Style Guide

The goal of this wiki is to provide a comprehensive survey of all publications in statistical machine translation. Anybody may edit it.

Publications — Each publication should be mentioned at least once in the description of the appropriate topic. Each contribution of a puplication should be summarized in one sentence, or grouped alongside very similar contributions.

Topics — You may add research topics. Each topic should be described in a number of sections:

• introductory definition of the topic
• review of all publications about this topic
• benchmarks used to evaluate the challenges posed by the topic (standard data sets and evaluation metrics, evaluation campaigns)
• discussion: what are the main fundings, hat are the open problems, how relevant is the topic today?
• reference to related topics
• new publications that have not yet been integrated

The topics are grouped in a hierarchical taxonomy, but this is mostly done to provide a coherent structure that allows for an overview of all topics. Topics have relationships that cross this hierarchy and often some work is relevant for more than one topic, and hence should be discussed multiple times.

Citations — The database of publications is maintained as a BibTeX file, which may be downloaded. Some of the BibTeX entries contain a URL of the pdf file of the publication. We aim to include such links for all publications. Citations of publications is close to LaTeX style using (:cite BibTex-ID :) for Author (Year) and (:citep BibTex-ID :) for (Author, Year). Muliple citation can be grouped together, their identifiers are to be separated by spaces.