Search

Transfer learning methods exploit similarities between different datasets to improve the performance of the target task by transferring knowledge from source tasks to the target task. "What to transfer" is a main research issue in transfer learning. The existing transfer learning method generally needs to acquire the shared parameters by integrating human knowledge. However, in many real applications, an understanding of which parameters can be shared is unknown beforehand. Transfer learning model is essentially a special multi-objective optimization problem. Consequently, this paper proposes a novel auto-sharing parameter technique for transfer learning based on multi-objective optimization and solves the optimization problem by using a multi-swarm particle swarm optimizer. Each task objective is simultaneously optimized by a sub-swarm. The current best particle from the sub-swarm of the target task is used to guide the search of particles of the source tasks and vice versa. The target task and source task are jointly solved by sharing the information of the best particle, which works as an inductive bias. Experiments are carried out to evaluate the proposed algorithm on several synthetic data sets and two real-world data sets of a school data set and a landmine data set, which show that the proposed algorithm is effective. [ABSTRACT FROM AUTHOR]

The stagnation phenomena regularly happen in solving complex optimization problems by a class of computational intelligence algorithms which lack theoretical guidance for their parameters setting. The search characteristics of these algorithms are analyzed in this paper, based on which an adaptive computational intelligence optimization algorithm called cloud drops algorithm is proposed by adopting a cloud model to express the randomness and fuzziness of its search process. The cloud drops algorithm is characterized by representing, mining and recreating the uncertain knowledge about its search process for the optimal solution. None of search parameters are predefined in the proposed algorithm and, whatever the initial solution set is, the whole system can adaptively approach to the global optimum. Based on the theory of stochastic processes, the almost sure convergence of the proposed algorithm is proved under certain conditions by introducing a martingale approach into traditional Markov Chain analysis. Two benchmark problems are tested with the proposed algorithm and the other two existing algorithms as a comparison. The results show that the proposed algorithm has faster convergence speed, better self-adaptability, and stronger ability to deal with stagnation phenomena effectively. [ABSTRACT FROM AUTHOR]