Candy papers, lol

Some papers are just so good that you have a great time reading them. They are like candy. The flavor is good, and then you get a sugar rush (insulin boost). Well, some papers are like candy, but in a healthy manner. When you read them, you can enjoy the great exposition of cleverness. Idea after idea, insight after insight, observation after observation. And then, you get the sugar rush from the motivation!

http://snap.stanford.edu/class/cs224w-readings/Pothen89Partition.pdf - This is a classic. One can even enjoy the old style Computer Modern Roman font by Don Knuth.

http://arxiv.org/pdf/1107.3231.pdf - This one is about triangles and social cohesion. It’s relatively new. It seems that social network analysis needs its new theoretical boost. A lot of research has been on the study of statistical properties of networks.

http://www.cs.cmu.edu/~mmcgloho/pubs/butterfly-kdd08.pdf - This is about weighted dynamical networks. It’s imperative to get a good understanding about to do social network analysis on weighted and dynamic graphs.

http://cs.stanford.edu/people/jure/pubs/powergrowth-tkdd.pdf - This one is about densification laws in dynamic networks. It seems that the distribution of degrees and weights changes over time, following a power law related to the size of the network! So, not only a static image (snapshot) of the graph has power law properties, also its image over time. And diameters do shrink with size, too. Ain’t this cool?

http://www.dcs.gla.ac.uk/~vincia/papers/sspsurvey.pdf - Social signal processing. In social interaction research, automatically detecting social interaction, and classifying it, has gotten its right relevance. We have been using the motes for this in the CompEpi group.

http://www.au.af.mil/au/awc/awcgate/lanl/social_balance_0405041.pdf - Social balance theory, or how people could adjust their preferences until social harmony is reached.

http://biology.mcgill.ca/faculty/loreau/pdfs/loreauecollett.pdf - Diversity and saturation in ecological communities.

http://toc.proceedings.com/13745webtoc.pdf - The table of contents of SocialCom 2011. Just for getting more ideas!

http://www.anesth.uiowa.edu/portal/LinkClick.aspx?fileticket=QisfiFHuQag%3D&tabid=878&mid=1919 - Surgical ICU survival guide. ICU stands for Intensive Care Unit. There may be several in a hospital, for different types of patients. This is interesting for my research.

http://ccn.aacnjournals.org/content/31/2/e1.long - Nurses do nap during night shifts.

http://or.journal.informs.org/content/56/6/1335.abstract - Weird one: we die because of our stupid decisions.

http://www.palgrave-journals.com/hs/index.html - Emerging journal. It may be interesting to send a paper here.

http://www.treesmapsandtheorems.com - How to write scientific or technical reports.

Leave a Comment

CompleNet ’12 proceedings: “Complex Networks”

http://www.springer.com/engineering/computational+intelligence+and+complexity/book/978-3-642-30286-2

They included my paper “the explanatory power of relations and an application to an economic network” in this book. I tell you, it’s an awesome paper, and hence, an awesome book! ;)

Leave a Comment

Some nice links related to my research and almost related to it…

http://www2.warwick.ac.uk/fac/sci/statistics/crism/workshops/infer/talks/held.pdf

These slides by Leonhard Held summarize many of his insights and developments for studying disease incidence through reported data. The slides are very clear and simple. I really liked the separation of the incidence in endemic and epidemic (occasional outbreaks). I also liked the incorporation of spatial and networked relations in the statistical models.

http://www.nature.com/srep/2011/110812/srep00062/pdf/srep00062.pdf

In this paper in Nature Scientific Reports, Meloni and others (including Vespignani from all these network measurement and analysis papers) study how changes in mobility patterns during epidemics outbreaks change their incidence and prevalence. The study is simulation based, but it incorporates something that many of us consider important yet have not incorporated in our models: that diseases change the behavior of people and, thus, how diseases spread in contact networks.

http://jgaa.info/accepted/2012/BachmaierBrandenburgBrunnerFulop2012.16.2.pdf

“Drawing recurrent hierarchies.” It’s a pretty neat compilation of visualization ideas for graphs where there are hierarchical relations, i.e. X > Y but often not vice-versa. Some ideas are pretty wild. Interesting would be to apply these ideas to describe temporal relations in data.

http://www.irit.fr/recherches/ASRTP11/possqcl.pdf

This is a logics paper. I like reading these from time to time. Theoretical (formal) fields use their own unique style of presentation and demonstrations that is often unseen in applied research. They are very rigorous and their results are very reputable or, more precisely, infallible. Applied research only uses terms regarding observable properties of systems (variables such sizes, counts, proportions, variations, etc) and intuitions (which are just relations, often vague, between known or observable properties). Theoretical research uses abstractions, mental objects, and so, it’s much more general and fail proof. It’s not subject to the problem of [non mathematical] induction, as opposed to applied research.

http://www.palgrave.com/business/collis/br/docs/sample.pdf

This is a book chapter on an introduction about how to do research. I really loved some of the tables, even if they are mostly for reference. (I.e. they are not rigorous.)

Leave a Comment

Updated my web page!

I updated my web page at the University of Iowa! I am not sure whether it looks better than before, but it is at least more complete than before. I added more details on my research.

http://www.cs.uiowa.edu/~mmonsalv/

Leave a Comment

CompleNet 2012 and Jakobsen

I attended the Third Workshop On Complex Networks conference, informally known as CompleNet 2012. I gave a talk as an author, and it was the perfect beginning of Spring break for me! I went to Melbourne, Florida; think about it, I live in Iowa! Of course, going to Florida was awesome. I wandered around the place and attended the conference. And I got to discover an awesome university: Florida Tech (The Florida Institute of Technology). Florida Tech is truly a tropical university, full of palm trees and near the beach. It’s botanical garden is a jungle. I think that is awesome. And what I loved the most was its eatery area. Loved the healthy food.
CompleNet 2012 was awesome. I met some super productive researchers and scholars, who gave talks on very interesting research topics and results. There were some first year PhD students whom had already several (interesting) papers published. That is something!
Conferences are great to update oneself with the state of the art in research and to share ideas. Just last weekend I attended the Jakobsen conference here at the University of Iowa. It was nice, in spite that is not a formal conference; it does not have proceedings or anything of the sort. Yet one can use the opportunity to listen to other graduate students’ research and share ideas with them. The Jakobsen conference, however, is about all academic areas, so it is not about sharing ideas in a multidisciplinary topic but about sharing ideas across fields. This is equally great! On top of that, I met other Fulbright grantees there. This is sure a small world!

Leave a Comment

Doing research

Lately I have been busy writing papers (for publication) and analyzing some data. This is very good yet tiresome! And what is truly tiresome is to clean some unready data sets. Good thing that the following researchers (grad students probably) in the group will have access to a clean and readable data set, with a few alternative views so their research is easier.

I have also been doing many more things. (I cannot afford to get bored, uh?) Besides attending social meetings and events, and doing exercise, I have been working on my web page. It will be ready (and nice) very soon.

Leave a Comment

Happy holidays!

2011 has almost ended. This is the year I started to do research seriously. I will keep doing that, but I will focus more on writing papers. I have noticed that researching something is really fun and one learns a lot of things in the way, but without a short term research goal (to get an important intermediate publishable result), one could work and work until getting something big, but nothing publishable during the way. And that is not healthy. So, I decided to organize my research projects as composed of intermediate milestones which consist in publishable results. That slightly reduces the flexibility of one’s research path to get to a result, but it pays off. Now, before I start working on something, I will write a potential timeline with research questions and choose the direction that leads to the better paper, without deviating too much from the final (normally decided by a funding agency) research goal.

I also learned how to organize narratives to be more convincing through clarity. This is crucial to the researcher, because using inadequate narrative structures can weaken the relevance or convincingness of a paper. The general key to clarity is to remember that your argumental structure is not about ideas but guidance. A structure based on ideas would say that paragraphs are organized like [intro], [idea1], [idea2], …, [end]. A structure based on guidance would say that paragraphs are organized like [what+importance], [top level explanation of the text], [go to idea1, work on simpler case], [give example], [show limits of simplification], [idea1], &c. Each part of one’s narrative must have a communication goal; to communicate an idea is just one part of this. Of course, communication goals aren’t to be chosen freely. One must remember that one should try to focus on one thing at a time, to make the life of the receptors easier. This is the way of simplification: work on simpler and simpler objects, until one gets to the details… but there are many strategies for this. But also, one must provide control too: one must prevent receptors get distracted by using too much detail at a time. For example, a landscape may consist of a day sky, a forest and a mountain chain at the background. Then, one describes the clouds and birds in the sky, and the colors; the forest becomes a bunch of trees of different shapes; and the mountain chain becomes a set of peaks with snow, hills, inclinations, &c. But one’s work is to also provide guidance, so when one describes the landscape, one could start with the forest, talk about how it ends at the base of the mountains, talk about the mountains, then talk about how their peaks reach the sky, talk about the sky, and why not, finish by talking about the greatness of the sky and recall how the narrative journey started… I could truly elaborate on this, but I am happy to say that I learned that it was all about guidance! And that is the objective of communication, not ideas! A person talks to others to induce something on them. That is the natural basis of communication. But as coordination increases, people’s minds get synchronized and all seems to be about ideas, which is a false appreciation of the broader picture.

Leave a Comment

Presented at ISDS’11, Atlanta

I think my talk went very well. They told me that it was good, and I talked to a lot of interested people at the end of it.

My presentation was divided into three parts:

  1. Forecasting syphilis in the US, in the short term. The goal here was to forecast syphilis one month ahead, weekly, instead of by months, quarters or years. Probably, month level forecasting was fine too, but I went with weeks. I showed some results that correlated different states for forecasting. I also explained how this played with overfitting in a very intuitive way, supported with a real example.
  2. The finding of bellwether states. Within the covariates used in the forecasts, a few of them were very frequent. They were the bellwether states. Here I explained how I obtained them, and how they were related to syphilis correlation clusters.
  3. And some issues about data quality. This part was really important. If I already delivered two important results, namely relating syphilis among states and identifying key predictive states, then I now added something of seemingly much greater value: how to evaluate data quality. Everyone was having problems with this; syphilis data looked too heterogeneous and unpredictable to work with, but now I showed a simple method to evaluate the quality of the source, at the weekly level only.

Obviously, having delivered three valuable results, my presentation looked powerful. And I did not mess it up by using the wrong form. I focused on one observation at a time (at several levels: I used a top-bottom approach), I explained everything in a simple, clear, and strong way (so all followed in a logically fluent way), and the slides were support to help understand everything (again, thinking about focusing in one thing at a time, minimizing distractions, yet not so simple as not be understandable by themselves, as my English is far from perfect).

Leave a Comment

On PowerPoint and technical presentations

This is by Edward Tufte: http://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=0001yB&topic_id=1&topic=Ask+E.T.

There are more columns here: http://www.edwardtufte.com/bboard/q-and-a?topic_id=1 These are all a very fun read, and very instructive, as they are about how to present information the right way.

Leave a Comment

A book about closing the gap between research and policy

http://www.idrc.ca/EN/Resources/Publications/Pages/IDRCBookDetails.aspx?PublicationID=851

The book is titled “THE KNOWLEDGE TRANSLATION TOOLKIT: Bridging the Know–Do Gap: A Resource for Researchers”, by Bennett and Jessani. It is about closing the gap between science and policy, when it comes to applying scientific recommendations to social realities. It seems very simple yet very comprehensive too, in that it also deals with public relations, conferences and corruption. And it never loses its focus: how to do all that through knowledge management.

Leave a Comment

Older Posts »
Follow

Get every new post delivered to your Inbox.