It occupied several whiteboards, in our office, for many months, but in the end... it was a dead end :p It all started out with me explaining random forests to my colleague a while back. We had crazy ideas about writing some awesome papers and wrote several whiteboards full of notes. Nobody dared to touch them and lately people really got annoyed by us taking up all the space. Still we had lots of fun thinking about it and we learned a lot about random forests and ferns in that time! I still have the feeling the idea holds some niceness, so let's see if at some point we can take up the idea and make something out of it! Would definitely look good on our resumes and working with him is the best thing that happens during a work day!

That's it for today I think. I really want to start coding a bit, wooooh! Exciting! See you all soon :)

Also, a bit late for Christmas, but a really cool video that my girlfriend dug up from somewhere in the internet! :)