computer systems seminar

clouds, crowds and phones -  the new tech revolution underway


latest news

Monday 17, January 5pm. Mail one page PDF of your project idea for mobile phone sensing to campbell@cs.dartmouth.edu.

Tuesday 2o, January 2 pm. Elevator pitch.

Tuesday 25, January 2 pm. Demo assignment one - see below. And, project group annoncements.

description

In this seminar, we will study the latest advances in mobile phone sensing and its interaction with the computing cloud and exploiting big sensor data (the crowd).

We will read the latest papers on this emerging field and program phones to test out some of the ideas that emerge in the seminar. A high degree of self learning is required for this seminar and a strong set of programming skills.

Assesemnt is based on presenting research papers, class involvment and programming projects. A large commitment in programming is required.
 

admin

Prerequisite: cs 23 and cs78 or cs 58
Instructor: Andrew Campbell
When:  Tuesday (presentations) and Thursday (lab)  2:00-3:50    
Where:  sensor lab (147 Sudikoff)

grading

Class participation: 20%; paper presentations and review 30%; project: programming 20% and final paper 30%.

Students will present papers and discuss the pros/cons of those papers in class. Students will also work on individual or group projects and write up a conference style paper. There is no midterm or final exam.

Top student will get an iPad (it's a large cellphone ;-) and an honourable mention on this webpage.

students

Xiaoyi Chen
Shaohan Hu     
Lei Liu     
Hong Lu     
Emiliano Miluzzo
Xiandong Ren     
Senate J. Taka     
Tianyu Wang     
John F. Williamson     
Rui Xie     
Ye Xu     
Xiaochao Yang
Mu Lin
Mashfiqui Rabbi

tips on presenting a paper

You don't want to repeat all the details in the paper because we've all read it and know the content. Here are some tips for your presentation:
  • What's the problem addressed in the paper and why is it important?
  • What's the proposed solution and why is it novel in comparison to the related work?
  • Are the assumptions made by the authors reasonable, is the methodology OK?
  • What are the design tradeoffs?
  • Present one or two of the more important results
  • What are your ideas for improving the ideas in the paper?
A good presentation gets the essence of the paper over and stimulates a discussion on the above topics. This is not a complete set of tips - I'm sure there are more.  Important: have a set of questions ready before your presentation and ask the class those questions at suitable points during your presentation: your job is to drive the discussion of the ideas in the paper - the good, bad, and the ugly. Practice your talk. Make sure you talk to the audience not the wall that the slides are projected on.

tips written review requirement

You need to write up one review of any paper you like and submit it before the last day of class. It would make sense to write the review up based on a paper you present in class. That way you can catch any discussion issues in your write up if you wish.

The basic idea is as follows. You are the member of a technical program committee for a conference and you get this paper to review. You need to argue what the good and bad points of the paper are and how it could be improved. What's the novelty of the paper, etc. You can use the “Tips for Presenting a Paper” as a means of structuring your write up. It would make sense to survey some of the papers cited (the important ones) to give you some more background to the paper. You could then incorporate that into the review that you will write.

I would assume the review would be around 3-5 pages. You could hand this in at any point. Perhaps the week after you have presented if you choose the paper you present in class to be your review paper.

Let me know if you have any questions.
 

phone programming assignments and resources

We can program iPhones or Andriod phones. You need to determine which phone you will use for the project and write some simple programs.  Checkout the phone resource information.

Assignment one (Due Tuesday 25, 2010)
 
I suggest that you write a program that continously reads, records and displays the location, time, accelometer. Get going on this initial assignment and complete it by Tuesday 25, 2010. You need to demo this to me during Tuesday's seminar.

projects

Project AnnonifyMe (Senate, John, Nic)
 
Project FaceBlur (Xiaochao, Xiandong, Xiaoyi)

Project WalkSafe (Tianyu, Lei)

Project Evolve (Hong, Ye)

Project VibN (Emiliano, Shaohan)

Each project needs to build a demonstrable system using the phone and backend. Each project should product a demo and conference quality paper as final output.

papers


Thursday  6, January - Starting Point

"2.5 billion smartphones shipped by 2014", Qualcomm CEO, New York Times, Jan 05, 2011

A Survey of Mobile Phone Sensing, Nicholas D. Lane, Emiliano Miluzzo, Hong Lu, Daniel Peebles, Tanzeem Choudhury, Andrew T. Campbell, , IEEE Communications Magazine, September, 2010.

How to Read a Paper, S. Keshav, ACM SIGCOMM Computer Communication Review, 2007

Tuesday  11, January - Phones, Crowds and Clouds

CrowdSearch: Exploiting Crowds for Accurate Real-Time Image Search on Mobile Phones, Tingxin Yan, Vikas Kumar, Deepak Ganesan (University of Massachusetts, Amherst), ACM MobiSys 2010. (Presenter: Tianyu Wang) 

MAUI: Making Smartphones Last Longer with Code Offload, Eduardo Cuervo (Duke University), Aruna Balasubramanian (University of Massachusetts, Amherst), Dae-ki Cho (University of California, Los Angeles), Alec Wolman, Stefan Saroiu, Ranveer Chandra, Paramvir Bahl (Microsoft Research), ACM MobiSys 2010.  (Presenter: Hong Lu)

Tuesday  18, January - Socio-Psycho Smart Phones

"EmotionSense: A Mobile Phones based Adaptive Platform for Experimental Social Psychology Research", Kiran K. Rachuri, Mirco Musolesi, Cecilia Mascolo, Jason Rentfrow, Chris Longworth, Andrius Aucinas, UbiComp 2010. ( Presenter: Lei Liu)

"Social Sensing to Model Epidemiological Behavior Change", Madan A., Cebrian M., Lazer D. and Pentland A., UbiComp 2010. (Presenter: Xiaochao Yang)

Tuesday  25, January - Complex Networks and Social Dyanmics

"Computation Social Science", D. Lazer, A. Pentland, L. Adamic, S. Aral, A.-L. Barabási, D. Brewer, N. Christakis, N. Contractor, J. Fowler, M. Gutmann, T. Jebara, G. King, M. Macy, D. Roy, M. Van Alstyne, Science 323, 721-724 (2009). (Presenter: Nic Lane)

"Scale-Free Networks: A Decade and Beyond", A.-L. Barabási,   Science 325, 412-413 (2009). (Presenter:  Emiliano Miluzzo)

"Understanding individual human mobility patterns", M. C. González, C. A. Hidalgo, A.-L. Barabási ,  Nature 453, 779-782 (2008). (Presenter:  Ye Xu )

Thursday 3, February - Smartphone Usage Patterns; Topic Models

"Diversity in Smartphone Usage", Hossein Falaki (University of California, Los Angeles), Ratul Mahajan, Srikanth Kandula, Dimitrios Lymberopoulos (Microsoft Research), Ramesh Govindan (University of Southern California), Deborah Estrin (University of California, Los Angeles), MobiSys 2010. (Presenter:  John)

"Discovery of Activity Patterns using Topic Models", T. Huynh, M. Fritz, B. Shiele (TU Darmstadt), UbiComp 2008, (Presenter: Shaohan)

5 minuate project update from each team.