|
|
|
 |
Search published articles |
 |
|
Showing 5 results for Motamed
Sara Motamed, Mahboubeh Yaghoubi,
Volume 16, Issue 1 (3-2025)
Abstract
Intelligence has long been an interesting and important topic in psychology and cognitive science. IQ is considered a basic measure of a person's cognitive abilities, which includes various aspects of reasoning, problem solving, memory, and overall intellectual ability. Considering the importance of IQ in cognitive and psychological evaluations, the main goal of this article was to provide a new and effective approach to improve the accuracy of estimating this measure through complex brain data processing. In this paper, we have analyzed and developed a hybrid model of GWO algorithm and CNN (GCNN) in order to estimate IQ using brain MRI images. The results of the experiments showed that the accuracy of the proposed model was significantly better than the traditional techniques, and this indicates the high capabilities of the model in interpreting complex medical data. By examining the results, we find that the accuracy of the proposed model with an estimation rate of 93.10% is better than other competing methods.
Mrs Sareh Bagheri Matak, Dr Elham Askari, Dr Sara Motamed,
Volume 16, Issue 2 (8-2025)
Abstract
Leukemia is one of the most common and dangerous types of cancer in the world. In many cases, the disease is curable if detected in its early stages. One of the effective tools for early detection is the analysis of microarray data, which measures the expression of thousands of genes simultaneously. However, the large volume of features and the presence of noise make the analysis process complex and time-consuming. Therefore, the selection of effective genes plays a key role in increasing the accuracy and reducing the computational cost of learning models. In this paper, a two-step hybrid approach is presented for feature selection and classification of leukemia types. In the first step, the features are filtered using the mutual information criterion and the genes with the highest correlation with the disease label are selected. In the second step, the XGBoost model is used to rank and stably select the features to identify the genes that are most important in different iterations. In the final stage, classification will be performed using the temporal fusion transformer method, which allows for fast and efficient learning of complex patterns among selected genes. Experimental results on real microarray datasets show that the proposed method outperforms the baseline methods with an accuracy of 99.2% and has been able to identify key genes effective in differentiating leukemia types by effectively reducing the data dimensions.
Sara Motamed,
Volume 16, Issue 2 (8-2025)
Abstract
This paper presents a constrained multi-objective deep reinforcement learning framework for urban traffic signal control. The problem is modeled as a constrained Markov decision process in which an agent simultaneously optimizes efficiency objectives while respecting explicit safety and fairness constraints. A dueling double deep Q-network (D3QN) is combined with a Lagrangian cost estimator to approximate both the reward value function and cumulative constraint costs. The state representation includes queue lengths, phase indicators and elapsed green times, and the action space consists of a small set of interpretable decisions such as extending the current green or switching to the next phase. The proposed controller is trained and evaluated in a SUMO-based microscopic simulation of a four-leg urban intersection under various traffic demand patterns. Its performance is compared with fixed-time, vehicle-actuated and unconstrained DQN controllers. Simulation results show that the proposed method can substantially reduce average delay and maximum queue length while keeping queue spillback and delay imbalance within predefined limits. These findings indicate that constrained multi-objective deep reinforcement learning offers a promising and practically deployable framework for safe and fair traffic signal control in congested urban networks, and can be extended to more complex corridors and network-wide settings in future work.
Saba Gholami, Sara Motamed, Elham Askari,
Volume 17, Issue 1 (5-2026)
Abstract
Autism spectrum disorder (ASD) is consistently associated with abnormal functional connectivity; resting-state fMRI data were obtained from the ABIDE dataset. Dynamic functional connectivity (DFC) was obtained in an autism-specific subnetwork consisting of 17 regions identified from previous static connectivity analyses. Time-varying connectivity matrices were estimated using a sliding window approach, and recurrent connectivity states were identified using a hidden Markov model. Dynamic measures included state occupancy rate, mean dwell time, and edge-level connectivity variability. Compared with controls, individuals with ASD showed a significant decrease in the occupancy of highly integrated connectivity states (ASD: 28.6 ± 7.4% vs. control: 36.9 ± 8.1%, p < 0.001) and longer dwell times in poorly integrated connectivity states (ASD: 42.3 ± 10.2 vs. control: 31.7 ± 9.5 s, p = 0.002). In contrast, edge-level connectivity variability was significantly increased in ASD, particularly in default mode-limbic connections. Importantly, increased connectivity variability in the default mode network significantly predicted ADOS total scores (β = 0.41), (p = 0.001). These findings suggest a dissociation between reduced network state flexibility and increased moment-to-moment connectivity variability in autism spectrum disorder (ASD).
Zinat Mehdidoust Jalali, Azamossadat Nourbakhsh, Sara Motamed,
Volume 17, Issue 2 (9-2026)
Abstract
Face recognition in real-world environments remains challenging due to domain shifts, limited labeled data, decision uncertainty, and constrained computational resources. This paper presents a novel neuro-symbolic framework that combines a SOAR-based cognitive meta-controller, cost-aware reinforcement learning, and domain adaptation to jointly improve recognition accuracy, computational efficiency, and inference latency. A face embedding network is first trained on the source domain and then adapted to the target domain using parameter-efficient fine-tuning (PEFT) and Domain-Adversarial Neural Networks (DANN). To enhance robustness under limited supervision and distribution shifts, the framework further incorporates few-shot learning and test-time adaptation (TTA). The proposed meta-controller is formulated as a Markov Decision Process (MDP), enabling dynamic allocation of computational resources based on image quality, uncertainty estimation, and recognition status. Symbolic knowledge encoded in the SOAR architecture guides the reinforcement learning policy through a neuro-symbolic gating mechanism, improving both interpretability and decision consistency. Experiments on the IJB-C and MegaFace benchmarks demonstrate significant improvements in face recognition and open-set recognition performance, while PEFT reduces trainable parameters by more than 95% without increasing inference latency. Overall, the proposed framework offers an accurate, adaptive, interpretable, and computationally efficient solution for real-world face recognition applications.
|
|
