IXAGroupEHUDiac: A Multiple Approach System towards the Diachronic Evaluation of Texts

Abstract

This paper presents our contribution to the SemEval-2015 Task 7. The task was subdivided into three subtasks that consisted of automatically identifying the time period when a piece of news was written (1,2) as well as automatically determining whether a specific phrase in a sentence is relevant or not for a given period of time (3). Our system tackles the resolution of all three subtasks. With this purpose in mind multiple approaches are undertaken that use resources such as Wikipedia or Google NGrams. Final results are obtained by combining the output from all approaches. The texts used for the task are written in English and range from the years 1700 to 2000.

A Bayesian Model of Diachronic Meaning Change

“…8 AMBRA (Zampieri et al, 2015) adopts a learning-to-rank modeling approach and uses several stylistic, grammatical, and lexical features. IXA (Salaberri et al, 2015) uses a combination of approaches to determine the period of time in which a piece of news was written. This involves searching for specific mentions of time within the text, searching for named entities present in the text and then establishing their reference time by linking these to Wikipedia, using Google n-grams, and linguistic features indicative of language change.…”

Interactive System for Reasoning about Document Age

“…On the other hand, the latest year is the one at which the ngram has been used for the last time with the frequency higher than . With a sufficiently low value of 11 , the oldest and the latest years of an ngram can be considered as the boundaries of the time period when the ngram has been in a relatively common use. Naturally, sometimes ngram may have lower frequency than within that time period, however, for simplicity, we assume the continuity of ngram use within its left and right boundaries.…”

“…These are selected as the minimum oldest 10 Called "Cut-off year view" in the system. 11 Currently, the value is set by the user. In the future we plan to offer automatic derivation of based on corpus-derived statistics.…”

“…Popescu and Strapparava [10] for example, have organized the Diachronic Text Evaluation challenge under the umbrella of SemEval workshop series to foster the development of algorithms capable of identifying the time interval a given document was published in. A number of research articles have also been published based on heuristic methods [3,6,7], language models [2,4,5] or utilizing diachronic frequencies of words [1,11] to determine document age. Despite these advances, still there is lack of a system for computing document age in an interactive way.…”

Modeling Language Change in Historical Corpora: The Case of Portuguese

“…The task proved to be a very challenging one 1 Results and methods are described in detail in the shared task report (Popescu and Strapparava, 2015). and the only team to participate in all three sub-tasks was the IXA team (Salaberri et al, 2015) who used external resources such as Google N-grams and Wikipedia Entity Linking to accomplish the task. The best performing system in the DTE task was the UCD team (Szymanski and Lynch, 2015) who achieved 54.2% precision in identifying the publication date of texts in an interval of 20 years (subtask 2) using Support Vector Machine (SVM).…”