Polynomial multiplication is a critical performance bottleneck in lattice-based cryptographic schemes, in which efficient and flexible hardware implementations are essential. This paper presents a high-performance and reconfigurable polynomial multiplication unit that integrates the cooptimization of algorithmic and architectural optimizations. As the core operation of polynomial multiplication, we propose the Karatsuba-based high-efficiency radix-4 number theoretic transform (KHNTT) algorithm, which reduces the computational complexity by leveraging low-bit-width arithmetic and parallel processing techniques. Based on this algorithm, we develop a unified radix-4 butterfly unit capable of supporting modular polynomial operations under various parameter configurations, and to further improve the computational efficiency and architectural flexibility, key algorithm modules such as modular reduction and data flow scheduling are deeply optimized. Furthermore, we introduce an iterative interleaving memory strategy combined with a bank resource sharing mechanism to eliminate memory access conflicts and improve data access efficiency. The proposed polynomial multiplication unit allows flexible configuration of the butterfly units and supports polynomial computations of various degrees without the need for recompilation. The FPGA implementation results demonstrate that, compared with existing solutions, our architecture achieves a$1.25\times $to$5.27\times $improvement in the polynomial multiplication throughput and a$1.14\times $to$2.7\times $enhancement in the area efficiency, fully leveraging the computational advantages of FPGAs.

MOTIVATION: Protein-protein interactions play a fundamental role in biological systems. Accurate detection of protein-protein interaction sites (PPIs) remains a challenge. And, the methods of PPIs prediction based on biological experiments are expensive. Recently, a lot of computation-based methods have been developed and made great progress. However, current computational methods only focus on one form of protein, using only protein spatial conformation or primary sequence. And, the protein's natural hierarchical structure is ignored. RESULTS: In this study, we propose a novel network architecture, HSSPPI, through hierarchical and spatial-sequential modeling of protein for PPIs prediction. In this network, we represent protein as a hierarchical graph, in which a node in the protein is a residue (residue-level graph) and a node in the residue is an atom (atom-level graph). Moreover, we design a spatial-sequential block for capturing complex interaction relationships from spatial and sequential forms of protein. We evaluate HSSPPI on public benchmark datasets and the predicting results outperform the comparative models. This indicates the effectiveness of hierarchical protein modeling and also illustrates that HSSPPI has a strong feature extraction ability by considering spatial and sequential information simultaneously. AVAILABILITY AND IMPLEMENTATION: The code of HSSPPI is available at https://github.com/biolushuai/Hierarchical-Spatial-Sequential-Modeling-of-Protein.

Abstract Binary vulnerability detection plays an important role in the field of program security. In order to deal with large-scale vulnerability detection tasks, more and more neural network technologies are applied to cross-architectures vulnerability detection. These technologies have significantly improved the accuracy of vulnerability detection. However, existing methods still face problems such as single extracted information, poor robustness against compilation optimization, and inability to perform cross-architectures vulnerability detection. Therefore, this paper proposes a cross-architectures vulnerability detection method based on attention mechanism and multi-feature fusion. This method can simultaneously obtain information such as assembly code, attribute control flow graph and function-level features for cross-architecture, cross-compilers and cross-optimization options vulnerability detection. Adding attention mechanism to GRU and GoogleNet improves the model to obtain semantic information and attribute information after the fusion of basic block-level and function-level features, and searches for Top-K suspected vulnerability functions and graph matching through deep neural network model to perform phased vulnerability detection. The experimental results show that this method achieves an accuracy of 95.79% and a Recall of 97.06%, which is better than the existing methods and performs well in vulnerability detection in real environments.

Code randomization, aka software diversification, is an effective way to mitigate code-reuse attacks. This mechanism diversifies the target software into heterogeneous variants, making a specific attack chain unfeasible for the transformed variants. Enhancing the heterogeneity of these variants is important for this method to reach its expected security level. Additionally, limiting their execution overhead is necessary to ensure the availability of this protection. However, finding the optimal subset among a large number of diversified variants is computationally difficult. This creates a dilemma in software diversification schema where enhancing heterogeneity and reducing execution overhead are both desired.To ensure a determined code quality control in generating software variants, we propose a guided software diversification mechanism (called GuiDiv). GuiDiv formalizes the iterate-used transformations into a branching process of a tree (called DivTree) and introduces an optimization process (called nodemerging) to guide the diversification. The optimizer evaluates the intermediate compilation results using multi-target evaluation functions in each iteration. This process aims to optimize the contribution of structural heterogeneity from redundant instructions, allowing variants with higher structural dissimilarity and lower overhead to have a higher chance of participating in the next iteration. We developed a proof-of-concept compilation module and used OpenSSL as the performance benchmark. In the evaluations, compared to related schemes, GuiDiv can bring higher control flow dissimilarity in most test cases. Regarding the variant heterogeneity, variants generated by GuiDiv exhibit significantly lower execution overhead.

In this paper, we address the problem of multi-view clustering (MVC), integrating the close relationships among views to learn a consistent clustering result, via triplex information maximization (TIM). TIM works by proposing three essential principles, each of which is realized by a formulation of maximization of mutual information. 1) Principle 1: Contained. The first and foremost thing for MVC is to fully employ the self-contained information in each view. 2) Principle 2: Complementary. The feature-level complementary information across pairwise views should be first quantified and then integrated for improving clustering. 3) Principle 3: Compatible. The rich cluster-level shared compatible information among individual clustering of each view is significant for ensuring a better final consistent result. Following these principles, TIM can enjoy the best of view-specific, cross-view feature-level, and cross-view cluster-level information within/among views. For principle 2, we design an automatic view correlation learning (AVCL) mechanism to quantify how much complementary information across views by learning the cross-view weights between pairwise views automatically, instead of view-specific weights as most existing MVCs do. Specifically, we propose two different strategies for AVCL, i.e., feature-based and cluster-based strategy, for effective cross-view weight learning, thus leading to two versions of our method, TIM-F and TIM-C, respectively. We further present a two-stage method for optimization of the proposed methods, followed by the theoretical convergence and complexity analysis. Extensive experimental results suggest the effectiveness and superiority of our methods over many state-of-the-art methods.

Emotion recognition in conversation is one of the essential tasks of natural language processing. However, this task’s annotation data is insufficient since such data is hard to collect and annotate. Meanwhile, there is large-scale data for conversational generation, and this data does not need annotation manually. But, whether the vector space between different datasets is similar will be a problem. Therefore, we utilize a same dataset to train the conversational generator and the classifier, and transfer knowledge between them. In particular, we propose an Emotion Recognition with Conversational Generation Transfer (ERCGT) framework to model the interaction among utterances by transfer learning. First, we train a conversational generator. In the second step, a transfer learning model is used to transfer the knowledge of generator to the emotion recognition model. Empirical studies illustrate the effectiveness of the proposed framework over several strong baselines on three benchmark emotion classification datasets.

Attacks launched from the inside of the cloud are threats not only to the cloud users but also to the cloud infrastructures. Although with trusted computing the cloud service providers can guarantee the trust and security of the cloud environment for the users, the trustworthiness of users is not properly assessed. Inspired by the concept of variable trust, the main contribution of this paper is that we propose a trust assessment method for cloud users based on deep learning. Firstly, we extract users’ activities from system logs and employ stacked LSTM (long short-term memory) neural network to model normal activity patterns to build trust profiles for different users. Secondly, the trust profile is capable of predicting future behavioural actions of the specific user, and by calculating the similarity between predicted actions and actual actions the trustworthiness of the user will be assessed with a baseline to detect the trust state of the cloud user dynamically. And in the end, we design and conduct experiments on a public dataset. The results of experiments indicate that when the user is in abnormal state, there are notable differences between predicted actions and user’s actual actions, which proves the efficiency of the proposed method.

Summary In the field of bioinformatics and DNA computing, simulated hybridization experiments can replace real molecular hybridization experiments to some extent, avoiding some disadvantages of the actual experimental design. However, the core techniques, which are employed by the popular DNA simulation software, are limited to the exponential computational complexity of the combinatorial problems. As a result, it is impossible to decide whether a specific hybridization among complex DNA molecules is effective or not within acceptable time. To address this common problem, we hereby introduce a new method based on the machine learning technique. First, a sample set is employed to train the boosted tree algorithm, which resulted in a corresponding machine learning model. Second, this model is applied to predict the classification results of molecular hybridization for a given group of DNA molecular coding. The experiment results showed that the new method had an average accuracy level of 94.2% and an average efficiency level 90 839 times higher than that of the existing representative approaches. Especially for the case study in this paper, the efficiency of the new method is 235 000, 250 000, and 990 000 times higher than that of the three existing methods, respectively. These experimental results indicate that our new approach can quickly and accurately determine the biological effectiveness of molecular hybridization for a given DNA design.

A current research trend is to combine multimedia data with artificial intelligence and process them on cloud servers. In this context, ensuring the security of multimedia cloud servers is critical, and the cyber mimic defence (CMD) technology is a promising approach to this end. CMD, which is an innovative active defence technology developed in China, can be applied in many scenarios. However, although the mathematical model is a key component of CMD, a universally acceptable mathematical model for theoretical CMD has not been established yet. In this work, the attack problems and modelling difficulties were extensively examined, and a comprehensive modelling theory and concepts were clarified. By decoupling the model from the input and output of the specific system scene, the modelling difficulties were effectively avoided, and the mathematical expression of the CMD mechanism was enhanced. Furthermore, the process characteristics of the attack behaviour were identified by using a specific mathematical mapping method. Finally, based on the decomposition problem of large prime factors and convolution operations, an intuitive and exclusive CMD mathematical model was proposed. The proposed model could clearly express the CMD mechanism and transform the problems of attack and defence in the CMD domain into corresponding mathematical problems. These aspects were considered to qualitatively assess the CMD security, and it was noted that a high level of security can be realized. Furthermore, the overhead of CMD was analyzed. Moreover, the proposed model can be directly programmed.