Space-time video super-resolution (STVSR) aims to generate videos with higher resolution (HR) and higher frame rate (HFR). In this paper, we propose a dual-attention STVSR model (DAST) to combine local features with global dependencies. DAST is a compact one-stage STVSR method that tightly combines spatial video super-resolution (S-VSR) and temporal video super-resolution (T-VSR) tasks. Specifically, we design a dual-attention network (DANet). First, we design a high-frequency filter attention block (HFAB) to extract local spatial-temporal information from video frames and enhance the model’s ability to capture local details. Then, the global dependencies in the features are captured by multi-head self attention (MSA) to construct global context information. In addition, a nonlinear residual block (NLRB) is designed to enhance the expressiveness of the model and make the input features more adaptable to each layer of the network. Extensive experiments demonstrate that our method achieves better performance in the STVSR task.

Video super-resolution (VSR) aims to reconstruct blurry low-resolution videos into explicit high-resolution videos.Most existing VSR networks have complex structures and high computational complexity, making it difficult to achieve real-time video processing and deployment on devices with limited computing performance.This paper proposes a lightweight video super-resolution network based on pyramid optical flow extraction and alignment called POFVSR. Specifically, the proposed POFVSR adopts an efficient frame-recurrent network framework. To enable the network to have faster video super-resolution processing capabilities, an efficient pyramid optical flow extraction module and a novel optical flow alignment method are designed to reduce the computational complexity of the POFVSR network while improving its optical flow extraction and alignment performance. Additionally, we employ efficient network structures and upsampling methods for feature fusion and reconstruction. Extensive experiments demonstrate that our proposed POFVSR surpasses existing mainstream lightweight video super-resolution networks regarding strong quantitative performance and visual qualities.

Neural Architecture Search (NAS) aims to identify high-performance networks within a defined search space. Training-free metrics have been proposed to estimate network performance without actual training, reducing NAS deployment costs. However, individual training-free metrics often capture only partial architectural features, and their estimation capabilities are different in various tasks. Combining multiple training-free metrics has been explored to enhance scalability across tasks. Yet, these methods typically optimize global metric combinations over the entire search space, overlooking the varying sensitivities of different architectures to specific metrics, which may limit the final architectures' performance. To address these challenges, we propose the Per-Architecture Training-Free Metric Optimization NAS (PO-NAS) algorithm. This algorithm: (a) Integrates multiple training-free metrics as auxiliary scores, dynamically optimizing their combinations using limited real-time training data, without relying on benchmarks; (b) Individually optimizes metric combinations for each architecture; (c) Integrates an evolutionary algorithm that leverages efficient predictions from surrogate models, enhancing search efficiency in large search spaces. Notably, PO-NAS combines the efficiency of training-free search with the robust performance of training-based evaluations. Extensive experiments demonstrate the effectiveness of our approach. Our code has been made publicly available at https://github.com/LMZ-Zhuo/PO-NAS.

Traditional bus systems with fixed routes and timetables struggle to accommodate dynamic and diverse passenger demands. Demand-Responsive Transit (DRT) offers a flexible solution through dynamic route planning. However, multi-route cooperative scheduling faces challenges such as high combinatorial optimization complexity and insufficient real-time responsiveness. We propose a Multi-Agent Deep Reinforcement Learning framework for cooperative optimization in dynamic multi-route DRT scheduling (MARL-DRT). The problem is modeled as a multi-agent Markov Decision Process (MDP) aimed at minimizing a weighted total cost, including operating costs, passenger waiting costs, trip cancellations, and real-time demand profit. We employ the Multi-Actor-Attention-Critic (MAAC) algorithm to solve the problem, where each agent dynamically generates station sequences through a policy network based on an encoder-decoder structure. A centralized critic network and the policy gradient method are used to improve global cooperation and scheduling stability. Extensive experiments on real-world and benchmark networks demonstrate that our algorithm outperforms baseline methods in total cost, responsiveness, and service quality, providing a more efficient DRT system with lower operational costs and higher passenger satisfaction.

In dealing with the expensive multiobjective optimization problem, some algorithms convert it into a number of single-objective subproblems for optimization. At each iteration, these algorithms conduct surrogate-assisted optimization on one or multiple subproblems. However, these subproblems may be unnecessary or resolved. Operating on such subproblems can cause server inefficiencies, especially in the case of expensive optimization. To overcome this shortcoming, we propose an adaptive subproblem selection (ASS) strategy to identify the most promising subproblems for further modeling. To better leverage the cross information between the subproblems, we use the collaborative multioutput Gaussian process surrogate to model them jointly. Moreover, the commonly used acquisition functions (also known as infill criteria) are investigated in this article. Our analysis reveals that these acquisition functions may cause severe imbalances between exploitation and exploration in multiobjective optimization scenarios. Consequently, we develop a new acquisition function, namely, adaptive lower confidence bound (ALCB), to cope with it. The experimental results on three different sets of benchmark problems indicate that our proposed algorithm is competitive. Beyond that, we also quantitatively validate the effectiveness of the ASS strategy, the CoMOGP model, and the ALCB acquisition function.

Objective value estimation based on computationally efficient surrogate models is widely used to reduce the computational cost in solving expensive multiobjective optimization problems (MOPs). However, due to the scarcity of training data and the lack of data sharing between training tasks in a surrogate-based system, the estimation effectiveness of the surrogate models might not be satisfactory. In this study, we present a novel surrogate methodology based on information transfer to deal with this problem. Particularly, in the proposed framework, the objectives of an MOP that may have little apparent similarity or correlation are linearly mapped to a number of related tasks. Afterward, the related tasks are used to train a multitask Gaussian process (MTGP). MTGP expands the training data leading to more confident learning of the parameters of the model. The predicted values of the objective functions can be obtained by a reverse mapping from the learned MTGP model. In this way, the computational burden of the expensive objective functions of an MOP can be substantially reduced while maintaining good estimation accuracy. MTGP facilitates mutual information transfer across tasks, avoids learning from scratch for new tasks, and captures the underlying structural information between tasks. The proposed surrogate approach is merged into MOEA/D to address MOPs. Experimental tests under various scenarios indicate that the resultant algorithm outperforms other state-of-the-art surrogate-based multiobjective optimization algorithms.

Video super-resolution (VSR) aims at recovering a natural and realistic high-resolution (HR) video frame from the corre-sponding low-resolution (LR) counterpart and its consecutive neighboring frames. The challenge is how to make full use of spatio-temporal coherence among the input LR frames. In this work, we propose a self-guided deformable convolutional gated recurrent unit (GRU) framework for VSR. Specifically, convolutional GRU can efficiently extract temporal features of the input LR frames. Deformable convolution (DConv) is utilized to spatially align the hidden states of GRU cells with the input feature maps. Moreover, we argue that the reference LR frame itself is efficient to guide the aggregated features learning frame-specific feature maps, which are then used to generate rich and more realistic textures towards the corresponding HR frame. Extensive experimental results on benchmark datasets demonstrate that the proposed framework achieves better performance than state-of-the-art methods and has higher model efficiency.

The challenge of Video super-resolution (VSR) is how to make full use of the spatial-temporal coherence among neigh-bouring LR frames to generate high-resolution (HR) prediction. In this study, we propose to use transformer on VSR to capture long-range temporal dependencies. Specifically, we first spatially divide LR images into patches and split each patch into sub-patches. Transformer encoders are applied to both the patches and sub-patches, such that the self-attention modules can extract both global and local correlations. To accelerate the training process and filter out irrelevant features, we only select top-k similar features for the attention scheme. We then feed the extracted long-range correlations into a temporal, spatial and channel attention fusion mod-ule’ which enhances the useful information along all three di-mensions' respectively. Extensive experiments on benchmark datasets show that the proposed model outperforms state-of-the-art VSR methods in terms of PSNR/SSIM values and vi-sual qualities.

Multiple-related tasks can be learned simultaneously by sharing information among tasks to avoid tabula rasa learning and to improve performance in the no transfer case (i.e., when each task learns in isolation). This study investigates multitask learning with conditional neural process (CNP) networks and proposes two multitask learning network models on the basis of CNPs, namely, the one-to-many multitask CNP (OMc-MTCNP) and the many-to-many MTCNP (MMc-MTCNP). Compared with existing multitask models, the proposed models add an extensible correlation learning layer to learn the correlation among tasks. Moreover, the proposed multitask CNP (MTCNP) networks are regarded as surrogate models and applied to a Bayesian optimization framework to replace the Gaussian process (GP) to avoid the complex covariance calculation. The proposed Bayesian optimization framework simultaneously infers multiple tasks by utilizing the possible dependencies among them to share knowledge across tasks. The proposed surrogate models augment the observed dataset with a number of related tasks to estimate model parameters confidently. The experimental studies under several scenarios indicate that the proposed algorithms are competitive in performance compared with GP-, single-task-, and other multitask model-based Bayesian optimization methods.

Neural architecture search (NAS) is an exciting new field in automating machine learning. It can automatically search for the architecture of neural networks. But the current NAS has extremely high requirements for hardware equipment and time costs. In this work, we propose a predictor based on Radial basis function neural network (RBFNN) as a surrogate model of Bayesian optimization to predict the performance of neural architecture. The existing work does not consider the difficulty of directly searching for neural architectures that meet the performance requirements of NAS in real-world applications. Meanwhile, NAS needs to execute multiple times independently when facing multiple similar tasks. Therefore, we further propose a multi-task learning surrogate model with multiple RBFNNs. The model not only functions as a predictor, but also learns knowledge of similar tasks jointly. The performance of NAS is improved by processing multiple tasks simultaneously. Also, the current NAS is committed to searching for very high-performance networks and does not take into account that neural architectures are limited by device memory during actual deployment. The scale of architecture also needs to be considered. We use a multi-objective optimization algorithm to simultaneously balance the performance and the scale, and build a multi-objective evolutionary search framework to find the Pareto optimal front. Once the NAS is completed, decision-makers can choose the appropriate architecture for deployment according to different performance requirements and hardware conditions. Compared with existing NAS work, our proposed MT-ENAS algorithm is able to find a neural architecture with competitive performance and smaller scale in a shorter time.

Neural Architecture Search (NAS) is studied to automatically design the deep neural network structure, freeing people from heavy network design tasks. Traditional NAS based on individual performance evaluation needs to train many networks generated by the search, and compare the performance of the networks according to their accuracy, which is very time-consuming. In this study, we propose to use a two-category comparator based random forest model as a surrogate to estimate the accuracy of the networks. thereby reducing heavy network training process and greatly saving search time. Instead of directly predicting the accuracy of each network, we propose to compare the relative performance between each two networks in our proposed two-category comparator. Furthermore, we implement the modeling process of the surrogate model in the sampling space of the original training data, which further accelerates the search process of the network in the NAS. Experimental results show that our proposed NAS framework can greatly reduce the search time, while the accuracy of the obtained network is comparable to that of other state-of-the art NAS algorithms.

Video super-resolution (SR) aims at generating high-resolution (HR) frames from consecutive low-resolution (LR) frames. The challenge is how to make use of temporal coherence among neighbouring LR frames. Most previous works use motion estimation and compensation based models. However, their performance relies heavily on motion estimation accuracy. In this paper, we propose a multi-scale pyramid 3D convolutional (MP3D) network for video SR, where 3D convolution can explore temporal correlation directly without explicit motion compensation. Specifically, we first apply 3D convolution into a pyramid subnet to extractmulti-scale spatial and temporal features simultaneously from the LR frames, such that it can handle various sizes of motions. We then feed the fused feature maps into an SR reconstruction subnet, where a 3D sub-pixel convolution layer is used for up-sampling. Finally, we append a detail refinement subnet based on the encoder-decoder structure to further enhance texture details of the reconstructed HR frames. Extensive experiments on benchmark datasets and real-world cases show that the proposed MP3D model outperforms state-of-the-art video SR methods in terms of PSNR/SSIM values, visual quality and temporal consistency, respectively.

A novel many-objective particle swarm optimization (PSO) algorithm called IDMOPSO is presented in this study to robustly and effectively address many-objective optimization problems (MaOPs). IDMOPSO is based on a performance indicator and direction vectors. A selection strategy based on the quality indicator Iε+ and Pareto dominance for personal best (pbest) particles is proposed to ensure the convergence and diversity of the algorithm and enhance the capability of local exploration. An external archive based on Iε+ and direction vectors is used to preserve the diversity of non-dominated solutions found in the search process. A multi-global optimal (gbest) particle selection method is developed to increase global search ability and ensure the particles' diversity. This method allows each particle to be assigned to a different gbest particle. This method differs from the traditional method, wherein only one gbest particle is allocated for the whole population of PSO. We aim to design a robust multi-objective evolutionary algorithm to deal with MaOPs. Extensive comparative experiments on DTLZ and DTLZ-1problems with varied numbers of objectives show that IDMOPSO is effective and flexible in addressing MaOPs. The influences and effectiveness of the proposed strategies are also analyzed in detail.

This paper proposes a Gaussian process (GP) based co-sub-Pareto front surrogate augmentation strategy for evolutionary optimization of computationally expensive multiobjective problems. In the proposed algorithm, a multiobjective problem is decomposed into a number of subproblems, the solution of each of which is used to approximate a portion or sector of the Pareto front (i.e., a subPF). Thereafter, a multitask GP model is incorporated to exploit the correlations across the subproblems via joint surrogate model learning. A novel criterion for the utility function is defined on the surrogate landscape to determine the next candidate solution for evaluation using the actual expensive objectives. In addition, a new management strategy for the evaluated solutions is presented for model building. The novel feature of our approach is that it infers multiple subproblems jointly by exploiting the possible dependencies between them, such that knowledge can be transferred across subPFs approximated by the subproblems. Experimental studies under several scenarios indicate that the proposed algorithm outperforms state-of-the-art multiobjective evolutionary algorithms for expensive problems. The parameter sensitivity and effectiveness of the proposed algorithm are analyzed in detail.

Artificial Bee Colony (ABC) algorithm is an optimization algorithm based on a particular intelligent behaviour of honeybee swarms. The standard ABC is weak at the local-search capability and precision. Extremal Optimization (EO) is a general-purpose heuristic method which has strong local-search capability and has been successfully applied to a wide variety of hard optimization problems. In order to strengthen the local-search capability of ABC, this work proposes a novel hybrid optimization method, called ABC-EO algorithm, through introducing EO to ABC. The simulation results show that the performance of the proposed method is as good as or superior to those of the state-of-the-art algorithms in complex numerical optimization problems.