SemEval 2015, Task 7: Diachronic Text Evaluation

Proceedings of the 9th International Workshop on Semantic Evaluation (SemEval 2015), pages 870–878, Denver, Colorado, June 4-5, 2015. c⃝2015 Association for Computational Linguistics SemEval 2015, Task 7: Diachronic Text Evaluation Octavian Popescu IBM Research Yorktown, NY 10598, USA o.popescu@us.ibm.com Carlo Strapparava FBK Povo, TN 38123, Italy strappa@fbk.eu Abstract In this paper we describe a novel task, namely the Diachronic Text Evaluation task. A cor- pus of snippets which contain relevant infor- mation for the time when the text was created is extracted from a large collection of newspa- pers published between 1700 and 2010. The task, subdivided in three subtasks, requires the automatic system to identify the time inter- val when the piece of news was written. The subtasks concern specific type of information that might be available in news. The intervals come in three grades: fine, medium and coarse according to their length. The systems par- ticipating in the tasks have proved that this a doable task with very interesting possible con- tinuations. 1 Introduction Language changes over the time, even over rela- tively small periods. For example, as the main in- tent of publishing newspapers is to disseminate in- formation to the population of a whole country, there is an objective pressure to impose a standard and to smooth over the dialectical differences. How- ever, since the late 1600s, each generation has read pieces of news containing new words, borrowed or invented, exhibiting new drifts in the meanings of old words, printed with different spelling etc. The examples (1), (2), (3) and (4) below exhibit a series of features which are useful to pin point the year when the respective piece of news was created. Well known global events, sense superseding, spe- cific spelling and new vocabulary entry words are all time relevant features. At a deeper level of anal- ysis, time is revealed also by the mentions of named entities, such as Security Council, the topic and the linguistic genre are also relevant features. 1. Dictator Saddam Hussein ordered his troops to march into Kuwait. After the invasion is condemned by the UN Se- curity Council, the US has forged a coalition with allies. Today American troops are sent to Saudi Arabia in Opera- tion Desert Shield, protecting Saudi Arabia from possible attack. circa 1990 2. We have cabled the English house from which we get it and expect a reply to-morrow. circa 1900 3. Occasional selfies are acceptable, but uploading a new picture of yourself every day is not necessary. circa 2014 4. .. .The House of Samuel Sandbroke was brokt and sev- eral Pistols discharged . .. Her Majesty, for the better Dis- covery of the Offenders, is pleased to promise Her most Gracious Pardon for the said Crime. circa 1705 While for humans it is relatively easy to notice the language differences between two texts, and even to be accurate in determining the period when a piece of news was written, for computational systems this task is challenging. On the other hand, with the …

Content truncated to 500 words