The memorization capability of large language models (LLMs) can enhance the continuity and personalization of services. However, it also gives rise to privacy risks. Grounded in privacy calculus theory, this study investigates how the memorization capability of medical LLMs affects users’ privacy disclosure willingness through two scenario-based experiments. The results indicate that the memorization capability significantly reduces the privacy disclosure willingness. When the memorization capability is enabled, users’ privacy disclosure willingness decreases by 20.9%. This effect is mainly realized through the mediating path of perceived risk. Moreover, the task capability plays a significant moderating role in this relationship. Additional analysis reveals that when the medical LLMs has the memorization capability, the negative relationship between privacy concerns and the privacy disclosure willingness only significant under specific conditions. However, while secrecy in disclosure consistently correlates negatively with disclosure willingness regardless of memorization capability.

As Large Language Models (LLMs) are integrated into high-stakes societal functions, their inherent biases have evolved from ethical concerns into critical, exploitable security vulnerabilities that undermine system integrity and trust. Traditional safety evaluations often fail to detect these subtle, context-dependent flaws. To address this, we introduce a scalable red-teaming framework designed to systematically attack and expose latent gender bias vulnerabilities in foundational models. Our framework operationalizes bias as an exploit, leveraging three distinct attack patterns—Latent Bias Elicitation, Forced-Choice Discrimination, and Stereotype-Amplifying Narrative Generation—to bypass safeguards and compel biased outcomes. We deployed this framework in a large-scale offensive campaign against a cohort of globally significant models, including the GPT, Claude, Gemini, and leading Chinese foundational model series. The attacks successfully manipulated all targets into producing statistically significant discriminatory outputs, proving that inherent bias is an operationally exploitable vulnerability. We discovered asymmetric weaknesses: English-centric models were attacked to exhibit strong male bias in Chinese contexts, while Chinese-centric models were vulnerable to similar male-biased exploits across both languages. This work provides concrete demonstration of socio-cultural bias as a potent and scalable attack vector, establishing the necessity of adversarial red-teaming for building trustworthy AI. All attack data and scripts are open-sourced to facilitate further security audits.

Differential privacy (DP) has achieved remarkable results in the field of privacy-preserving machine learning. However, existing DP frameworks do not satisfy all the conditions for becoming metrics, which prevents them from deriving better basic private properties and leads to exaggerated values on privacy budgets. We propose Wasserstein differential privacy (WDP), an alternative DP framework to measure the risk of privacy leakage, which satisfies the properties of symmetry and triangle inequality. We show and prove that WDP has 13 excellent properties, which can be theoretical supports for the better performance of WDP than other DP frameworks. In addition, we derive a general privacy accounting method called Wasserstein accountant, which enables WDP to be applied in stochastic gradient descent (SGD) scenarios containing subsampling. Experiments on basic mechanisms, compositions and deep learning show that the privacy budgets obtained by Wasserstein accountant are relatively stable and less influenced by order. Moreover, the overestimation on privacy budgets can be effectively alleviated. The code is available at https://github.com/Hifipsysta/WDP.

Continual learning aims to learn new knowledge from a sequence of tasks without forgetting. Recent studies have found that projecting gradients onto the orthogonal direction of task-specific features is effective. However, these methods mainly focus on mitigating catastrophic forgetting by adopting old features to construct projection spaces, neglecting the potential to enhance plasticity and the valuable information contained in previous gradients. To enhance plasticity and effectively utilize the gradients from old tasks, we propose Gradient Projection in Common Null Space (GPCNS), which projects current gradients into the common null space of final gradients under all preceding tasks. Moreover, to integrate both feature and gradient information, we propose a collaborative framework that allows GPCNS to be utilized in conjunction with existing gradient projection methods as a plug-and-play extension that provides gradient information and better plasticity. Experimental evaluations conducted on three benchmarks demonstrate that GPCNS exhibits superior plasticity compared to conventional gradient projection methods. More importantly, GPCNS can effectively improve the backward transfer and average accuracy for existing gradient projection methods when applied as a plugin, which outperforms all the gradient projection methods without increasing learnable parameters and customized objective functions. The code is available at https://github.com/Hifipsysta/GPCNS.

Continual learning emerges as a framework that trains the model on a sequence of tasks without forgetting previously learned knowledge, which has been applied in multiple multimodal scenarios. Recently, prompt-based continual learning has achieved excellent domain adaptability and knowledge transfer through prompt generation. However, existing methods mainly focus on designing the architecture of a generator, neglecting the importance of providing effective guidance for training the generator. To address this issue, we propose Generating Prompts in Latent Space (GPLS), which considers prompts as latent variables to account for the uncertainty of prompt generation and aligns with the fact that prompts are inserted into the hidden layer outputs and exert an implicit influence on classification. GPLS adopts a trainable encoder to encode task and feature information into prompts with reparameterization technique, and provides refined and targeted guidance for the training process through the evidence lower bound (ELBO) related to Mahalanobis distance. Extensive experiments demonstrate that GPLS achieves state-of-the-art performance on various benchmarks. Our code is available at https://github.com/Hifipsysta/GPLS.

Change address identification is one of the difficulties in bitcoin address clustering as an emerging social computing problem. Most of the current-related research only applies to certain specific types of transactions and faces the problems of low recognition rate and high false positive rate. We innovatively propose a clustering method based on multiconditional recognition of one-time change addresses and conduct experiments with on-chain bitcoin transaction data. The results show that the proposed method identifies at least 12.3% more one-time change addresses than other heuristics. On top of the multi-input heuristic clustering method, the proposed method also improves the address clustering performance by 5.7%, achieves optimal recognition results compared with similar methods, and significantly reduces the false positive rate of recognition results. This work provides the technical basis for antimoney laundering efforts based on entity identification. Code and data could be accessed from https://github.com/ECNU-Cross-Innovation-Lab/BitcoinAddressClustering.

In recent years, it is a difficult issue to integrate the deep cross-fertilization and interpretable cognitive modeling methods from the basic theory of emotional psychology with deep learning and other algorithms. To address this problem, a cognitive model that integrates the VGG-facial action coding system (FACS)-OCC model based on fer2013 expression features and the OCC-pleasure-arousal-dominance (PAD)-openness, conscientiousness, extraversion, agreeableness, and neuroticism (OCEAN) fusion of the basic theory of emotional psychology, namely, a computational affection-based OCC-PAD-OCEAN federation cognitive modeling (OPO-FCM), is constructed. By constructing this model and performing formal proof algorithms, it is shown that the OPO-FCM can acquire expression features in video streams, complete the acquisition of expression features in videos by training a deep neural network, map expressions to the PAD emotion space through the established expression–basic emotions–emotion space mapping relationship, and finally complete the mapping of the average emotion over a period time. The information of personality space is obtained through it. Finally, the experimental simulation of the model is conducted, and the results show that the average accuracy of the valid tested personalities is 79.56%. This article takes the knowledge-driven approach of emotional psychology as a starting point and combines deep learning techniques to construct interpretable cognitive models, thus providing new ideas for future cross-innovation between computer technology and psychology theory.

Fusion technology is crucial for multimodal sentiment analysis. Recent attention-based fusion methods demonstrate high performance and strong robustness. However, these approaches ignore the difference in information density among the three modalities, i.e., visual and audio have low-level signal features and conversely text has high-level semantic features. To this end, we propose a non-homogeneous fusion network (NHFNet) to achieve multimodal information interaction. Specifically, a fusion module with attention aggregation is designed to handle the fusion of visual and audio modalities to enhance them to high-level semantic features. Then, cross-modal attention is used to achieve information reinforcement of text modality and audio-visual fusion. NHFNet compensates for the differences in information density of different modalities enabling their fair interaction. To verify the effectiveness of the proposed method, we set up the aligned and unaligned experiments on the CMU-MOSEI dataset, respectively. The experimental results show that the proposed method outperforms the state-of-the-art. Codes are available at https://github.com/skeletonNN/NHFNet.

Abstract Skeleton‐based action recognition has been extensively studied in recent years and applied in virtual reality, detection systems and other cases with strong requirements for low cost as well as high accuracy, but most of the existing methods mainly focus on complex architecture of deep neural networks without considering computation efficiency. To balance accuracy and computation cost well, this paper proposes a simple and efficient attention‐enhanced semantic‐guided graph convolutional network (AeS‐GCN) for skeleton‐based action recognition. Firstly, we fuse semantics of joint type and frame index and dynamics together as representation of skeleton. Then, we use spatial attention block (SAB) to explore important features in spatial structure, in which adaptive GCN layer is adopted to adaptively model skeleton topology structure. Next, we use temporal attention block (TAB) to extract latent temporal information. The model proposed is a lightweight network and achieves the state‐of‐the‐art performance on mainstream datasets with less parameters and less computational complexity.

Spectrum detection plays an important role in 5G communication and Internet of things (IoT) networks which is one of the core technologies in wireless communication network. The process of spectrum detection largely depends on energy sensing as this process needs to detect the energy of different frequency bands in the information transmission. In this paper, we propose an energy-aware scheme based on spectrum detection in wireless sensor networks applications, which can save the cost of data collection and computing power of DDoS attack de-tection without collecting a large number of traffic characteris-tics and large-scale deep learning model training. By detecting the information transmission energy in different frequency channels, the busy and idle state of the channel is judged. On the basis of the busy channel detection, the scope of DDoS attack is further judged by the topology of wireless sensor networks. Sim-ulation results show that the proposed energy-aware algorithm can achieve the expected efficiency and accuracy, and is suitable for low load and high security requirements of wireless sensor networks.

With the continuous development and popularity of depth cameras, skeleton-based human action recognition has attracted people's wide attention. Graph Convolutional Network (GCN) has achieved remarkable performance. However, the existing methods do not better consider the semantic characteristics, which can help to express the current concept and scene information. Semantic information can also help with better granularity classification. In addition, most of the existing models require a lot of computation. What's more, adaptive GCN can automatically learn the graph structure and consider the connections between joints. In this paper, we propose a relatively less computationally intensive model, which combines semantic and adaptive graph network (SAGN) for skeleton-based human action recognition. Specifically, we mainly combine the dynamic characteristics and bone information to extract the data, taking the correlation between semantics into the model. In the training process, SAGN includes an adaptive network so that we can make attention mechanism more flexible. We design the Convolutional Neural Network (CNN) for feature extraction on the time dimension. The experimental results show that SAGN achieves the state-of-the-art performance on NTU-RGB+D 60 and NTU-RGB+D 120 datasets. SAGN can promote the study of skeleton-based human action recognition. The source code is available at https://github.com/skeletonNN/SAGN.

With the development of Web 2.0, traditional customers have increasingly transferred to online purchase and created a large volume of User Generated Content (UGC) on the Internet, which brought traditional customer relationship management great challenges and attract many scholars' attention on customer review. Most of the previous researches focus on the influence of customer's review behavior on customer's purchase behavior, but little researches explore the impact on the reverse direction. In this paper, our study seeks insights into analyzing the impact of customer's purchase behavior on its review behavior and discovering how this effect could be fully utilized to predict customer review's churn in the next stage. Based on data from Dianping.com, a famous comprehensive website which contains review and purchase platforms, we build the Logit regression model, considering customer's own factors, review behavior and purchase behavior and finding the impact of user's purchase behavior on its review behavior. Finally, we also use ten-fold cross-validation to prove the stability of our model. Our study can provide a theoretical basis for research on User Generated Content.

With the rapid growth of online daily deals, the behaviors of consumers in such online group buying commerce have become popular research topics. We examine the effects of the period, price, discount rate and product category on sales in the context that local consumers and visitors purchase the restaurant coupons in a group buying electronic marketplace. Applying the conjoint analysis algorithm, we study the actual transactional data sets on the Dianping.com, one of the largest online group buying commerce businesses. The results show that there is a clear discrepancy between the preferences of the two consumer groups. The local consumers tend to think high of discounts while the visitors likely care more on product category. In addition, the behaviors of two groups are different in product attribute selections. The findings could be useful to the businesses in terms of understanding the customers' purchasing and product designs.

Network value has attracted more and more attentions from scholars and enterprisers. Information through network can essentially impact everyone, while everyone in the network has the structure function to the network and may also influence others. But the contribution to the network is not the same for every one. Based on the complex network theory and customer activity, a customer network value model is proposed, which is from the influence value and existence value. Then a case study was done with the data of a city branch of China Telecom. Finally, the management application based on customer network value is proposed.

Customer churn prediction model is hot research topic in recent years. Most of the researchers have paid much attention on how to construct novelist data mining algorithm for the prediction model, while less research concerns the choosing of input variables for the churn prediction model. This paper focuses on how to select effective input variables for the telecommunications customer churn model. We proposed a procedure to select the input variables step by step, and proved the effect by comparative experiment using the data from one telecom carrier.

When the customer churn prediction model is built, a large number of features bring heavy burdens to the model and even decrease the accuracy. This paper is aimed to review the feature selection, to compare the algorithms from different fields and to design a framework of feature selection for customer churn prediction. Based on the framework, the author experiment on the structured module with some telecom operator's marketing data to verify the efficiency of the feature selection framework.

Decision tree, neural network and logistic regression were applied frequently as models of customer churn prediction, but the application of them has been mature and they are difficult to be improved. In this paper, Bayesian networks, support vector machines, rough sets and survival analysis were selected for experimental comparison study. An integrated contrast among the four models from the applicability of model in theory and experimental comparison has been processed. Overall, of the four models the Bayesian network model performed best while the survival analysis did worst.

The use of peer-to-peer (P2P) applications is growing dramatically, particularly for BitTorrent system. In order to gain insights into BitTorrent systems and the network traffic load they place on ISPs, we have undertaken an measurement study. Our experimental evaluation is ISP oriented instead of peer oriented, which enables us to study the global characteristics of BitTorrent system. We have developed a dedicated BitTorrent sniffer platform to collected extensive packet across a large ISP network. The measurement results bring important insights into BitTorrent systems. Specifically,our results show that 1) BitTorrent flashcrowd appears around midnight due to the BitTorrent users habits of behavior; 2) workload generated by BitTorrent extention - DHT is much more than it generated by calssic BitTorrent; 3) overhead rather than data transfer is the dominant component of the totle BitTorrent traffic due to DHT; 4) Zipf and pareto are the suitable model to characterize the distribution of visits to both trackers and BitTorrent Websites.Insights obtained in this study will be valuable for the management and development of future P2P file transform systems.

Globalization requires enterprises to suit different customers' demands and integrate different countries' resources. Current research on SOA could not provide an appropriate way for support. This paper proposes a SOA (SOA-RIPAS) in globalization. It describes customers semantically and recommends worldwide products/services to customers. It develops products and new application services by semantically conversing and compositing resources from worldwide Application Service Providers (ASP) and Manufacturers.

Recommender systems apply statistical and knowledge discovery techniques to the problem of making product recommendations during a live customer interaction and they are achieving widespread success in e-commerce nowadays. In this article, we present a novel product recommendation approach, which involves customer value hierarchy model into traditional recommender systems. This approach is divided into two phases. Product categories are recommended using the collaborative filtering algorithm in the first phase. In phase II, product items are recommended, based on customer value hierarchy model, to customers whose purchasing goals are met by these products’ attributes. In contrast to traditional approaches, which provide recommendation by the opinions of customers with the similar purchasing behavior, the proposed approaches root out customer’s purchasing motivation and maximize customer satisfaction.

This paper presents a hybrid approach for building a binary classifier. The approach is the combination of the k-nearest neighbor algorithm, handling separately m 1-dimensional data sets divided from a data set in m-dimension, and the logistic regression method. This hybrid KNN-LR classifier improves the performance of the logistic regression in classification accuracy in some situations where the predictor and target variables exhibit complex nonlinear relationships. The results of the experiment on four benchmark data sets show the proposed approach compares favorably with the well-known classification algorithms such as C4.5 and RBF. Furthermore, its effectiveness is illustrated by its application in customer churn prediction based on real-world customer data sets.

Abstract:Artificial intelligence (AI) has been applied to the telecommunications industry for more than a decade. The purpose of this paper is to examine the application of AI in the telecommunications industry sector. Our research finds that AI's first main application in telecommunications is in the network management area. Expert systems and machine learning are the two AI techniques that have been widely used in telecommunications, while machine learning and distributed artificial intelligence are the two AI techniques which are most promising for the future. The research also finds that different AI techniques have their unique applications in the telecommunications industry.

For the purpose of improving the predictive accuracy and interpret ability of churn prediction model, TreeLogit model, which integrates the advantage of AD tree model and logistic regression model, is proposed in this paper to predict customers' churn propensities. Compared with TreeNetreg, a model which won the gold prize in the 2003 mobile customer churn prediction modeling contest (The Duke/NCR Teradata Churn Modeling Tournament) , the overall predictive accuracy of TreeLogit model is not inferior. In fact, the predictive accuracy of TreeLogit is superior to the one of TreeNetreg on two important observation points (i.e. the captured rates in 10% and 50% of the most likely churn groups respectively)

Operations support systems (OSS) is one kind of enterprise information systems (EIS), which is becoming increasingly popular in telecommunication industry. However, the academic research in OSS is sparse. In this paper, a general review and research taxonomy of OSS are provided.