Language in a (Search) Box: Grounding Language Learning in Real-World Human-Machine Interaction

Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4409–4415 June 6–11, 2021. ©2021 Association for Computational Linguistics 4409 Language in a (Search) Box: Grounding Language Learning in Real-World Human-Machine Interaction Federico Bianchi∗ Bocconi University Milano, Italy f.bianchi@unibocconi.it Ciro Greco Coveo Labs New York, USA cgreco@coveo.com Jacopo Tagliabue Coveo Labs New York, USA jtagliabue@coveo.com Abstract We investigate grounded language learning through real-world data, by modelling a teacher-learner dynamics through the natu- ral interactions occurring between users and search engines; in particular, we explore the emergence of semantic generalization from un- supervised dense representations outside of synthetic environments. A grounding domain, a denotation function and a composition func- tion are learned from user data only. We show how the resulting semantics for noun phrases exhibits compositional properties while be- ing fully learnable without any explicit la- belling. We benchmark our grounded seman- tics on compositionality and zero-shot infer- ence tasks, and we show that it provides better results and better generalizations than SOTA non-grounded models, such as word2vec and BERT. 1 Introduction Most SOTA models in NLP are only intra- textual. Models based on distributional seman- tics – such as standard and contextual word em- beddings (Mikolov et al., 2013; Peters et al., 2018; Devlin et al., 2019) – learn representations of word meaning from patterns of co-occurrence in big cor- pora, with no reference to extra-linguistic entities. While successful in a range of cases, this ap- proach does not take into consideration two fun- damental facts about language. The first is that language is a referential device used to refer to extra-linguistic objects. Scholarly work in psy- cholinguistics (Xu and Tenenbaum, 2000), formal semantics (Chierchia and McConnell-Ginet, 2000) and philosophy of language (Quine, 1960) show that (at least some aspects of) linguistic meaning can be represented as a sort of mapping between lin- guistic and extra-linguistic entities. The second is ∗Corresponding author. Authors contributed equally and are listed alphabetically. that language may be learned based on its usage and that learners draw part of their generalizations from the observation of teachers’ behaviour (Tomasello, 2003). These ideas have been recently explored by work in grounded language learning, showing that allowing artificial agents to access human ac- tions providing information on language meaning has several practical and scientific advantages (Yu et al., 2018; Chevalier-Boisvert et al., 2019). While most of the work in this area uses toy worlds and synthetic linguistic data, we explore grounded language learning offering an example in which unsupervised learning is combined with a language-independent grounding domain in a real- world scenario. In particular, we propose to use the interaction of users with a search engine as a setting for grounded language learning. In our setting, users produce search queries to find products on the web: queries and clicks on search results are used as a model for the teacher-learner dynamics. We summarize the contributions of our work as follows: 1. we provide a grounding domain composed …

Content truncated to 500 words