Sequential recommendation has attracted a lot of attention from both academia and industry, however the privacy risks associated with gathering and transferring users’ personal interaction data are often underestimated or ignored. Existing privacy-preserving studies are mainly applied to traditional collaborative filtering or matrix factorization rather than sequential recommendation. Moreover, these studies are mostly based on differential privacy or federated learning, which often lead to significant performance degradation, or have high requirements for communication. In this work, we address privacy-preserving from a different perspective. Unlike existing research, we capture collaborative signals of neighbor interaction sequences and directly inject indistinguishable items into the target sequence before the recommendation process begins, thereby increasing the perplexity of the target sequence. Even if the target interaction sequence is obtained by attackers, it is difficult to discern which ones are the actual user interaction records. To achieve this goal, we introduce a novel sequential recommender system called CoLlaborative-cOnfusion seqUential recommenDer (CLOUD) , which incorporates a collaborative confusion mechanism to modify the raw interaction sequences before conducting recommendation. Specifically, CLOUD first calculates the similarity between the target interaction sequence and other neighbor sequences to find similar sequences. Then, CLOUD considers the shared representation of the target sequence and similar sequences to determine the operation to be performed: keep, delete, or insert. A copy mechanism is designed to make items from similar sequences have a higher probability to be inserted into the target sequence. Finally, the modified sequence is used to train the recommender and predict the next item. We conduct extensive experiments on three benchmark datasets. The experimental results show that CLOUD achieves a maximum modification rate of 66.57% on interaction sequences and obtains over 99% recommendation accuracy compared to the state-of-the-art sequential recommendation methods. This proves that CLOUD can effectively protect user privacy at minimal recommendation performance cost, which provides a new solution for privacy-preserving for sequential recommendation. Our implementation is available at https://github.com/weiwang0927/CLOUD .

Modern recommender systems heavily rely on high-quality representations learned from high-dimensional sparse data. While significant efforts have been invested in designing powerful algorithms for extracting user preferences, the factors contributing to good representations have remained relatively unexplored. In this work, we shed light on an issue in the existing pairwise learning paradigm (i.e., embedding collapse), that the representations tend to span a subspace of the whole embedding space, leading to a suboptimal solution and reducing the model capacity. Specifically, we show that alignment of positive pairs is equivalent to a low-pass filter causing users and items to collapse to a constant vector. While negative sampling can partially mitigate this issue by acting as a high-pass filter to balance the spectrum, leading to an incomplete collapse. To tackle this issue, we present a novel learning paradigm DirectSpec, which directly optimizes the spectrum distribution to ensure that users and items effectively span the entire embedding space. We demonstrate that many self-supervised learning algorithms without explicit negative sampling can be considered as special cases of DirectSpec. Furthermore, we show that optimizing the spectrum inappropriately could also be detrimental to data representation, where the key lies in a dynamic balance between alignment of positive pairs and spectrum balancing. Finally, we propose an enhanced and practical implementation DirectSpec + to balance the embedding spectrum more adaptively and effectively. We implement DirectSpec + on two popular recommender models: matrix factorization and LightGCN. Our experimental results demonstrate its effectiveness and efficiency over competitive baselines.

It has been shown that the effectiveness of graph convolutional network (GCN) for recommendation is attributed to the spectral graph filtering. Most GCN-based methods consist of a graph filter or followed by a low-rank mapping optimized based on supervised training. However, we show two limitations suppressing the power of graph filtering: (1) Lack of generality. Due to the varied noise distribution, graph filters fail to denoise sparse data where noise is scattered across all frequencies, while supervised training results in worse performance on dense data where noise is concentrated in middle frequencies that can be removed by graph filters without training. (2) Lack of expressive power. We theoretically show that linear GCN (LGCN) that is effective on collaborative filtering (CF) cannot generate arbitrary embeddings, implying the possibility that optimal data representation might be unreachable.

While Graph Convolutional Networks (GCNs) have shown great potential in recommender systems and collaborative filtering (CF), they suffer from expensive computational complexity and poor scalability. On top of that, recent works mostly combine GCNs with other advanced algorithms which further sacrifice model efficiency and scalability. In this work, we unveil the redundancy of existing GCN-based methods in three aspects: (1) Feature redundancy . By reviewing GCNs from a spectral perspective, we show that most spectral graph features are noisy for recommendation, while stacking graph convolution layers can suppress but cannot completely remove the noisy features, which we mostly summarize from our previous work; (2) Structure redundancy . By providing a deep insight into how user/item representations are generated, we show that what makes them distinctive lies in the spectral graph features, while the core idea of GCNs (i.e., neighborhood aggregation) is not the reason making GCNs effective; and (3) Distribution redundancy . Following observations from (1), we further show that the number of required spectral features is closely related to the spectral distribution, where important information tends to be concentrated in more (fewer) spectral features on a flatter (sharper) distribution. To make important information be concentrated in as few features as possible, we sharpen the spectral distribution by increasing the node similarity without changing the original data, thereby reducing the computational cost. To remove these three kinds of redundancies, we propose a Simplified Graph Denoising Encoder (SGDE) only exploiting the top- K singular vectors without explicitly aggregating neighborhood, which significantly reduces the complexity of GCN-based methods. We further propose a scalable contrastive learning framework to alleviate data sparsity and to boost model robustness and generalization, leading to significant improvement. Extensive experiments on three real-world datasets show that our proposed SGDE not only achieves state-of-the-art but also shows higher scalability and efficiency than our previously proposed GDE as well as traditional and GCN-based CF methods.

Out-of-Domain (OOD) text detection has attracted significant research interest. However, conventional approaches primarily employ Cross-Entropy loss during upstream encoder training and seldom focus on optimizing discriminative In-Domain (IND) and OOD representations. To fill this gap, we introduce a novel method that applies supervised contrastive learning (SCL) to IND data for upstream representation optimization. This effectively brings the embeddings of semantically similar texts together while pushing dissimilar ones further apart, leading to more compact and distinct IND representations. This optimization subsequently improves the differentiation between IND and OOD representations, thereby enhancing the detection effect in downstream tasks. To further strengthen the ability of SCL to consolidate IND embedding clusters, and to improve the generalizability of the encoder, we propose a method that generates two different variations of the same text as "views". This is achieved by applying a twice "dropped-out" on the embeddings before performing SCL. Extensive experiments indicate that our method not only outperforms state-of-the-art approaches, but also reduces the requirement for training a large 354M-parameter model down to a more efficient 110M-parameter model, highlighting its superiority in both effectiveness and computational economy.

With the tremendous success of Graph Convolutional Networks (GCNs), they have been widely applied to recommender systems and have shown promising performance. However, most GCN-based methods rigorously stick to a common GCN learning paradigm and suffer from two limitations: (1) the limited scalability due to the high computational cost and slow training convergence; (2) the notorious over-smoothing issue which reduces performance as stacking graph convolution layers. We argue that the above limitations are due to the lack of a deep understanding of GCN-based methods. To this end, we first investigate what design makes GCN effective for recommendation. By simplifying LightGCN, we show the close connection between GCN-based and low-rank methods such as Singular Value Decomposition (SVD) and Matrix Factorization (MF), where stacking graph convolution layers is to learn a low-rank representation by emphasizing (suppressing) components with larger (smaller) singular values. Based on this observation, we replace the core design of GCN-based methods with a flexible truncated SVD and propose a simplified GCN learning paradigm dubbed SVD-GCN, which only exploits K-largest singular vectors for recommendation. To alleviate the over-smoothing issue, we propose a renormalization trick to adjust the singular value gap, resulting in significant improvement. Extensive experiments on three real-world datasets show that our proposed SVD-GCN not only significantly outperforms state-of-the-arts but also achieves over 100x and 10x speedups over LightGCN and MF, respectively.

As much as Graph Convolutional Networks (GCNs) have shown tremendous success in recommender systems and collaborative filtering (CF), the mechanism of how they, especially the core components (\textiti.e., neighborhood aggregation) contribute to recommendation has not been well studied. To unveil the effectiveness of GCNs for recommendation, we first analyze them in a spectral perspective and discover two important findings: (1) only a small portion of spectral graph features that emphasize the neighborhood smoothness and difference contribute to the recommendation accuracy, whereas most graph information can be considered as noise that even reduces the performance, and (2) repetition of the neighborhood aggregation emphasizes smoothed features and filters out noise information in an ineffective way. Based on the two findings above, we propose a new GCN learning scheme for recommendation by replacing neihgborhood aggregation with a simple yet effective Graph Denoising Encoder (GDE), which acts as a band pass filter to capture important graph features. We show that our proposed method alleviates the over-smoothing and is comparable to an indefinite-layer GCN that can take any-hop neighborhood into consideration. Finally, we dynamically adjust the gradients over the negative samples to expedite model training without introducing additional complexity. Extensive experiments on five real-world datasets show that our proposed method not only outperforms state-of-the-arts but also achieves 12x speedup over LightGCN.

Prediction of bus travel time and/or delay time is a useful tool for passengers who want to plan their journey, e.g., when they should leave from the origin bus stop, what they will do after arriving at the destination bus stop, and so on. Many studies have tackled this task using probe data and/or the real time data provided by automatic vehicle location (AVL) systems. Most of them only targeted a small number of routes, short time periods, e.g. less than one week, and used few machine learning models to evaluate their methods. However, different routes generally show different characteristics. In fact, there are big differences between urban routes and rural routes. Furthermore, the performance of machine learning models also varies according to the data dealt with by the models. In this paper, we propose prediction models for bus delay over all intervals between pairs of adjacent bus stops. To build the models, we use one month of bus probe data, which includes more than 80 routes, and apply several machine learning models: linear regression (LR), artificial neural network (ANN), support vector regression (SVR), random forest (RF), and gradient boosting decision tree (GBDT). Experimental results demonstrate the superiority of the GBDT-based prediction model and the effects of considering travel time over prior intervals.

Government 2.0 activities have become very attractive and popular these days. Using platforms to support the activities, anyone can anytime report issues or complaints in a city with their photographs and geographical information on the Web, and share them with other people. Since a variety of reports are posted, officials in the city management section have to check the importance of each report and sort out their priorities to the reports. However, it is not easy task to judge the importance of the reports. When several officials work on the task, the agreement rate of their judgments is not always high. Even if the task is done by only one official, his/her judgment sometimes varies on a similar report. To remedy this low agreement rate problem of human judgments, we propose a method of detecting signs of danger or unsafe problems described in citizens' reports. The proposed method uses a machine learning technique with word feature selection. Experimental results clearly explain the low agreement rate of human judgments, and illustrate that the proposed machine learning method has much higher performance than human judgments.

Government 2.0 activities have become very attractive and popular these days. Using platforms to support the activities such as FixMyStreet, SeeClickFix, or CitySourced, anyone can anytime report issues or complaints in a city with their photographs and geographical information on the Web, and share them with other people. On the other hand, unlike telephone calls, the concreteness of a report depends on its reporter; the actual status and demand to the status may not be described clearly or either one may be miss-described in the report. It may accordingly happen that officials in the city management section can not understand the actual status or a demand to the status from the report. To solve the problems, it is indispensable to complement missing information and estimate the actual status or the demand to the status from ambiguous information in the report. This paper proposes novel methods to detect segments related to an actual status and the demand to the status in a report. The methods combine empirical rules with several machine learning techniques that actively use dependency relation between words. Experimental results illustrate the validity of the proposed methods.

In Web retrieval, it is often the case that the results given by search engines are not just what we want. To solve this problem there have been many studies on improving queries to be submitted to search engines. However, they are still insufficient due to lack of consideration on time information. To remedy this we propose a user-schedule-based Web page recommendation method. The method makes use of a user's schedule to make the recommendation suite to his/her requests. In addition, an algorithm for extracting related words is introduced as a key technique in this method. Some preliminary experiments show very promising results in recommending Web pages relevant to users requests. We also confirmed that this algorithm outperforms a method using mutual information.