Proteins serve as the essential executors of cellular activities, and their mislocalization often resulting in diverse diseases. Traditional methods for determining protein subcellular localization are noted for their time-consuming, labor-intensive, and complex nature. To address these challenges, this study introduces SubLoc, a deep learning-based algorithm for predicting protein subcellular localization. The methodology comprises three key steps: Firstly, leveraging the deep pretrained protein language model ProtT5 to derive protein embedding vectors, thereby capturing intricate patterns and biological functionalities of protein sequences. Secondly, constructing a 3D protein structure graph model using amino acid residue contact relationships within the sequence, which is subsequently processed by a graph convolutional network to effectively manage spatial structural information. Lastly, employing a bidirectional gated recurrent unit and multi-head attention mechanism to analyze sequence features, integrating both structural and sequence data for enhanced subcellular localization prediction. Experimental results demonstrate that SubLoc exhibits exceptional performance in localizing proteins across 10 subcellular compartments, outperforming all comparative methods in terms of precision, recall, and MCC average values. Notably, SubLoc achieves particularly notable results in identifying Cytoplasm and Mitochondrion locations.

To reduce traffic accidents caused by drowsy driving, computer vision techniques have been widely adopted to analyze facial cues and assess driver fatigue levels. However, existing methods suffer from several limitations. Most approaches extract features from only one or two facial regions (e.g., the eyes and mouth), which restricts their ability to capture variations in drowsiness expression across different fatigue levels within the same driver, as well as individual differences among drivers. Moreover, although facial regions are inherently structured, current models typically lack structural awareness, resulting in weakly structured feature representations that may lose subtle drowsiness-related details. To address these issues, we propose a Feature-Enhanced Channel Topology Graph Convolutional Network (FE-CTGCN). The proposed framework consists of three key modules: the Feature-Enhanced Global–Local Information Module (FEGL), the Multi-source Information Representation Module (MSIR), and the Channel-wise Topology Graph Convolution Module (CTGCN). The FEGL module extracts visual features from five local facial regions and the entire face, with an emphasis on enhancing discriminative patterns across different fatigue states. The MSIR module introduces an attention-based feature fusion mechanism that not only integrates multi-source features via attention but also captures temporal dynamics, enabling effective modeling of correlations between global and local facial cues. Together, FEGL and MSIR address intra-driver variations across drowsiness levels and inter-driver differences. In addition, the CTGCN module constructs a topology-aware graph where nodes correspond to the five local facial features and the global facial representation. By modeling spatial relationships among these nodes, it facilitates structured information exchange and builds a strongly structured facial feature space that enhances internal feature integration. Experimental results demonstrate that FE-CTGCN achieves superior detection performance compared to existing methods, validating its effectiveness for driver drowsiness detection. To facilitate reproducibility and further research, the source code is available athttps://github.com/zzs-code/FECTGCN.git

Promoter components play a critical role in the regulation of gene expression, directly determining the expression intensity of downstream target genes. Opting for high-quality promoters is essential for synthetic biology. In the existing literature, the method of designing high-quality promoters using Generative Adversarial Networks (GANs) is constrained by training difficulties, mode collapse, implicit generation, and information loss in deep networks after vector representation of promoters. Diffusion models are more easily trainable, possess elegant mathematical explanations, and can directly model the target distribution, potentially overcoming the issues faced by GANs mentioned above. In this paper, we propose a new method for promoter design, PromoterDiff, based on the diffusion model. Specifically, we adjusted the convolutional network structure of the diffusion model and introduced a bridging structure to adapt to the diffusion and reconstruction steps of the diffusion model when dealing with DNA sequence tensors containing a large number of zero elements. We used natural promoters of Escherichia coli as the training set to train the model, resulting in the successful design of 14,080 entirely new promoters. Through the analysis of the motifs, k-mer frequencies, and motif spacing constraints of these promoters, we confirmed that the diffusion model can capture the characteristics of natural promoters, and the quality of the generated promoters surpasses existing models. Through biological experiments, it has been confirmed that 83% of the promoter sequences designed by PromoterDiff surpass the activity of natural promoters, exceeding the current leading model by 13%. This underscores the immense potential of applying the diffusion model to de novo promoter design. This paper also provides a clear mathematical representation of the de novo promoter design task. According to our survey, this is the first time the diffusion model has been applied in the field of de novo promoter design.

Knowledge tracing(KT) is one of the supporting technologies for adaptive learning. It obtains learners' knowledge level by analyzing their online historical answer records, thereby predicting their future answer performance. Since there are a large number of continuous or repeated exercises in KT datasets, and traditional graph structures can usually only represent one-to-one or one-to-many relationships. Some existing models tend to ignore the potential connections between knowledge concept, and thus fail to accurately obtain high-level representations of students' knowledge state. Hypergraphs can directly represent many-to-many relationships and have advantages in extracting high-level information and processing complex relationships. In this paper, we propose a KT method (HTKT) based on Hypergraph Transformer. This method is based on the Encoder-only architecture and introduces a hypergraph attention mechanism to obtain a high-order feature representation of students' knowledge state. Experiments on multiple public and general datasets show that our proposed HTKT method outperforms existing baseline methods.

Understanding protein function is crucial for comprehending life at the molecular level. Currently, less than 0.1% of proteins having experimental GO annotations. Traditional experimental methods are time-consuming and expensive. To narrow this gap, employing accurate and efficient computational methods can fill the void in automated protein function prediction (AFP). We have developed a new method for predicting protein function using sequence and predicted structural information. We use the large-scale pretrained language model ESM2 to pretrain protein sequences and an encoder to capture contextual information. Using the 3D structural data of proteins generated by AlphaFold2, combined with the sequence, as input for the Graph Convolutional Neural Network, to infer the probabilities of Gene Ontology (GO) annotations for the proteins. Compared to earlier methods, our model achieves better performance. Evaluations on the human dataset show AUPR improvements of 9%, 9.4%, and 20.7% in the BP, MF, and CC branches, respectively, demonstrating that our model is an effective tool for predicting protein function.

Knowledge tracing is a key technology in online education platforms such as Intelligent Tutoring Systems (ITSs) and Massive Open Online Courses (MOOCs), which model the state of an individual’s knowledge based on the learner’s historical sequence of interactions to predict future performance. However, existing knowledge-tracing approaches lack attention to high- and low-order features and exercise relevance, while few models consider the real-world situation of online tutoring systems where learners can only interact with a limited number of exercises and data is often sparse. Therefore, this paper proposes a knowledge tracing method based on the fusion of exercise relevance and hybrid attention network. Firstly, the parallel GRU captures the high and low order features of learners, which are processed by the extraction network to fit the information representation of learner-interacted exercises. Second, the correlation-enhancing features of target exercises and historical interaction exercises are captured. Finally, a prediction network is used to fuse multiple features in order to effectively capture learners’ knowledge states and improve prediction accuracy in the face of sparse data. Extensive experiments on real online education datasets have shown that ERKT achieves better prediction results compared to existing mainstream methods.

Knowledge Tracing (KT) involves using deep neural networks (DNNs) to track students’ learning progress, but over-fitting can be an issue with small datasets. Adversarial examples have been introduced to improve generalization, but they may overlook individual student differences. To address this issue, we propose a new model called Customized Adversarial Training Knowledge Tracing (CATKT). The model generates unique adversarial perturbations for each sample based on the characteristics of knowledge tracking tasks, thereby better adapting to students’ learning traits and enhancing the effectiveness and accuracy of knowledge tracking tasks. Specifically, CATKT can dynamically adjust the level of perturbations according to the difficulty of the knowledge, adding non-uniform and effective perturbations to each interaction embedding, and replacing the original labels with adaptively smoothed labels to improve task accuracy. Experimental results show that CATKT outperforms previous knowledge tracking methods in terms of performance and provides new ideas and methods for teaching assessment and personalized learning in the education field.

To reduce the repetitive and complex work of instructors, exam paper generation (EPG) technique has become a salient topic in the intelligent education field, which targets at generating high-quality exam paper automatically according to instructor-specified assessment criteria. The current advances utilize the ability of heuristic algorithms to optimize several well-known objective constraints, such as difficulty degree, number of questions, etc., for producing optimal solutions. However, in real scenarios, considering other equally relevant objectives (e.g., distribution of exam scores, skill coverage) is extremely important. Besides, how to develop an automatic multi-objective solution that finds an optimal subset of questions from a huge search space of large- sized question datasets and thus composes a high-quality exam paper is urgent but non-trivial. To this end, we skillfully design a reinforcement learning guided Multi-Objective Exam Paper Generation framework, termed MOEPG, to simultaneously optimize three exam domain-specific objectives including difficulty degree, distribution of exam scores, and skill coverage. Specifically, to accurately measure the skill proficiency of the examinee group, we first employ deep knowledge tracing to model the interaction information between examinees and response logs. We then design the flexible Exam Q-Network, a function approximator, which automatically selects the appropriate question to update the exam paper composition process. Later, MOEPG divides the decision space into multiple subspaces to better guide the updated direction of the exam paper. Through extensive experiments on two real-world datasets, we demonstrate that MOEPG is feasible in addressing the multiple dilemmas of exam paper generation scenario1. 1https://github.com/researcher-tiger/MOEPG

Graph anomaly detection, here, aims to find rare patterns that are significantly different from other nodes. Attributed graphs containing complex structure and attribute information are ubiquitous in our life scenarios such as bank account transaction graph and paper citation graph. Anomalous nodes on attributed graphs show great difference from others in the perspectives of structure and attributes, and give rise to various types of graph anomalies. In this paper, we investigate three types of graph anomalies: local, global, and structure anomalies. And, graph neural networks (GNNs) based anomaly detection methods attract considerable research interests due to the power of modeling attributed graphs. However, the convolution operation of GNNs aggregates neighbors information to represent nodes, which makes node representations more similar and cannot effectively distinguish between normal and anomalous nodes, thus result in sub-optimal results. To improve the performance of anomaly detection, we propose a novel community-aware attributed graph anomaly detection framework (ComGA). We design a tailored deep graph convolutional network (tGCN) to anomaly detection on attributed graphs. Extensive experiments on eight real-life graph datasets demonstrate the effectiveness of ComGA.

Network representation aims to learn low-dimensional vector representations of network nodes while preserving the inherent properties of the network. For all its popularity, majority of the existing methods focus on exploitation of diverse information, including network topology and semantic information on nodes of network, and ignore their implicit semantics. For example, we all know the saying that birds of a feather flock together. More concretely, semantic information of one node can be influenced by its neighbors' semantic information. Furthermore, even two nodes are not directly connected, they may have similar implicit semantic information (i.e., high-order semantic proximity). Thus, they should be close in the represented vector space. To this end, we propose a Deep Semantic Network Representation approach (DSNR) in the self-translation framework from sequence to sequence. To excavate the implicit semantic information of nodes and capture the high-order semantic proximity, three key components make our approach effective, i.e., aggregation of nodes neighbors' semantic information and enhancement to the semantic feature representations of nodes by a deep autoencoder, integration of nodes semantic information in node identity sequence to generate node semantic sequence, and translation from node semantic sequence to node identity sequence to capture the high-order semantic proximity in an attention-enhanced seq2seq framework. Extensive experiments based on three real-world datasets have verified the effectiveness of our proposed approach11Code is available at https://github.com/DASE4/DSNR.