A Computational Model of Human Preferences for Pronoun Resolution

Abstract

We present a cognitive computational model of pronoun resolution that reproduces the human interpretation preferences of the Subject Assignment Strategy and the Parallel Function Strategy. Our model relies on a probabilistic pronoun resolution system trained on corpus data. Factors influencing pronoun resolution are represented as features weighted by their relative importance. The importance the model gives to the preferences is in line with psycholinguistic studies. We demonstrate the cognitive plausibility of the model by running it on experimental items and simulating antecedent choice and reading times of human participants. Our model can be used as a new means to study pronoun resolution, because it captures the interaction of preferences.

Discourse‐level adaptation in pronoun comprehension

“…If so, this would modulate the interpretation of ambiguous pronouns, and lead to predictions about pronouns that could facilitate or inhibit comprehension. Indeed, Seminck and Amsili (2017) created a computational model of pronoun resolution that learnt two human biases (subject bias and parallel function bias) on the basis of the statistics exhibited in a corpus. (See Table 1 for a list of possible frequency metrics at the discourse level).…”

The Past, Present and Better Future of Feedback Learning in Large Language Models for Subjective Human Preferences and Values

“…Ficler and Goldberg (2017) focus on modulating formality depending on context, while others focus on the personalisation of language models, such as reflecting author personality in machine translation Mirkin and Meunier, 2015;Rabinovich et al, 2017); providing financial recommendations via chat bots (Den Hengst et al, 2019); or enabling customised online shopping (Mo et al, 2016). Most studies target human preferences assumed to be commonly-held and stable, such as word order (Futrell and Levy, 2019), sense making (De Deyne et al, 2016;Seminck and Amsili, 2017) and vocabulary matching (Campano et al, 2014;Dubuisson Duplessis et al, 2017). In contrast, Nguyen et al (2017) and Kreutzer et al (2017) acknowledge the noisiness of human feedback but attempt to extract a single, unified preference.…”