2016 Review and 2017 Preview

2016 Review

• January
• Frontier of Computer Science, Special issue for smart city submission due
• February
• G-SPASE Symposium 2016
• December
• PhD program progress report session II

Objectives in 2016

• Apply for DC2 or any similar financial aids.
• Failed to apply for it, but it turns out that I could earn as much as I could from DC2 with the combination of other opportunities.
• Publish a paper or give a presentation at an international conference.
• Did not accomplish. There are some papers in production.

2017 Preview

• February
• G-SPASE symposium
• March - May
• Paper submission
• June
• Dissertation draft submission
• July
• Dissertation presentation

Objectives in 2017

• Graduate

Paper review checklist

http://www.sciencemag.org/careers/2016/09/how-review-paper?utm_source=sciencemagazine&utm_medium=facebook-text&utm_campaign=reviewpaper-7757
I read this article and found convincing to create a checklist for reviewing papers.

I usually follow these steps:

1. Skim the paper. Highlight or add notes on potential problem/concern and interesting parts
2. Re-read the paper while following the checklist below and start writing the review comments on a Word document.
3. Make sure the review covers all of the highlights and the notes

Checklist

• [ ] Skim
• [ ] Ask questions
• [ ] Introduction
• [ ] Clear key ideas and contributions?
• [ ] Logic toward hypothesis makes sense?
• [ ] Related works
• [ ] Covers sufficient area?
• [ ] Covers sufficient papers?
• [ ] Logical critique?
• [ ] Model / Method
• [ ] Easy to follow?
• [ ] Innovative?
• [ ] Experiment
• [ ] Is the experiment setting fair?
• [ ] Discussion
• [ ] Is it critically discussed?

Writing review

1. summarize the contribution
2. briefly cover the background of the study.
3. assess the significance of the paper.

Finishing PhD

Some say that “finishing” Ph.D. program is like a marathon, while others say it’s like a set of sprints. Both have reasonable metaphors and it takes both strategies.

Why Ph.D. program is like a marathon:

• Every day counts. We can’t waste a few days and try to catch up later.

Why Ph.D. program is like a set of sprints

• We should not cruise through years. These are conference paper deadlines, journal paper deadlines, and the defense. We need short-term schedules and pace-making.

How to spend time
Morning (9:00 ~ 12:00)

1. Wake up early and start the best day.
2. Spend fixed amount of time for writing.
3. Refer “How to write a lot (http://www.apa.org/pubs/books/4441010.aspx)” book. I try to spend 1 hour/day for writing when I do not have upcoming deadlines. If there is an upcoming deadline, I would write as much as I should finish it during the daytime.
4. Spend fixed amount of time for reading papers.
5. Make sure to take notes and always think about how the paper you are reading contributes your paper. If you come up with nothing, stop reading it.
6. Work on long term tasks. Sometimes you work on tasks that require weeks to finish. It is important to keep working on it every day in order to make progress.
7. An important note is that these activities pay off after a long time. This means if you skip these by oversleep or something, you would not notice the effect. That’s why you need decisiveness to keep working on the morning activity.

Day (13:00 ~ 18:00)

1. Finish tasks (different from "doing"). Now it’s very important to finish tasks. Utilize GTD and PM tools like Asana to track your productivity.

Night (20:00 ~ after dinner and gym)

1. Write down final thoughts (paper, presentation, blog)
2. Work on MOOC program.
3. Stop rushing
4. Prepare for better tomorrow. Adjust the task schedule to make sure you can start your best day tomorrow.

How to learn skills

• MOOCs
• MOOCs are good opportunities to learn latest (especially computer science) skills
• Reading papers / books / tutorials
• Papers provide new techniques with the philosophy behind the new techniques. Books provide the history of the subject and it’s usually easier to learn techniques from books than papers. Tutorials allow you to practice techniques with clean datasets.
• How to read a lot of papers
• 10 papers a week
• 2 papers a day
• 1 paper / 30 mins
• Take a formatted note
• Highlight papers you will read next
• How to write a lot of papers
• Stick with writing schedule

Mac Setup

Last Update: 2016-11-19

1. Back up data directories
Make sure Dropbox / Google Drive / One Drive are in sync
[Documents]
.ssh keys


2. Run a clean install of macOS
Run software update

3. System Configuration
Activate FileVault 2 in System Preferences > Security
Modify caps lock key to control
Change host name at Sharing

4. Install Apps
From App Store:
Kindle.app
The Unarchiver.app
Xcode
CotEditor
ClamXav
BetterTouchTool.app
Line.app
Slack.app

From Web
Adobe CC
Clipy
Cyberduck.app http://cyberduck.ch/
JetBrains Toolbox
- DataGrip
- IntelliJ Idea
- Pycharm
Dropbox.app http://www.getdropbox.com/
Skype.app
GitHub.app
Google Chrome.app
Google Drive.app
Google Earth.app http://earth.google.com/
Google Japanese Input
iStat Menues.app
Mendeley Desktop.app
Microsoft Office 2016
RStudio.app
VLC.app
Sequel Pro.app
Steam.app
Docker.app
QGIS.app
Parallels Desktop
-Install Windows 10
-ArcGIS 10.3

Homebrew
-emacs
-zsh
-r
-python3
-qt5
-pyqt5
-mysql
-postgis
-git
-mecab
-mecab-ipadic
-mongodb

2015 Review and 2016 Preview

2015 Review

• January
• Attended a PhD defence session of my fellow student for the first time. Got an idea how the presentation for the defence should be organized.
• February
• Started working as RA at CSIS.
• March
• Missed paper submission due of popular Computer Science conferences.
• April
• Started working as RA at lab.
• May
• G-SPASE; a collaborative educational project.
• September
• Started working as engineering intern at Nightley Inc. Having an opportunity that a portion of my work can be "cashed out" with a start-up companies changed my prospect after graduation. Now I think it is also pretty exciting to work at industry or even start up my own company.
• October
• GISA 2015. A domestic conference that I consider as my home ground. But maybe I should go to other area sometime.
• November
• CSIS Days 2015
• December
• PhD program progress report session I. A very brief session.

Objectives in 2015

• Apply for DC2 or any similar financial aids.
• Failed to apply for it, but it turns out that I could earn as much as I could from DC2 with combination of other opportunities.
• Publish a paper or give a presentation at an international conference.
• Did not accomplish. 

2016 Preview

• January
• FCS Spatial Issue due
• ISPRS Congress 2016 abstract due
• February
• G-SPASE Symposium 2016
• KDD 2016 paper due
• March
• Ubicomp 2016 paper due
• April
• ISPRS Congress 2016 paper due
• May
• SIGSPATIAL 2016 paper due
• October
• WWW 2017 paper due
• December
• PhD program progress report session II

My life works around paper due dates in the next year. I think I can enjoy it.

2014 Review and 2015 Preview

Review 2014

January 2014 - March 2014
Finished the rest of applications for Ph.D programs in US
I was called for a Skype interview of one university, however, I could not received good results with financial aid.

May 2014
Submitted the capstone project report.
Graduate from the University of Maryland.
Though I was hoping and going to keep working at START after graduation, I was suspended because my work authorization card had not arrived. For the time I could not work, I went over some MOOCs and watched the World Cup games.

July 2014
Returned to work at START.

August 2014
Left to Japan to take an entrance exam at the University of Tokyo. Back to Japan for the first time in two years.

September 2014
Trip to France with my mother.

October 2014
Started PhD program in the University of Tokyo. I have been mainly focusing on learning Machine Learning.

Objectives in 2014
Graduate from the University of Maryland <- done
Publish a paper about the human trafficking project <- could not finish due to my leave & difficulties to work off-site
Enter a Ph.D program / get a full-time job <- entered a PhD program.

2015 Preview
January 2015 to March 2015
Call for DC2 candidate will be out.

February to March 2015
Deadlines for popular international conferences in Computer Science. It would be great if I could make it to one of them.

July 2015
Progress report and presentation to graduate committee.

Objectives in 2015
Apply to DC2 or any similar financial aids.
Publish a paper or give a presentation at an international conferences.

How to allow ArcGIS to build a large transportation network dataset

Sometimes when building a large network dataset (in my case 100M ~ edges) in ArcGIS, "Build network dataset" tool returns "read/write error". I at first thought it was because the feature class source was somehow corrupt, but it turned out to be insufficient temporary directory space.

Problem: ArcMap or ArcCatalog returns “read/write error” while building a transportation network dataset.
Cause: insufficient temporary directory space
Solution: Secure enough free space in TEMP directory.

A new window opens and displays the contents of the System Temporary Directory. When building a very large network dataset (for example, all of the streets in France), hundreds of gigabytes of temporary space may be required to process a task of this size. If there is not enough disk space in the current TEMP directory to perform the task, it is necessary to change the TEMP directory to a location that has sufficient memory. The following steps show how to make this change.

1. Right-click Computer and click Properties.
2. Click Advanced system settings on the left panel to display the System Properties dialog box, and click the Advanced tab.
3. Click the Environment Variables button.
4. In the Environment Variables dialog box, under User variables for , both TMP and TEMP may listed. This is the common location; each different login account is assigned its own temporary location.
5. Click the Edit button. In the Edit User Variable dialog box, under Variable value, specify the path to the directory that Windows and many other programs uses for temporary files. Change this to a location that has sufficient disk space to process the network dataset. 
6. Repeat this process for both the TMP and TEMP directories.
7. Restart Windows to begin using the new value for temporary files. Restart any running programs for the new value to take effect.

http://support.esri.com/de/knowledgebase/techarticles/detail/40743

Another lesson to keep free space in the boot drive.