The video game industry has emerged as a significant economic force, driving extensive research on optimizing the gaming environment and improving gaming experiences. Among these endeavors, player behavior representation learning has become a critical way to model valuable player properties and is beneficial for a wide range of downstream tasks. However, some common factors, such as login rewards and daily tasks, can trigger similar behaviors among different players, which are informative and noisy for learning high-quality player behavior representations. Existing methods ignore the low signal-to-noise ratio in player behavior data and waste too much modeling capacity on less informative behaviors, resulting in their learned representations being noisy. In this paper, we propose a novel model for Denoised Player Behavior representation Learning, namely DPBL, which consists of two key modules. The first module extracts various player behavior patterns and isolates them from less informative noise. The second module utilizes the extracted patterns to refine the embedding of each behavior and eliminates noise. To optimize DPBL, two contrastive learning strategies are proposed to identify the noise that should be eliminated and to learn distinguishable representations, respectively. With the above design, DPBL is capable of mitigating the impact of noise in the data and learning high-quality representations that effectively capture player characteristics. We conducted extensive experiments on two real-world datasets, and DPBL outperforms all baselines on various downstream tasks with an improvement of 1.4% ∼ 18.1%. The results also show that DPBL achieves an improvement of 5.6% ∼ 23.0% in the denoising experiments, which proves that DPBL is more robust to noisy behaviors. Code is available at https://github.com/LwbXc/DPBL.

Adversarial purification has emerged as a critical defense mechanism against various adversarial attacks on deep neural networks, however, such category of diffusion-based purification in the pixel space confronts two challenges while applying into today’s large vision transformers: i) the distributional divergence between adversarial and clean examples in high-dimensional image manifold; and ii) the prohibitive computational cost while processing high-resolution images. To address the predicament, we in this paper focus on the transformer-based visual encoders commonly employed in large Vision-Language Models (VLMs), and propose a novel Latent Diffusion-based Purification (LDP) mechanism through leveraging the latent space to pave the gap between adversarial distribution and clean distribution. Resorting to projecting adversarial input into a low-dimensional latent representation, our proposed LDP not only suppresses the off-manifold perturbations to achieve accelerated denoising, but also preserves critical semantic features by aligning input with the priori high-quality visual representations. Multi-facet experiments over both proactive robustness enhancement and post-attack purification demonstrate that our LDP has a superior performance in terms of effectiveness and efficiency.

Trajectory anomaly detection, aiming to estimate the anomaly risk of trajectories given the Source-Destination (SD) pairs, has become a critical problem for many real-world applications. Existing solutions directly train a generative model for observed trajectories and calculate the conditional generative probability$P(T \vert C)$as the anomaly risk, where$T$and$C$represent the trajectory and SD pair respectively. However, we argue that the observed trajectories are confounded by road network preference which is a common cause of both SD distribution and trajectories. Existing methods ignore this issue limiting their generalization ability on out-of-distribution trajectories. In this paper, we define the debiased trajectory anomaly detection problem and propose a causal implicit generative model, namely CausalTAD, to solve it. CausalTAD adopts do-calculus to eliminate the confounding bias of road network preference and estimates$P(T\vert do(C))$as the anomaly criterion. Extensive experiments show that CausalTadcan not only achieve superior performance on trained trajectories but also generally improve the performance of out-of-distribution data, with improvements of 2.1% ~ 5.7% and 10.6% ~ 32.7% respectively.

Trajectory similarity computation as the fundamental problem for various downstream analytic tasks, such as trajectory classification and clustering, has been extensively studied in recent years. However, how to infer an accurate and robust similarity over two trajectories is difficult due to the some trajectory characteristics in practice, e.g. non-uniform sampling rate, nonmalignant fluctuation, and noise points, etc. To circumvent such challenges, we in this paper introduce the adversarial training idea into the trajectory representation learning for the first time to enhance the robustness and accuracy. Specifically, our proposed method AdvTraj2Vec has two novelties: i) it perturbs the weight parameters of embedding layers to learn a robust model to infer an accurate pairwise similarity over each two trajectories; and ii) it employs the GAN momentum to harness the perturbation extent to which an appropriate trajectory representation can be learned for the similarity computation. Extensive experiments using two real-world trajectory datasets Porto and Beijing validate our proposed AdvTraj2Vec on the robustness and accuracy aspects. The multi-facet results show that our AdvTraj2Vec significantly outperforms the stat-of-the-art methods in terms of different distortions, such as trajectory-point addition, deletion, disturbance, and outlier injection.

To circumvent the grave problem of present adversarial training methods, i.e. distortion of classification surface, we in this paper propose a generated Triplet-based adversarial training method-GTAT, in which a Generator generates a semi-hard Triplet by design, rather than directly invoking the existing clean examples and adversarial examples. Through this kind of generated semi-hard Triplet constraint, GTAT can reshape the classification boundaries appropriately across various classes, arising from two-facet synergies: i) pull the intra-class examples together with tight distances; and ii) push away the inter-class examples with broad distances. This synergy will simplify and broaden the classification surfaces across different classes. Extensive experiments on the popular MNIST and CIFAR-10 datasets show that our proposed GTAT significantly outperforms other state-of-the-art adversarial training methods. We believe GTAT opens a door for the adversarial training from a new horizon of rationally generating semi-hard Triplet-satisfied adversarial training (retraining) examples, instead of straightly performing retraining on the generated adversarial examples and existing clean examples, or on the generated adversarial examples only.

In this paper, we propose a new task namely trajectory cross-modal retrieval which achieves the cross-modal search between coordinate trajectories and images containing trajectories. Nevertheless, trajectory cross-modal retrieval is rather challenging in learning the representations of each modality and reduce the cross-domain discrepancy caused by the inconsistent data distribution at the same time. we proposes a cross-modal retrieval model TrajCross based on multi-level representation for trajectory cross-modal retrieval. Specifically, TrajCross extracts the location features and the shape information respectively for the represention of multi-modal data. we adopt a contrastive learning method to achieve semantic preservation among similar multi-modal data. Extensive experiments show that TrajCross significantly outperforms state-of-the-art cross-modal retrieval methods.

Social media has became a critical manner for people to acquire information in daily life. Despite the great convenience, fake news can be widely spread through social networks, causing various adverse effects on people's lives. Detecting these fake news or misinformations has proved to be a critical task and draws attentions from both governments and individuals. Recently, many methods have been proposed to solve this problem, but most of them rely on the body content of the news, ignoring the social context information such as the comments. We argue that the comments of a specific news contain common judgements of the whole society and could be extremely useful for detecting fake news. In this paper, we propose a new method TRANSFAKE which jointly models the body content and comments of news systemically, and detects fake news with multi-task learning framework. TRANSFAKE model is a Transformer-based model. It takes different modalities as input and employs multiple tasks, i.e. rumor score prediction and event classification, as intermediate tasks for extracting useful hidden relationships across various modalities. These intermediate tasks promote each other and encourage TRANSFAKE making the right decision. Extensive experiments on two standard real-life datasets demonstrate that TRANSFAKE outperforms state-of-the-art methods. It improves the detection accuracy by margins as large as ~12.6% and F1 scores as large as ~15%.

Recently, a new methodology using generative adversarial network (GAN) has been proposed to produce adversarial examples, which breaks the limitations of the previous methods dependent on different norm-levels. It can efficiently generate perturbations for any instance once the generator is trained, arising from the learning to approximate the distribution of real instances. However, there are still two shortcomings for this category of GAN-based method: i) the predicted label in attacking stage totally depend on a fixed or randomly-chosen label in training stage, which cannot tackle the elasticity problem on how to elastically produce adversarial example with any arbitrarily-assigned label in targeted attack scene when the generator has finished training; and ii) it only considering the produced adversarial example is as close as the real instances, which cannot guarantee the generated adversarial example is visually indistinguishable from its corresponding original instance perceptually. The aboved two disadvantages make this kind of method lack of flexibility and covertness. To circumvent these two predicaments, we in this paper propose a simple and easy-to-use adversarial example generating framework AdvCGAN through training a conditional generative adversarial network under the co-consideration on the similarities in data distributions and the image labels between the adversarial examples and the original instances to be imperceptible to humans. Concretely, our proposed AdvCGAN trains the conditional GAN with both image data and label (normal and attack) information, by which the generator can utilizing the guidance of label information to appropriately produce the adversarial example with any specific label in attacking stage. Extensive experiments using the commonly used MNIST and CIFAR-10 datasets show that our proposed AdvCGAN significantly outperforms other methods in terms of multi-facet evaluation. The results exhibit that our AdvCGAN can elastically produce more realistic adversarial examples with any arbitrarily-assigned attack label and achieve higher attack accuracy, especially in targeted attack.

Recommending citywide POIs where users would visit in the next future benefits many location-based businesses and individuals. To train a decent recommendation model, adequate historical data is usually a prerequisite. However, historical check-ins are usually distributed unevenly, which leads to many cities suffering from data scarcity. To make matters worse, transferring knowledge from data sufficient cities is challenging due to the varying distribution of POIs and city structures. Most of existing next POI recommendation methods assume that the training data is adequate and can not solve these problems. In this paper, we propose a novel meta-learning enhanced neural ordinary differential equation (ODE) method, namely METAODE, which models city-invariant information and city-specified information separately to achieve accurate citywide next POI recommendation. For transferring knowledge from data sufficient cities, METAODE learns city-invariant information including the representation of POIs categories and user groups to extract user preference. Basing on that, METAODE employs a GRU-ODE-Bayes model for city-specified information modeling. It can not only capture the sequential relationships within the historical check-ins but also model the irregular-sampled timestamp in the continuous timeline. Moreover, METAODE leverages meta-learning mechanism to optimize the parameters on various data sufficient cities and train a well-generalized initialization, which can be effectively adapted to data insufficient cities to enhance recommendation performance. Extensive experiments on real-world datasets demonstrate the effectiveness of METAODE. Comparing with the state-of-the-art baselines, METAODE achieves 6.21% and 14.77% improvements on HR and NDCG, respectively.