Search

Crowdsourcing platforms provide an efficient and cost-effective means to acquire the extensive labeled data necessary for supervised learning. However, the labels provided by untrained crowdsourcing workers often contain a considerable amount of noise. Although the application of ground truth inference algorithms to deduce integrated labels effectively enhances label quality, a certain level of noise persists. To further diminish the noise within crowdsourced labeling, this paper introduces a novel Small Loss-based Noise Correction algorithm (SLNC). SLNC first filters the crowdsourced data, leveraging the characteristic of neural networks to preferentially fits clean samples, thereby obtaining relatively clean and noisy sets. It then employs data augmentation techniques to enhance the clean set and subsequently trains the corrector on this augmented set to rectify the noisy set. SLNC has been evaluated using 16 simulated and two real-world datasets. The results indicate that SLNC surpasses comparative algorithms in the quality of the final labels.

The optimal foraging algorithm (OFA) was proposed by summarizing the rules of the animal foraging behavior in a group. Therefore OFA also has the defects of the swarm intelligence algorithm, such as easy to trap into local optimum and low convergence accuracy. In order to overcome these defects, an optimal foraging algorithm based on differential evolution (DEOFA) is proposed. The differential evolution mechanism contains mutation and crossover operators. The mutation and crossover operators are used to accelerate the convergence speed and global search capability of the OFA. The mutation operator is adopted to perform mutation operations centered on the optimal individual of each iteration to raise the convergence accuracy of the OFA. The test results of 30 benchmark functions show that the performance of DEOFA is better than nine compared algorithms in search accuracy, convergence speed and robustness. In order to verify the effectiveness of the DEOFA in solving practical problems, DEOFA is applied to solve the 0-1 knapsack problem. The test results in the six examples of 0-1 knapsack problems indicate that the DEOFA achieves better performance in accuracy, stability and high dimension.

Many optimization algorithms have been proposed to solve hybrid flowshop scheduling problem (HFSP). However, with the development of industry and society, labor right and labor safety have become important problem to consider in production scheduling. So the green HFSP considering makespan, noise and dust pollution becomes an urgent problem to be solved. In this paper, the rider optimization algorithm (ROA) is modified into the multi-objective rider optimization algorithm (MOROA) using Pareto archive and neighborhood sorting techniques. The Pareto archive and neighborhood sorting technology make the Pareto optimal solution set of MOROA have higher coverage and more solutions. Then MOROA is discretized into discrete MOROA (DMOROA) to solve the HFSP considering makespan, noise and dust pollution. DMOROA is tested on 10, 30 and 50 jobs HFSP considering makespan, noise and dust pollution. The test results are compared with two multi-objective algorithms to verify the performance of DMOROA. And the test results verify that the DMOROA is superior to the comparison algorithms in search accuracy, number of non-dominated solutions, diversity of solution set and stability. Therefore, DMOROA is effective in solving multi-objective HFSP considering makespan, noise and dust pollution.

Characterized by less parameter settings, easy implementation, and strong optimization capacity, the crow search algorithm has been successfully applied to solve the optimization problem. As the basic crow search algorithm is a new kind of swarm intelligent algorithm only based on the crow's memory foraging mode, it also contains defects like slow search speed and low optimization precision in later iterations, which are especially obvious for the optimization of high-dimensional functions. In order to overcome these shortcomings, a new crow search algorithm based on neighborhood search of non-inferior solution set (NICSA) is proposed. The proposed algorithm makes the crow individual choose the memory search mode or neighborhood search mode automatically in the course of evolution by the determination factor of non-inferior solution. With this strategy, the local exploitation and the global exploration of the algorithm became more balanced. In the neighborhood search, the selectivity factor is used to guide non-inferior solutions to adaptively execute neighborhood search of Levy flight or Gaussian flight, to enhance the neighborhood searchability of the algorithm and improve the optimization precision. The result of simulation experiments with 23 benchmark test functions verifies that the proposed algorithm has good optimization effect in the aspects of search veracity, convergence rate and robustness.