The First Workshop on Context Representation in User Modeling

This position paper aims to encourage researchers in the field of context-aware public transport information systems to incorporate human-centred approaches more deeply into their methodologies. Current context-aware systems in this domain often take a representational view and employ a data-first approach. Drawing on insights from previous work, we propose a distinction between the objective context and the experienced context. The experienced context incorporates interactions and perceptions to reflect better how we, as humans, experience the world. To measure this experienced context, we advocate for using qualitative research methods for HCI. To demonstrate this approach, we present the results of a focus group study on context in public transport. The results reveal that emerging experiences are shaped by a combination of various factors. These findings highlight the importance of incorporating user perspectives in designing context-aware systems. Therefore, in this paper, we take the position that if we want to improve the context-aware public transport information systems, we need to understand what travellers truly experience during their journey.

The ability to initiate gait involves a complex coordination between posture and movement, known as anticipatory postural adjustments (APAs). The emotional context in which gait initiation occurs can impact several spatio-temporal parameters, particularly the duration of APAs. While previous studies have used biologically relevant stimuli to induce emotions, such as images of pleasant or unpleasant scenes, to the best of our knowledge, the impact of the emotional context induced by music on gait initiation has not been explored yet. This paper presents a new dataset collected to study this impact. Objective biomechanical and physiological data were collected from participants during and after music listening, and subjective emotional responses were assessed using questionnaires. We also focused on two factors, liking judgment and familiarity, known to modulate emotions. Our preliminary analyses shows the impact of the emotional context induced by music on gait initiation, and confirms the strong importance of liking judgment and familiarity on the emotional context.

Monitoring and predicting user engagement is important to gauge the overall health of an E-commerce platform. A healthy active user-base indicates that the platform is able to retain users and is performing well on the user satisfaction metric. To measure the long-term user satisfaction, predicting return rate of a user is important. A frequent return of the user indicates that they are overall satisfied with the platform. To this end, we consider the problem of predicting user's return time on the platform given their historical interactions. The current state-of-the-art models for user return time prediction are based on recurrent neural network, which models the sequence of user interactions and predicts the return time using a Temporal Point Process based formulation. However, it is well-known that the inference time for these models grow as the sequence length increases, due to the complex recurrent and gating mechanisms, which deems them unfit to be used in a real-time prediction setting. Towards this end, we propose a lightweight and simple neural bag-of-words based model for user return time prediction, which considers the user activity trail as a bag-of-words embedding model and performs simple aggregation operation, used for the prediction. We perform experiments on interaction log data from a major e-commerce company, and the proposed bag-of-words model outperforms the complex recurrence based neural network by 6.14% and 4.81% on average, in terms of the Root-Mean-Squared-Error and Mean-Absolute-Error, respectively. We also compare the inference time of our method to recurrent neural network based method, with an overall reduction of 78.5% in terms of the wall-clock time.

Recent studies show that recommender system can be improved by using transformer-based approaches. Such models attempt to predict the next item based on the user's history of past interactions. Most of the recently proposed methods only consider a sequence of items and do not consider the moment of prediction. Although time-aware models have been developed in recent years, this area of research appears unstructured and difficult to replicate. In this paper, we demonstrate a simple yet effective method for making most next-item recommenders time-aware. We provide extensive experiments on two commonly used sequential recommenders, namely GRU4Rec and TiSASRec. Our results on four real-world datasets demonstrate the effectiveness of the proposed approach.