Diversifying Personalized Recommendation with User-session Context

Abstract

Recommender systems (RS) have become an integral part of our daily life. However, most current RS often repeatedly recommend items to users with similar profiles. We argue that recommendation should be diversified by leveraging session contexts with personalized user profiles. For this, current session-based RS (SBRS) often assume a rigidly ordered sequence over data which does not fit in many real-world cases. Moreover, personalization is often omitted in current SBRS. Accordingly, a personalized SBRS over relaxedly ordered user-session contexts is more pragmatic. In doing so, deep-structured models tend to be too complex to serve for online SBRS owing to the large number of users and items. Therefore, we design an efficient SBRS with shallow wide-in-wide-out networks, inspired by the successful experience in modern language modelings. The experiments on a real-world e-commerce dataset show the superiority of our model over the state-of-the-art methods.

An Attention-Based Recommender System to Predict Contextual Intent Based on Choice Histories across and within Sessions

“…There are many existing techniques that investigate the combination of the user's long-term preference and the short-term intent [13,[21][22][23][24][25][26][27]. In addition, the user's preference for items constantly evolves over time [28], while these works implicitly assume the user preference to be stationary, which seems to be an unrealistic assumption in many scenarios, particularly in the news or shopping related scenarios [5][6][7]13,29]. Therefore, such methods perform poorly when preferences, in fact, are context-sensitive and transient.…”

“…Therefore, such methods perform poorly when preferences, in fact, are context-sensitive and transient. There are session-based approaches [7,[29][30][31][32] that focus on the interest drifts, to solve the problem of the static user preference and intent, while these methods assume that each item in one session has the same influence, which is usually an invalid assumption, especially when considering a user's interactions with items of diverse characteristics. One of the main attractions of search-engine advertising is that search engines take into account the "current interest" of the user and thereby can deliver the right message, ad, or product just at the right time.…”

ragamAI: A Network Based Recommender System to Arrange a Indian Classical Music Concert

“…The proposed ragamAI framework relies on the influence of two separate models: 1. a deep attention model to capture the importance of sequence of ragams, and 2. an embedding model to capture the importance of hand picked features to train. Unlike other methods, which predict next event or item in a given sequence [6], [9], [17]- [19], the proposed model(s) Figure 1 A. Raaga network Networks or graphs have been considered as a promising framework to study variety of applications like influence modeling [20], community detection [21], and recommender systems [22]. Their organization of nodes and edges help to study the structural organization and positional values of entities (nodes and communities).…”

“…FPMC models the user preference on a list and the preference transitions to recommend next set of items. • SWIWO [17]: Session-based Wide In Wide Out(SWIWO) is a deep learning based recommender model that uses a both user and item based feature set to predict the next item. To adapt to our problem, we eliminate the user feature section and add a softmax layer to predict the score of items for a given sequence.…”

Sequential Recommender Systems: Challenges, Progress and Prospects

“…In the real world, some user-item interaction sequences are strictly ordered while others may not be, namely, not all adjacent interactions are sequentially dependent in a sequence [4]. For instance, in a shopping sequence S2 = {milk, butter, flour}, it does not matter whether to buy milk or butter first, but the purchase of both items leads to a higher probability of buying flour next; namely, there is no strict order between milk and butter, but flour sequentially depends on the union of them.…”