As a computing accelerator, a large-scale photonic spatial Ising machine has great advantages and potential due to its excellent scalability and compactness. However, the current fundamental limitation of a photonic spatial Ising machine is the configuration flexibility for problem implementation in the accelerator model. Arbitrary spin interactions are highly desired for solving various non-deterministic polynomial (NP)-hard problems. In this paper, we propose a novel quadrature photonic spatial Ising machine to break through the limitation of the photonic Ising accelerator by synchronous phase manipulation in two sections. The max-cut problem solution with a graph order of 100 and density from 0.5 to 1 is experimentally demonstrated after almost 100 iterations. Our work suggests flexible problem solving by the large-scale photonic spatial Ising machine.

No abstract

Federated Learning (FL) has become an active and promising distributed machine learning paradigm. As a result of statistical heterogeneity, recent studies clearly show that the performance of popular FL methods (e.g., FedAvg) deteriorates dramatically due to the client drift caused by local updates. This paper proposes a novel Federated Learning algorithm (called IGFL), which leverages both Individual and Group behaviors to mimic distribution, thereby improving the ability to deal with heterogeneity. Unlike existing FL methods, our IGFL can be applied to both client and server optimization. As a by-product, we propose a new attention-based federated learning in the server optimization of IGFL. To the best of our knowledge, this is the first time to incorporate attention mechanisms into federated optimization. We conduct extensive experiments and show that IGFL can significantly improve the performance of existing federated learning methods. Especially when the distributions of data among individuals are diverse, IGFL can improve the classification accuracy by about 13% compared with prior baselines.

No abstract

This paper proposes an extended VlseKriterijuska Optimizacija I Komoromisno Resenje (VIKOR) method based on the particle swarm optimization (PSO) algorithm for solving multicriteria group decision-making problems with probabilistic linguistic information. First, we define the novel operations of probabilistic linguistic term sets and then prove the corresponding properties.Second, we apply a modified PSO algorithm to the consensus reaching process to improve the collective consensus level. In the context of probabilistic linguistic information, each participant can be recognized as a particle moving towards the best position. The consensus level can be regarded as the objective function that is used to construct the fitness function. In the update function, the trust relationship and the similarity measure between experts are exploited to determine the adjustment coefficient. The new consensus model based on PSO can ensure that the ultimate evaluation achieves a high level of consensus. Afterward, we propose the extended VIKOR method to obtain the optimal solution, which not only avoids the loss of decision information, but also considers the separation of each alternative from the positive ideal solution and the negative ideal solution when criteria are interactive. The advantages of