Arash Sharifi

Reinforcement learning (RL) has shown significant potential for dealing with complex decision-making problems. However, its performance relies heavily on t

Unsupervised Domain Adaptation (UDA) techniques in real-world scenarios often encounter limitations due to their reliance on reducing distribution dissimilarity between source and target domains, assuming it leads to effective adaptation. However, they overlook the intricate factors causing domain shifts, including data distribution variations, domain-specific features and nonlinear relationships, thereby hindering robust performance in challenging UDA tasks. The Neuro-Fuzzy Meta-Learning (NF-ML) approach overcomes traditional UDA limitations with its flexible framework that adapts to intricate, nonlinear domain gaps without rigid assumptions. NF-ML enhances domain adaptation by selecting a UDA subset and optimizing their weights via a neuro-fuzzy system, utilizing meta-learning to efficiently adapt models to new domains using previously acquired knowledge. This approach mitigates domain adaptation challenges and bolsters traditional UDA methods’ performance by harnessing the strengths of multiple UDA methods to enhance overall model generalization. The proposed approach shows potential in advancing domain adaptation research by providing a robust and efficient solution for real-world domain shifts. Experiments on three standard image datasets confirm the proposed approach’s superiority over state-of-the-art UDA methods, validating the effectiveness of meta-learning. Remarkably, the Office+Caltech 10, ImageCLEF-DA and combined digit datasets exhibit substantial accuracy gains of 30.9%, 6.8% and 10.9%, respectively, compared with the best-second baseline UDA approach.

One of the advanced machine learning branches is emotional learning based on brain emotional learning (BEL), which has been widely used for almost three decades. BEL mimics the emotional learning mechanism in the mammalian brain, which has the superior features of fast learning, quick reacting, and low computational complexity. The original BEL model inspired by the limbic system is composed of two neural network components, namely the amygdala (AMYG) and orbitofrontal cortex (ORBI), which interact with each other. Using a fuzzy extreme learning machine (FELM), a brain-inspired emotional learning model was developed in this study to predict noisy, chaotic time series. The exchange of information between AMYG and ORBI is facilitated by an online sequential type-one FELM with interactive recurrent memory (OIRMS-T1FELM). The CenteroMedial (CM) section of AMYG uses type-one FELM (T1FELM) and interval type-two FELM (IT2FELM). Hence, the proposed model is named BEL-OIRMS-T1/IT2 FELM. The impact of noise on time series targets was determined by altering the various standard deviation of Gaussian noise levels (Sig). The experimental results show that the interaction between AMYG and ORBI leads to the reduction of RMSE in their outputs. Also, when IT2FELM is used instead of T1FELM in the CM of AMYG, the RMSE and MAPE of the final prediction output are lower, and the R is higher. In this state, by increasing the noise level, the proposed method with MF=2 performs better than other states.

Early diagnosis of neurodegenerative diseases has always been a major challenge that physicians and medical practitioners face. Therefore, using any method or device that helps with prognostics is of great importance. In recent years, deep neural networks have become popular in medical fields, and the reason is that these networks can help diagnose diseases quickly and precisely. In this research, two novel models based on a CNN-LSTM network are introduced. The main goal is to classify three neurodegenerative diseases, including ALS, Parkinson’s disease, and Huntington’s disease, from one another and from healthy control patients using the gait signals, which are transformed into spectrogram images. In the first model, the spectrogram images derived from the gait signals are fed into a CNN-LSTM network directly. This model achieved 99.42% accuracy. In the second model, the same input data was used to be classified using a CNN-LSTM network, which uses wavelet transform as a feature extractor before the LSTM unit. During the experiments with the second model, the detail sub-bands were eliminated one by one, and the classification results were compared. Comparing these two models has shown that using the wavelet transform and, in particular, the approximation sub-bands can result in a lighter and faster prognosis with nearly 103 times fewer training parameters overall. The classification result using only approximation sub-bands was 95.37%, using three sub-bands was 94.04%, and including all sub-bands was 94.53%, which is remarkable.

In recent years, the development of smart cities has accelerated. There are several issues to handle in smart cities, one of the most important of which is efficient resource allocation. For the modeling of smart cities, multi-agent systems (MASs) can be used. In this paper, an efficient approach is proposed for resource allocation in smart cities based on the multi-agent credit assignment problem (MCA) and bankruptcy game. To this end, the resource allocation problem is mapped to MCA and the bankruptcy game. To solve this problem, first, a task start threshold (TST) constraint is introduced. The MCA turns into a bankruptcy problem upon introducing such a constraint. Therefore, based on the concept of bankruptcy, three methods of TS-Only, TS + MAS, and TS + ExAg are presented to solve the MCA. In addition, this work introduces a multi-score problem (MSP) in which a different reward is offered for solving each part of the problem, and we used it in our experiments to examine the proposed methods. The proposed approach is evaluated based on the learning rate, confidence, expertness, efficiency, certainty, and correctness parameters. The results reveal the better performance of the proposed approach compared to the existing methods in five parameters.