Dimensions of Interpersonal Relationships: Corpus and Experiments

Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 2307–2316 Copenhagen, Denmark, September 7–11, 2017. c⃝2017 Association for Computational Linguistics Dimensions of Interpersonal Relationships: Corpus and Experiments Farzana Rashid and Eduardo Blanco Human Intelligence and Language Technologies Lab University of North Texas Denton, TX, 76203 farzanarashid@my.unt.edu, eduardo.blanco@unt.edu Abstract This paper presents a corpus and exper- iments to determine dimensions of inter- personal relationships. We define a set of dimensions heavily inspired by work in so- cial science. We create a corpus by retriev- ing pairs of people, and then annotating dimensions for their relationships. A cor- pus analysis shows that dimensions can be annotated reliably. Experimental results show that given a pair of people, values to dimensions can be assigned automatically. 1 Introduction The task of information extraction (IE) consists in creating structured representations from unstruc- tured text. These representations usually consist of relations explicitly stated in text, and involve two or more arguments. For example, IE sys- tems would extract SPOUSE(John, Mary) or MAR- RIED(John, Mary, 1994) from John and Mary have been married since 1994. IE systems have a long history, and became popular after eval- uations such as MUC (Grishman and Sundheim, 1996) and ACE (Doddington et al., 2004). Traditional IE systems are supervised and ex- tract relations defined before training takes place (Peng and McCallum, 2004). More recently, open IE systems have been proposed to extract all re- lations explicitly stated in text in an unsuper- vised manner and without defining relations a priori (Mausam et al., 2012). Regarding inter- personal relations—relations that take as argu- ments two people—both IE approaches extract re- lations such as RELATIVE, FRIEND and COMMU- NICATES WITH. Open IE systems are domain in- dependent and would extract, in principle, rela- tions such as CLASSMATES and ADVISOR from students’ diaries or biographies of scientists. While useful for applications such as ques- tion answering (Yao and Van Durme, 2014), these dyadic relations only provide a generic under- standing of the relationship between two people. For example, COMMUNICATES WITH may relate people who have an intense or superficial rela- tionship (e.g., engaged couples talking about wed- ding plans vs. home owners discussing remodels with contractors), pleasure- or task-oriented rela- tionships (e.g., friends planning a backpacking trip vs. software developers discussing the next deliv- ery), and may be spatially near or distant (e.g., inviduals having an in-person meeting vs. those exchanging emails or talking on the phone). These elemental properties of interpersonal re- lationships are called dimensions in social science, and have been studied for decades (Wish et al., 1976). In those studies, the goal is to understand how relationships are perceived by people, not to extract them. Note that unlike interpersonal re- lationships, their dimensions are usually implic- itly stated in text, thus extracting them is chal- lenging. Also, extracting dimensions of interper- sonal relationships requires text understanding be- yond the event in which two people participate. As shown above, two people who communicate may have different dimension values depending …

Content truncated to 500 words