Blog

Catching Cyclical Citations

Citations are a big part of a researcher's popularity and credibility. The more citatons an author has, the more likely their work is impactful and influential. Therefore, it is not difficult to imagine that some authors might be tempted to artificially inflate their citation count.

Generally speaking, there are two well-known ways an author can inflate citation count:

Paper Factories - Paper factories are when researchers, often professors, amass a large number of PhD students, post-docs, and interns to constantly churn out papers, often with little to no original content. Even when the researcher contributed barely anything to the paper, they are still listed as a co-author, and thus are able to increase their citation count anytime the paper is cited. However, these are difficult to catch, as the author's contribution is often not defined, and even then, it would take time to read through each and every paper the author produced.

Citation Rings - Citation Rings are when researchers agree with others to cite each other - or in other words, collude with each other. This is also difficult to catch, because in can be hard to tell a genuine citation from a fabricated one.

However, there is a subset of Citation Rings which is relatively easy to track, which is Cyclical Citations. As the name implies, this is when a researcher cites a previous paper they wrote in the current paper. This is not inherently bad, as it is common for researchers to build upon their previous work. But when a researcher does this excessively, it can be a sign of inflating citation count.

In order to see how many of a researcher's citations are self-citations, I wrote a python script in a few hours that scrapes the researcher's Google Scholar page. After getting a list of publications, it retrieves the citations for each paper from semantic scholar. Then, it checks how many of the citations have the researcher listed as an author.

Writing the script was harder than I initially thought, mainly becuase Google Scholar heavily limits web-scrpaing ability. Using proxies, I was able to get around the limitations at the cost of speed. Hence the script can take between 3-10+ minutes to run, depending on the number of papers the researcher has.

The results are interesting. I won't name specific researchers, but I found that some researchers had over 10% of their citations as self-citations. I have released the beta version of the code on my github, and I encourage others to check it out if they suspect a researcher of inflating their citation count. I will work on an optimized version of the code in the future when I get more time.

Hacking into Warframe's Communication Servers

Warframe is a free-to-play space-themed third-person shooter developed by Digital Extremes, with millions of players across the globe. Not only is the game one of the most popular games ever created - consistently being in the top 30 active players for the past decade - but also my most played game.

Back when I used to code for fun, I'd often write c++ scripts to create tools for various things in my life such as music and chess, and Warframe was no exception. The habit persisted and eventually my suite of tools grew into excel damage calculators, java script discord bots, and python scripts for in-game image processing.

Thus it comes to no surprise that when I heard a rumor a few months ago that there was a vulnerability in Warframe's authentication process, I was instantly intruiged. Access into Warframe's communication server would allow for the creation of trading bot faster than any other bot on the market. In the past few years, multiple bots have been developed by various players (including me), but lack of internal access has made the bots limited and laggy.

Such an exploit would not affect gameplay, nor 99.99% of the player base. However, a bot with instant access to the server would provide a handful of players a trading edge over the ~200 or so players that trade high value items (often worth thousands of dollars) on a daily basis.

Unfortunately, the only person who knew about the vulnerability was a supposed 17-year old Russian hacker by the name Redlim, who was unwilling to share the details with anyone. This was expected as he was making a fortune charging $100/month for access to his bot.

[THIS ARTICLE IS IN PROGRESS]

Swagnik Roychoudhury

S.R

Blog