Accurate recommendation and reliable explanation are two key issues for modern recommender systems. However, most recommendation benchmarks only concern the prediction of user-item ratings while omitting the underlying causes behind the ratings. For example, the widely-used Yahoo!R3 dataset contains little information on the causes of the user-movie ratings. A solution could be to conduct surveys and require the users to provide such information. In practice, the user surveys can hardly avoid compliance issues and sparse user responses, which greatly hinders the exploration of causality-based recommendation. To better support the studies of causal inference and further explanations in recommender systems, we propose a novel semi-synthetic data generation framework for recommender systems where causal graphical models with missingness are employed to describe the causal mechanism of practical recommendation scenarios. To illustrate the use of our framework, we construct a semi-synthetic dataset with Causal Tags And Ratings (CTAR), based on the movies as well as their descriptive tags and rating information collected from a famous movie rating website. Using the collected data and the causal graph, the user-item-ratings and their corresponding user-item-tags are automatically generated, which provides the reasons (selected tags) why the user rates the items. Descriptive statistics and baseline results regarding the CTAR dataset are also reported. The proposed data generation framework is not limited to recommendation, and the released APIs can be used to generate customized datasets for other research tasks.
This study explores how L1 and L2 Chinese speakers use world knowledge and classifier information to predict fine-grained referent features. In a visual-world-paradigm eye-tracking experiment, participants were presented with two visual objects that were denoted by the same noun in Chinese but matched different shape classifiers. Meanwhile, they heard sentences containing world knowledge triggering context and classifiers. The effect of world knowledge has been differentiated from word-level associations. Native speakers generated anticipations about the shape/state features of the referents at an early processing stage and quickly integrated linguistic information with world knowledge upon hearing the classifiers. In contrast, L2 speakers show delayed, reduced anticipation based on world knowledge and minimal use of classifier cues. The findings reveal different cue-weighting strategies in L1 and L2 processing. Specifically, L2 speakers whose first languages lack obligatory classifiers do not employ classifier cues in a timely manner, even though the semantic meanings of shape classifiers are accessible to them. No evidence supports over-reliance on world knowledge in L2 processing. This study contributes to the understanding of L2 real-time processing, particularly in L2 speakers’ utility of linguistic and non-linguistic information in anticipating fine-grained referent features.