Speech represents a fundamental medium for conveying human emotions and, as a result, speech-based emotion recognition (SER) systems have become pivotal in advancing human-computer interaction (HCI) across a range of applications. While significant progress has been made in speech emotion recognition over recent years, existing solutions still face several key challenges, in that they:$(i)$rely excessively on subjectively annotated (discrete) labels during training,$(ii)$often overlook the label ambiguity of speech samples that express more than one class of emotions, and$(iii)$underutilize unlabeled or ambiguous speech, for which typically a label distribution (or so-called soft labels) is available. To address these issues, we propose in this paper a novel SER model that explicitly handles ambiguous speech samples and overcomes the shortcomings outlined above. Central to our approach is a novel real-time soft-label correction strategy designed to refine the annotations assigned to ambiguous speech. The proposed model leverages both, (explicitly) labeled as well as ambiguous samples and applies the dynamic soft-label correction strategy alongside an enhanced inter-class difference loss function to iteratively optimize the label distributions during training. We theoretically demonstrate that our method is capable of approximating the true emotional distribution of speech even in the presence of label noise, suggesting that utilizing ambiguous speech samples without explicit emotion labels still contributes toward more effective emotion recognition. Furthermore, we integrate the representational power of convolutional neural networks (CNNs) with the contextual modeling capabilities of Wav2Vec 2.0 to enable a comprehensive extraction of spatio-temporal speech features. Experimental results on the IEMOCAP multi-label dataset confirm the effectiveness of our approach, achieving state-of-the-art performance with significant improvements in weighted accuracy (WA) and unweighted accuracy (UA) over competing methods.

Abstract Relation extraction is a research hot-spot in the field of natural language processing, and aims at structured knowledge acquirement. However, existing methods still grapple with the issue of entity overlapping, where they treat relation types as inconsequential labels, overlooking the fact that relation type has a great influence on entity type hindering the performance of these models from further improving. Furthermore, current models are inadequate in handling the fine-grained aspect of entity positioning, which leads to ambiguity in entity boundary localization and uncertainty in relation inference, directly. In response to this challenge, a relation extraction model is proposed, which is guided by relational semantic cues and focused on entity boundary localization. The model uses an attention mechanism to align relation semantics with sentence information, so as to obtain the most relevant semantic expression to the target relation instance. It then incorporates an entity locator to harness additional positional features, thereby, enhancing the capability of the model to pinpoint entity start and end tags. Consequently, this approach effectively alleviates the problem of entity overlapping. Extensive experiments are conducted on the widely used datasets NYT and WebNLG. The experimental results show that the proposed model outperforms the baseline ones in F1 scores of the two datasets, and the improvement margin is up to 5.50% and 2.80%, respectively.

The Trusted Platform Module (TPM) enhances system security by offering features such as a root of trust, secure storage, and authentication mechanisms. While TPM has been widely adopted, there has been no detailed exploration of a firmware-based TPM implementation specifically designed for the RISC-V architecture. In this paper, we present the design and implementation of a firmware TPM system for RISC-V, named rTPM. rTPM leverages DRAM latency-based Physical Unclonable Functions (PUFs) and PMP (Physical Memory Protection) to achieve secure NVRAM storage. Addressing challenges such as the need for additional security hardware extensions and the requirement for a Trusted Execution Environment (TEE) in firmware-based TPMs, we develop a secure communication mechanism across different privilege levels, tailored to the RISC-V security architecture. In addition, we proposed new solutions to address rollback attacks and the absence of secure clocks. Our prototype demonstrates that, although rTPM incurs approximately 200 ms of additional overhead during startup and data read/write operations, it significantly improves message transmission efficiency. As a result, rTPM achieves a performance enhancement of nearly 2-3 times compared to TPM emulators when executing related instructions, while also providing enhanced security.

Abstract With the rapid development of information technology, the problem of “information overload” emerges when users choose cloud services. How to integrate multi‐source information to achieve accurate service recommendation is an urgent problem to be solved by current recommendation systems. This article proposes a cloud service recommendation method based on extended multi‐source information fusion. First, we propose a score prediction based on matrix decomposition and topic matrix, and we fully mine existing explicit data and feedback data, such as user ratings, social trust information, reviews, and user personalized preferences and so on. Second, in order to solve the problems of data sparseness and cold start of the system, we integrate the score, social trust information and review into a comprehensive model through collaborative filtering (CF), and propose a multi‐source information fusion recommendation method. The CF fusion method mainly combines two parts: social matrix decomposition and topic matrix decomposition. Finally, in order to further improve the accuracy and scalability, the implicit feature matrix is integrated into the user rating matrix, and the original CF enhancement based on scoring matrix decomposition is a matrix decomposition method that can learn implicit features. Experimental results show that compared with other recommendation algorithms, the cloud service recommendation method proposed in this article can improve the recommendation accuracy and allow users to choose satisfactory cloud services.

Cloud services are cloud computing resources and applications deployed on the Internet or cloud computing platform, and users can access the required cloud services at any time. However, users face the diversity and complexity of quality of service (QoS) when evaluating and selecting cloud services. Therefore, it is important to study and establish an effective and objective trust model to improve user satisfaction and interaction success rate. In this paper, a model based on combining weights and gray correlation analysis is proposed. Firstly, direct trust, recommendation trust, and reputation together form a comprehensive trust, resulting in a more accurate overall trust. Second, rough set theory and analytic hierarchy process (AHP)-based method are used for the direct trust. Meanwhile, the degree of similarity recommendation trust is calculated by a gray relational analysis method. In order to ensure the accuracy of direct trust, this paper proposes a dynamic trust update mechanism. Finally, the simulation experiment is carried out to verify that the cloud services trust evaluation model (CSTEM) is more robust than the other three methods. It protects against malicious entities; at the same time, it can increase user satisfaction and interaction success rate.

Driven by the energy crisis and global warming problem, smart grid was proposed in the early 21th century as a solution for the sustainable development of human society. With the two-way communication infrastructure available in smart grids, a current challenge is to interpret and gain knowledge from the collected grid big data to optimize grid operations. Service recommendation techniques provide promising tools to discover knowledge from the grid data, and recommend energy-aware products/services/suggestions to the smart grid participators. This paper is among the first to investigate the prospective of introducing service recommendation techniques into the smart grid demand side management (DSM). In the first part of the paper, the backgrounds of smart grid DSM and service recommendation techniques are reviewed, followed by the presentation and discussion of key technologies that can facilitate the development of smart grid recommender systems. An outline on potential application scenarios of smart grid recommender systems as well as future challenges are also provided.

With the rapid development of electronic commerce (E-commerce), information overload has become an issue in people's daily lives. Personalized products and services have thus drawn wide attentions, and personalized recommendation techniques provide effective tools to capture the user's interests and find out most relevant information to the user. In this paper, a new personalized recommendation algorithm based on improved twin support vector machine (TWSVM) is proposed. Firstly, we introduce the smoothing techniques to TWSVM (STWSVM). Then, the primal quadratic programming problems of TWSVM are transformed to be smooth unconstrained minimization problems. Followed by this, we apply the sample dynamic update strategy and STWSVM on the personalized recommendation, and compare the proposed method with conventional methods including the correlation-based, back propagation (BP)-based, and SVM-based recommended methods. The experimental results show that the proposed method has superior performance than the other methods.

Personalized recommendation system (PRS) is an effective tool to automatically extract meaningful information from the big data of the users. Collaborative filtering is one of the most widely used personalized recommendation techniques to recommend the personalized products for users. In this paper, a PRS model based on the support vector machine (SVM) is proposed. The proposed model not only considers the items’ content information, but also the users’ demographic and behavior information to fully capture the users’ interests and preferences. Meanwhile, an improved particle swarm optimization (PSO) algorithm is applied to optimize the SVM’s learning parameters. The efficiency of the proposed method is verified by multiple benchmark datasets.

The open nature of recommender systems makes them vulnerable to shilling attacks. Biased ratings are introduced in order to affect recommendations, have been shown to cause great harm to collaborative filtering algorithms. Most of previous research focuses on the differences between genuine profiles and attack profiles, ignoring the group characteristics in an attack. There exists class unbalance problems in SVM based detecting methods, that is, the detecting performance is not good when the amount of samples of attack profiles in training set is small. In this paper, we study the use of SVM based method and group characteristics in attack profiles to detect attack profiles. Based on this, a two phase detecting method SVM-TIA is proposed. In the first phase, Borderline-SMOTE method is used to alleviate the class unbalance problem in classification; a rough detecting result is obtained in this phase; the second phase is a fine-tuning phase whereby the target items in the potential attack profiles set are analysed. We conduct experiments on the MovieLens 100K Dataset and compare the performance of SVM-TIA with other shilling detecting methods to demonstrate the effectiveness of the proposed approach.