July 18, 2006

Notes from Monday 2007-07-17's talks at ACL/COLIONG 2006

Notes from the second machine translation session at COLING/ACL in Sydney Australia. If you aren't a computational linguist, this will probably not interest you. Even if you are, I am not making any promises...

read more (1585 words)

July 5, 2006

How to use Excel charts in LaTeX documents on Mac OSX

I use LaTeX to write most of my research-related papers. I really like LaTeX. I think it has a high learning curve, but it does give very nice results, makes formatting something you don't have to think about, and is great with references and citations. I love how you can easily re-organize your paper's structure, and LaTeX (with BiBTeX and friends) will always make sure that your section and figuring numbering will be correct.

What I don't really like about LaTeX is that it is very hard to set up, and doesn't integrate well with modern tools. Copy and paste for images is unthinkable in this model. The only real graphic format that is well supported is Encapsulated Postscript (.eps) files. You can argue that using pdf files for graphics works, if you use pdflatex, but that opens up a completely different can of worms. I like sticking with regular LaTeX. You can always convert a .dvi file to ps or pdf, you can use pdflatex with .eps files, but you can not go the other way around: you can't make a .dvi with a LaTeX file that includes .pdf files for graphics, at least in my experience.

I wish LaTeX would support better graphic formats, like SVG. I use Inkscape a lot on Linux, and like it. Since LaTeX does use EPS though, isn't it nice that under Mac OSX we can print anything to a PostScript file? Here is a description (sans photos, for now at least) of how I recently included some Excel charts in a LaTeX paper I'm writing.

Create a .ps file for the chart

In Excel, select the chart you want to include. Print it, and in the print options box click "Page Setup...". Under the "Page" tab, make sure that Orientation is set to "Portrait", then click over to the "Chart" tab. For "Printed chart size" select "Custom". Click "OK" to get back to the regular print options, and click on the PDF drop-down menu. Select "Save PDF as PostScript..." and save the file as a .ps file somewhere.

For reference, I based these directions on some information I found for windows users on the same topic.

Convert the .ps file to a .eps file

Now we have to convert the .ps file to an .eps file. While this is supposed to be a process you can do by hand, I've never had much luck doing it by hand. Instead, I suggest you use ghostview or The Gimp to convert to an .eps file. I have installed The Gimp on my Mac using fink, so that is the route that I took. Be sure to check "Encapsulated Postscript" in the Gimp when you save, and you might not want to have any offset at all. Also, you should crop the image to the correct size - The Gimp's auto-crop feature worked well for me.

Save the file with a .eps extension.

Add graphic to your LaTeX file

I use the graphicx package for graphics usually, and just include the resulting file. I usually have to reduce the size of the charts, but it all works really well.

I hope some people find this information useful. I know that I will in a few months when I ask myself "Now, how did I include that Excel chart in my last paper again?"

October 13, 2005

Visit to Hiroshima City University

On Wednesday morning I went to Hiroshima City University (広島市立大学) and met with Professor Hidetsugu Nanba (難波 英嗣), and was introduced to Dr. Teruaki Aizawa (相沢 輝昭 先生). I gave my multi-lingual multi-document summarization talk, as well as a brief demo of NewsBlaster and a short talk about graduate student life at Columbia University. Professor Nanba and one of his students, Ms. Taniguchi, gave a talk about their work on PRESRI, an automated survey paper generation system, although currently it identifies survey papers.

Afterwards I met with some of the students, and saw a brief presentation on some of their hetergeneous database searching research that also uses some interesting citation analysis methods.

I had some Hiroshima Yaki (a type of okonomiyaki) at Micchan (a chain) at Hiroshima station. Then I caught my 6:30pm Nozomi to Tokyo, and arrived at Shin-Yokohama at 10:10pm. I wish they had electrical outlets so I could use my laptop for longer than the hour that the battery lasts, but I really like the Shinkansen rail system. It is very comfortable and convenient.

October 12, 2005

Lab visit at Mie Daigaku (三重大学)

Tuesday I visisted Dr. Masui (桝井先生) at Mie Daigaku. He introduced me to Dr. Atsuo Kawai (河合 敦夫 先生) and I was able to see some demonstrations of the research that they are doing. They have some very interesting work on identifying errors in Japanese and English (particles and determiners respectively) and some interesting Q&A work, as well as automatically learning features for nouns from the internet. Some of the Q&A work uses patterns, similar to Sasha's genus / type identification for ACQUAINT.

October 7, 2005

Kyoto Daigaku, Kinkakuji, back to Tokyo

Small camera phone picture of Kinkakuji

On Friday I left the Keihana area and headed to Kyoto to meet with Professor Toru Ishida, whom I was introduced to by Professor Sal Stolfo. Professor Ishida spent some time at Columbia University years ago. We had a nice lunch and talked about the upcoming "Language Grid" project and other things. I should contact Isahara-san at NiCT.

After meeting at Kyoto Daigaku for lunch, I headed out to Kinkakuji. It was really nice - the golden pavilion is just so dominating. I took some pictures with my Japanese cell phone, but it turns out that the camera on that thing is like 64x64 or something! It's horrible! While walking to the temple I met a family of three from Utah who had visited Kiyomizutera and Ginkakuji earlier in the day, and were finishing up with Kinkakuji. It seems like they were having a really nice vacation in Japan, which is nice to see.

After Kinkakuji, I headed back to Kyoto station and had a coffee. I tried to connect to the free wireless network in Kyoto, but you had to register (in person!) to get an account to set up a VPN. So that didn't go anywhere. Anyway, I caught my Shinkansen at 7pm, and made it back home by 10:00pm.


October 6, 2005

Talk and interview at ATR

I gave a talk today at ATR (also in Kyoto) and met with some people there: Dr. Yutaka Sasaki, the head of the Natural Language Processing department. I also spoke briefly with Dr. Eiichiro Sumita, a principal researcher in the NLP group, and I heard about Dr. Kiyonori Ohtake's work on Japanese paraphrasing / simplification.

After lunch with Dr. Sasaki, I met with Yoshinobu Tonomura, the deputy director at NTT. We talked for about half an hour I believe, and I think he was impressed with (if nothing else) my Japanese.

October 5, 2005

Shinkansen to Kyoto, talk at NTT

I took the Shinkansen to Kyoto today, from Shin-Yokohama. It took about 20 minutes, not including the ridiculously long time it took me to get to Shin-Yokohama. To get there, I took the Toyokosen to Kikuna from Jiyugaoka, about 25 minutes, then at Kikuna realized that I had forgotten my passport (which I will need to exchange my order for the JR Rail Pass I bought) so I turned right back around and was back at Jiyugaoka 25 minutes later. A short 10 minute taxi ride to the apartment and back to get my passport, and then another 25 minutes to Kikuna, from there one stop on the JR Yokohama line, and I had about 2 minutes to spare to catch the last Hikaru Shinkansen (the fastest type you can get with a JR Rail Pass.)

Of course, that train had no reserved seats remining so I had to chance the unreserved seating, but since it was only the second stop, there was plenty of room. I arrived at Kyoto at about 11:45am, and arrived at NTT at about 12:30pm.

The talk went well, and I met with Hideki Isozaki, supervisor group leader of the Knowledge processing group, Eisaku Maeda, Tatsuto Takeuchi (HR, Research Planning Supervisor), Naonori Ueda (Executive Manager), and Hajime Tsukada (Senior Research Scientist.) I really got the feeling that I would be welcome there. It was a nice visit, I saw a demo of their Q&A system, and heard about some knowledge ontology work that they are doing.

Go to Page: 1  2