What does BERT know about books, movies and music? Probing BERT for Conversational Recommendation

Abstract

Heavily pre-trained transformer models such as BERT have recently shown to be remarkably powerful at language modelling by achieving impressive results on numerous downstream tasks. It has also been shown that they are able to implicitly store factual knowledge in their parameters after pre-training. Understanding what the pre-training procedure of LMs actually learns is a crucial step for using and improving them for Conversational Recommender Systems (CRS). We first study how much off-the-shelf pre-trained BERT "knows" about recommendation items such as books, movies and music. In order to analyze the knowledge stored in BERT's parameters, we use different probes (i.e., tasks to examine a trained model regarding certain properties) that require different types of knowledge to solve, namely content-based and collaborative-based. Content-based knowledge is knowledge that requires the model to match the titles of items with their content information, such as textual descriptions and genres. In contrast, collaborative-based knowledge requires the model to match items with similar ones, according to community interactions such as ratings. Both are important types of knowledge for a CRS-a system that should ideally be able to explain recommendations, match a user's textual description of their information need to relevant items, and elucidate a user's interests. We resort to BERT's Masked Language Modelling (MLM) head to probe its knowledge about the genre of items, with cloze style prompts. In addition, we employ BERT's Next Sentence Prediction (NSP) head and representations' similarity (SIM) to compare relevant and non-relevant search and recommendation query-document inputs to explore whether BERT can, without any fine-tuning, rank relevant items first. The insights we gain from these probes help us understand BERT's limitations and strengths for conversational recommendation. Finally, we study how BERT performs in a conversational recommendation downstream task. To this end, we fine-tune BERT to act as a retrieval-based conversational recommender system. Overall, our analyses and experiments show that: (i) BERT has knowledge stored in its parameters about the content of books, movies and music; (ii) it has more content-based knowledge than collaborative-based knowledge; and (iii) fails on conversational recommendation when faced with adversarial data.

What BERTs and GPTs know about your brand? Probing contextual language models for affect associations

“…Similarly, (Petroni et al, 2019) presents an in-depth analysis of relational knowledge present in pre-trained LMs. Penha and Hauff (2020) probe the contextual LMs (BERT and RoBERTa) for the conversational recommendation of books, movies, and music. Our work seeks to apply the idea of probing to a relatively unexplored area of affect analysis.…”

What BERT Based Language Model Learns in Spoken Transcripts: An Empirical Study

“…There has also been research on probing LM on application specific representations such as question answering (van Aken et al, 2019), information retrieval (Yilmaz et al, 2019), recommendation systems (Penha and Hauff, 2020), dialog systems (Wu and Xiong, 2020) etc. The entire spectrum of these works aims to understand the learning capability and properties encoded in the LM along with discovering their shortcomings.…”

NUANCED: Natural Utterance Annotation for Nuanced Conversation with Estimated Distributions

“…Conversational recommendation focus on combining the recommendation system with online conversation to capture user preference (Fu et al, 2020;Sun and Zhang, 2018;. Previous works mostly focus on learning the agent side policy to ask the right questions and make accurate recommendations, such as Li et al, 2020;Penha and Hauff, 2020). Chit-Chat (Adiwardana et al, 2020;Roller et al, 2020) is the most free form dialogue but almost with no knowledge grounding or state tracking.…”