In app recommendation, user cold-start remains a fundamental challenge in recommender systems. Existing approaches primarily focus on efficiently leveraging limited data or transferring knowledge from active users to alleviate the user cold-start problem, yet they often overlook the influence of feature interactions on user cold-start. We group features according to their semantic types and identify an interesting phenomenon: user attribute features and behavioral sequence features interfere with each other, thereby constraining the model's ability to represent cold-start users effectively. We attribute this issue to differences in the latent space distributions and learning complexities of the two feature types, which hinder the model from accurately capturing cold-start users' interests. To address this challenge, we propose the AFIM architecture, which decouples the learning of user attribute and behavior sequential features. AFIM leverages a lightweight attention module to explicitly capture user interests from behavioral sequences, thereby reducing the learning burden on downstream recommendation networks. Additionally, it incorporates feature decoupling and dynamic fusion modules to mitigate learning bias arising from heterogeneous feature spaces. Extensive experiments on two public datasets and two industrial datasets demonstrate that AFIM consistently outperforms SOTA baselines, highlighting its effectiveness in user cold-start scenarios.

Message passing scheme is a general idea for Graph Neural Networks (GNNs) to learn node representations. During message passing, given a target node, we transform and aggregate the feature vectors of its neighbors and generate a representation vector for the target node. However, real-world graph data is usually constructed from complicated scenarios based on manually pre-defined rules; it is often the case that noisy information gets involved in message passing, thereby resulting in sub-optimal performance for GNNs and also impacting their trustworthiness and reliability. In this study, we present an effective learnable threshold technique that explicitly optimizes heterogeneous graph structure with the goal to maximize performance improvement of GNNs for downstream tasks. We give an explanation about the design of the learnable threshold and show the ability that our model can be applied to large-scale graphs. Experiments on seven datasets show that our model has a powerful ability to deal with homogeneous graphs with low homophily ratio and dense graphs. With the verification of robustness analysis, our model can resist the noisy information, which proves the robustness of our model.

Carpooling Route Planning(CRP) has become an important issue with the growth of low-carbon traffic systems. We investigate a novel, meaningful and challenging scenario for CRP in industry, calledMulti-Candidate Carpooling Route Planning(MCRP) problem, where each passenger may have several potential positions to get on and off the car. We surprisingly notice that this problem can be easily generalized for similar services such as express, takeout, or crowdsensing services, which means MCRP is a new fundamental combinatorial optimization problem. Traditional graph search algorithms or indexing methods are usually time and space consuming or perform poorly, which are not suitable for solving the problem. In this paper, we propose an end-to-end encoder-decoder model to plan a route for each many-to-one carpooling order with various data-driven mechanisms such as graph partitioning and feature crossover. The encoder is a filter-integrated Graph Convolution Network with external information fusion combining a supervised pre-training classification task, while the latter mimics a pointer network with a rule-based mask mechanism and a domain feature crossover module. We validate the effectiveness and efficiency of our model based on both synthetic and real-world datasets.

Graph Neural Networks (GNNs) have seen significant achievements in semi-supervised node classification. Yet, their efficacy often hinges on access to high-quality labeled node samples, which may not always be available in real-world scenarios. While active learning is commonly employed across various domains to pinpoint and label high-quality samples based on data features, graph data present unique challenges due to their intrinsic structures that render nodes non-i.i.d. Furthermore, biases emerge from the positioning of labeled nodes; for instance, nodes closer to the labeled counterparts often yield better performance. To better leverage graph structure and mitigate structural bias in active learning, we present a unified optimization framework (SCARCE), which is also easily incorporated with node features. Extensive experiments demonstrate that the proposed method not only improves the GNNs performance but also paves the way for more fair results.

Uncovering temporal dependency patterns behind event sequences plays a key role in predicting event types and event times. Recently, transformers based models have been used to describe point processes, such as the Transformer Hawkes Processes (THP models). However, existing THP models assume that data are collected in a central server and can be always seen during model training. Indeed, event sequence data are often located at different data centers which can not be shared directly due to the risk of privacy leakage. To this end, we combine in this paper the THP models with federated learning, enabling collaborative learning from a large amount of distributed event sequence data. Experiments show that our approach surpasses single data source training while preserving data privacy. For clients lacking certain types of event sequence data, our method performs much more stable than previous centralized training models.

Link prediction, which aims to forecast unseen connections in graphs, is a fundamental task in graph machine learning. Heuristic methods, leveraging a range of different pairwise measures such as common neighbors and shortest paths, often rival the performance of vanilla Graph Neural Networks (GNNs). Therefore, recent advancements in GNNs for link prediction (GNN4LP) have primarily focused on integrating one or a few types of pairwise information. In this work, we reveal that different node pairs within the same dataset necessitate varied pairwise information for accurate prediction and models that only apply the same pairwise information uniformly could achieve suboptimal performance. As a result, we propose a simple mixture of experts model Link-MoE for link prediction. Link-MoE utilizes various GNNs as experts and strategically selects the appropriate expert for each node pair based on various types of pairwise information. Experimental results across diverse real-world datasets demonstrate substantial performance improvement from Link-MoE. Notably, Link-Mo achieves a relative improvement of 18.71% on the MRR metric for the Pubmed dataset and 9.59% on the Hits@100 metric for the ogbl-ppa dataset, compared to the best baselines. The code is available at https://github.com/ml-ml/Link-MoE/.

Existing online advertising systems employ separate models for each task and site, resulting in a large number of models that require significant computing power and human effort to train and deploy. Moreover, separate models have limitations in sharing cross-scenario information. To address these issues, we propose a unified sequential recommendation model called SCRIPT. It takes cross-scenario user behavior sequences as input and explicitly incorporates meta information that characterizes scenario features, such as domain, site, and behavior types. Inspired by the advances of the pretrain and prompt paradigm, we generate scenario-aware and personalized prompts based on the user profile and meta information of candidate items. This allows the model to leverage the knowledge learned during pre-training and adapt it to serve different downstream tasks. Extensive experiments on two public dataset and a production dataset demonstrate that our model achieves state-of-the-art performance on multiple downstream recommendation tasks.

Social networks are widely recognized as highly effective information sources for social recommendation services. However, previous social recommendation methods assumed that a user’s preference factor and social trust factor shared a common latent feature space. Additionally, few studies have explored the incorporation of social information into the item domain for recommendations. To address these gaps, we propose ExpoEv, a deep collaborative filtering recommendation model that integrates social exposure based on feature evolution for social recommendation services. Specifically, we propose a social exposure module for both user and item domains that considers the number of items that a user’s social friends interact with. Furthermore, we introduce a feature evolution component that enables the incorporation of social exposure information with social trust and attribute factors in the context of social recommendation services. Experiments demonstrate the effectiveness of our model in the quality of recommendation service.

In large-scale e-commerce live-stream recommendation, streamers are classified into different levels based on their popularity and other metrics for marketing. Several top streamers at the head level occupy a considerable amount of exposure, resulting in an unbalanced data distribution. A unified model for all levels without consideration of imbalance issue can be biased towards head streamers and neglect the conflicts between levels. The lack of inter-level streamer correlations and intra-level streamer characteristics modeling imposes obstacles to estimating the user behaviors. To tackle these challenges, we propose a curriculum multi-level learning framework for imbalanced recommendation. We separate model parameters into shared and level-specific ones to explore the generality among all levels and discrepancy for each level respectively. The level-aware gradient descent and a curriculum sampling scheduler are designed to capture the de-biased commonalities from all levels as the shared parameters. During the specific parameters training, the hardness-aware learning rate and an adaptor are proposed to dynamically balance the training process. Finally, shared and specific parameters are combined to be the final model weights and learned in a cooperative training framework. Extensive experiments on a live-stream production dataset demonstrate the superiority of the proposed framework.