Monday, Sep. 08, 2008

Lost in Translation

In 2005, three MIT graduate students skeptical of the standards at some academic conferences wrote a computer program to randomly generate papers, complete with charts and diagrams. They mad-libbed academic jargon — which can sound like gibberish — into sentences that literally were gibberish, such as, "The model for our heuristic consists of four independent components: simulated annealing, active networks, flexible modalities, and the study of reinforcement learning." They even included fictitious citations. The students submitted two of the papers to the World Multi-Conference on Systemics, Cybernetics and Informantics — and one of them was accepted. Conference organizers, who insisted the paper was one of a small number approved without being reviewed first, subsequently revised their acceptance procedures.