MIT alums celebrate 10th anniversary of bogus CompSci paper generator with cheeky new tool

Creators of SCIgen computer science paper generator return with SCIpher tool for sending phony calls for papers from fictitious conferences


Left to right: Dan Aguayo, Max Krohn, and Jeremy Stribling in 2005

Credit: Photo by Frank Dabek

Three MIT grads this week are celebrating the 10th anniversary of their clever SCIgen program, which randomly generates computer science papers realistic enough to get accepted by sketchy technical conferences and publishers, with a brand new tool designed to poke even more fun at such outfits.

Just a bit late for April Fool’s Day, the new SCIpher program from the MIT Computer Science and Artificial Intelligence Lab alums enables users to hide messages inside randomly-generated calls for papers from phony conferences whose names are so ridiculous that they sound legit. An MIT spokesman says the new tool is really just a way for geeky friends to mess with each other.

For instance, pop the following text into the SCIpher decoder to find out my secret message:

The First Annual AVQYK Symposium on linear-time, vertical sharing economy

Dear list owner and all!

Many information theorists would agree that, had it not been for cloud-based models, the understanding of thin clients might never have occurred. To put this in perspective, consider the fact that famous information theorists mostly use hierarchical databases to solve this issue. Without a doubt, two properties make this approach perfect: the new framework of scholars locates superpages, and also the new framework of security experts allows perfect technology. Thus, IPv4 and e-commerce offer a viable alternative to the understanding of IPv4.

The objective of this conference is to supply a seminar for surmounting the significant obstacles in the development, exploration, and improvement of atomic algorithms and scalable archetypes. Without a doubt, original papers are released on omniscient virtualization, perfect user interface design, and cloud-based natural language processing. The subject of AVQYK is ' interfering flexible services and modular configurations for experts ', sharing the convergence of cloud-based information, mobile epistemologies, and IPv7 in proving perfect frameworks of parallel cryptography. Thus AVQYK provides innovative, half-baked, and forward-thinking submissions on arguing any event-driven methods to all aspects covering the motif of this conference.

* Camila Hampton - University of Nebraska-Lincoln
A understanding of the Turing machine
* Prof. Oliver Knight - Ohio State University
Understanding of neural networks
* Assistant Professor Derek Lowe - University of British Columbia
A methodology for the understanding of interrupts
* Prof. Alfie Mahajan - Hasselt University
On the natural unification of information retrieval systems and hierarchical databases
* Ray Guzman - Korea University
A robust unification of systems and agents
* Caryn Tapia - University of Arkansas - Fayetteville
SMPs now considered harmful
* Billy Contreras - Kyoto Prefectural University of Medicine
The World Wide Web no longer considered harmful
* Artem Knapp - National Central University
A case for operating systems
* Terrence Schwartz - University of Texas Medical Branch at Galveston
Randomized algorithms now considered harmful
* Cheri Mittal - Jagiellonian University
A unproven unification of kernels and consistent hashing

Ambimorphic computer vision
Logical e-learning
Fuzzy software engineering, and exhaustive computer vision
Extremely partitioned hardware and architecture
Relational internet of things

Steering Committee:
Professor Rosalind Horne, University of Technology Sydney
Dr. Maxim Medina, University of British Columbia
Dylan Hartman, Royal Veterinary College University Of London

AVQYK in previous years:
University of Twente
Benxi, China

Advisor Committee:
Prof. Jerald Ranga (Louisiana State University)
Aleksandra Gillespie (State University of New York Upstate Medical University)
Jeremiah Harrell (University of Essex)

General Co-Chairs:
Lecturer Toby Hatfield - Medical College of Wisconsin
Graham Hawkins - Wageningen University and Research Centre

Important dates:
June 6, 2015: works due
June 28, 2015: notification of acceptance
July 17, 2015: final submissions due
July 25, 2015: conference date

We are giving you this call for papers, assuming that you will consider submitting abstracts to this special issue. As a guideline, only research communications will be considered (no works). By comparison, submissions of revisions, abstracts and abstracts are also taken.

While SCIpher is intended as just a bit of fun, SCIgen was a low-budget effort to point out serious flaws in the world of academic journal publishers and conferences, which pepper researchers with calls for papers and charge for content they often clearly don't read before accepting.

Jeremy Stribling MS ’05 PhD ’09 (now at crypto company Keybase), Dan Aguayo ’01 MEng ’02 (now at Meraki) and Max Krohn PhD ’08 (now runs Keybase) back in 2005 developed SCIgen over the span of just a couple weeks. Despite their rush job, the program was good enough to produce seemingly real compsci papers that include graphs, figures and citations. MIT's press office describes it as being almost like a Mad Libs for academic papers, and notes that it stemmed from work by Krohn at online study guide SparkNotes. (I could swear half the press releases we receive are generated by SCIgen...)

In April of 2005 the team’s submission, Rooter: A Methodology for the Typical Unification of Access Points and Redundancy,” was accepted as a non-reviewed paper to the World Multiconference on Systemics, Cybernetics and Informatics (WMSCI). After the hoax was revealed, the conference nixed the team's invite, but the students raised funds and showed up anyway -- armed with fake names, business cards and 'staches -- and presented on a variety of bogus topics.

The impact of SCIgen has been real. The IEEE wound up pulling its sponsorship from WMSCI and worked with Spring Publishing to remove nonsensical papers from their sites (you can read Springer's paper ON the papers here). And Springer recently released SciDetect, an open-source tool for spotting SCIgen papers.

“Our initial intention was simply to get back at these people who were spamming us and to maybe make people more cognizant of these practices,” says Stribling, in a statement. “We accomplished our goal way better than we expected to.”

Look to Reddit's Ask Me Anything (AMA) section today (April 14, 2pm EST) for a Q&A session with the MIT grads.

Enjoy a video recap of the SCIgen adventure, with the students donning disguises, below:

This story, "MIT alums celebrate 10th anniversary of bogus CompSci paper generator with cheeky new tool" was originally published by Network World.

ITWorld DealPost: The best in tech deals and discounts.