Takis

Presentation Slides: View Here

Office Hours:

Contact:

I'm Professor and Chair of Computer Science at Wellesley College, studying online social media, primarily related to the propagation of information and misinformation, prediction of political events, and in developing tools that help users evaluate the trustworthiness of information. In particular, with my Wellesley colleagues and students, we have been studying the problem of propaganda and online misinformation since 2002.

* The paper "Of course it is true; I saw it on the Internet: Critical thinking in the Internet Era" was the first to raise concerns about our challenges to evaluate the validity of what we read on the internet, many years before "fake news" became an issue. It pointed to the fact that people confuse the process of searching online with the critical thinking that is required to make sense of the results of the search.

* It was followed by the 2005 paper on "Web Spam, Propaganda and Trust" that shows the close relationship between propaganda in society and search engine manipulation (web spam) on the internet. It also describes how Google and other search engines have been manipulated by advertisers and propagandists since the early 2000's.

* A follow-up version, "Web Spam, Social Propaganda and the Evolution of Search Engine Rankings" shows that search engines are evolving their ranking methods to outsmart the propagandists and spammers. This is not an easy task, however, and the result is an arms race that most people are unaware of.

* Defending against propaganda is particularly important in democratic elections when propagandists can alter public perception of candidates and even flip elections to their benefit. Multiple evidence of such manipulation was uncovered during the 2010 MA special election for the seat of Ted Kennedy. In "From Obscurity to Prominence in Minutes: Political Speech and Real-Time Search" (WebScience 2010 Best Paper Prize) we reveal the first political "Twitter-bomb" by bots, and a Twitter-enabled Google bomb.

* The previous paper was the inspiration for many projects to defend against misinformation including Indiana University's Truthy and Botometer, and our TwitterTrails system. (Here is a short description on how TwitterTrails works.)

* In "The infamous "pizzagate" conspiracy theory: Insights from a TwitterTrails investigation" we study how the 2016 conspiracy theory developed on Twitter and identify who started the rumor and the role that pro-Erdogan Turkish trolls and journalists played in its spreading (see video).

* As a response to the early 2010's plethora of reports about predicting electoral results using social media, we published "Limits of Electoral Predictions" warning against the flaws of such predictions. The paper "How (Not) to Predict Elections" describes the conditions under which such predictions should be evaluated: One should predict BEFORE the elections; one should not do a post-election analysis on how one could have predicted!

Our electoral unpredictability "theorem" says that one cannot predict elections using just online social media data because the stakes are too high. If it were possible, spammers and propagandists would manipulate the data so that their side appears as predicted to win, which makes prediction impossible. We summarize our observations in a SCIENCE article entitled Social Media and the Elections

* More recently, in "The Fake News Spreading Plague: Was it Preventable?" we show how the method of spreading "fake news" on Facebook in 2016 was identical to the method used in the first political "Twitter-bomb" in 2010. The disturbing fact is that techniques such as the ones that the Russians and other propagandists deployed in 2016 were taught by Americans in 2010 and they can be redeployed at any time.

* Since the problem of online misinformation became obvious, much research is done in trying to develop a technical solution. As I explain in Technology, Propaganda, and the Limits of Human Intellect, purely technical will not work, they may even backfire. Education is still an essential component of any solution. Especially, epistemological education: Understanding why we believe what we believe.

More information about my current work is in "Separating Truth From Lies" and in my Publications, News coverage and Video/Audio pages.

In the past, I have done research in Visualization, Parallel Computing, Image Dithering and CS Education.

Current Positions:

Past Positions:

I studied Mathematics at the University of Athens and Computer Science at Brown. I received my M.Sc. and Ph.D. in Computer Science from Dartmouth, and have been a visiting scientist at MIT, the Sydney University in Australia, and at Harvard. Regarding professional associations, I'm a senior member of the ACM, and a member of LACS, IEEE Computer Society, SIGWEB, SIGCSE and SIGACT's Electronic Publication Board. I also serve on the program committees for the WWW (Social Networks Track), AAAI-Web, WEBIST, ICWI and ICWSM conferences.

My research has been supported by the National Science Foundation, Brachman-Hoffman Fellowships, and Wellesley College.

On a personal note, here is a digital copy of the book "History of the Metaxa Family from 1081 until 1864" (in Greek) by Epaminondas Metaxas, MD (Athens, 1893)