MobileConvRec: A Conversational Dataset for Mobile Apps Recommendations

TL;DR

This work introduces MobileConvRec, a large-scale benchmark that unifies historical mobile app interactions with multi-turn conversational data to advance conversational mobile app recommendations. It builds the dataset by transforming the MobileRec sequential data into grounded dialogues via a theoretical framework that leverages global user preferences and aspect-based sampling, then realizes natural language with off-the-shelf LLMs. MobileConvRec comprises 12.2K dialogs, 156K turns, 11.8K users, and 1,730 apps across 45 Google Play categories, enriched with rich app metadata and APKs for security analyses. Baseline experiments with GPT-2 and Flan-T5 demonstrate the dataset’s utility for evaluating both generation and ranking in conversational mobile app recommendation settings and establish reference performance for future work.

Abstract

Existing recommendation systems have focused on two paradigms: 1- historical user-item interaction-based recommendations and 2- conversational recommendations. Conversational recommendation systems facilitate natural language dialogues between users and the system, allowing the system to solicit users' explicit needs while enabling users to inquire about recommendations and provide feedback. Due to substantial advancements in natural language processing, conversational recommendation systems have gained prominence. Existing conversational recommendation datasets have greatly facilitated research in their respective domains. Despite the exponential growth in mobile users and apps in recent years, research in conversational mobile app recommender systems has faced substantial constraints. This limitation can primarily be attributed to the lack of high-quality benchmark datasets specifically tailored for mobile apps. To facilitate research for conversational mobile app recommendations, we introduce MobileConvRec. MobileConvRec simulates conversations by leveraging real user interactions with mobile apps on the Google Play store, originally captured in large-scale mobile app recommendation dataset MobileRec. The proposed conversational recommendation dataset synergizes sequential user-item interactions, which reflect implicit user preferences, with comprehensive multi-turn conversations to effectively grasp explicit user needs. MobileConvRec consists of over 12K multi-turn recommendation-related conversations spanning 45 app categories. Moreover, MobileConvRec presents rich metadata for each app such as permissions data, security and privacy-related information, and binary executables of apps, among others. We demonstrate that MobileConvRec can serve as an excellent testbed for conversational mobile app recommendation through a comparative study of several pre-trained large language models.

Figures (5)

• Figure 1: A sample natural language dialog between the user and the system might unfold as follows: The system draws insights from the user's historical interactions, proactively elicits the user's needs explicitly, and synergizes this information to make more meaningful recommendations. Additionally, the user may pose follow-up questions regarding the recommended app.
• Figure 2: Overview of the framework: It transforms a conventional sequential recommendation dataset into a conversational recommendation dataset.
• Figure 3: Top-10 categories in the dataset.
• Figure 4: The distribution of the number of turns per dialog.
• Figure 5: The distribution of the number of words per turn.