CerviLearnNet: Advancing cervical cancer diagnosis with reinforcement learning-enhanced convolutional networks Shakhnoza Muksimova, Sabina Umirzakova $ ^{*} $ , Seokwhan Kang, Young Im Cho $ ^{**} $ Department of Computer Engineering, Gachon University, Sujeong-gu, Seongnam-si 461-701, Gyeonggi-do, South Korea This paper proposes a novel framework called RL-CancerNet, which integrates reinforcement learning (RL) with convolutional neural networks (CNNs) to improve the accuracy of cervical cancer diagnosis. The framework uses a modified EfficientNetV2 model with novel r-by-object supporter blocks to enhance the model's ability to learn critical feature information. The model was trained and tested on two publicly available datasets, SipaKMeD and Herlev, to assess its performance and compare it with existing methods. The results show that RL-CancerNet achieves a total recognition accuracy of 99.32%, which is a significant improvement over the baseline EfficientNetV2 model with 98.53% accuracy. The model also demonstrates high precision, recall, and F1 score, indicating its effectiveness in classifying cervical cancer cells. The study highlights the potential of RL-enhanced CNNs in improving the accuracy and reliability of cervical cancer diagnosis. The proposed method addresses the challenge of class imbalance in the datasets, which is a critical factor affecting the model's ability to accurately distinguish between normal and abnormal cervical cells. The results show that the model's specificity has been improved to 99.35%, indicating its reliability in distinguishing between normal and abnormal samples. The study concludes that RL-CancerNet is a promising approach for improving the accuracy and reliability of cervical cancer diagnosis.CerviLearnNet: Advancing cervical cancer diagnosis with reinforcement learning-enhanced convolutional networks Shakhnoza Muksimova, Sabina Umirzakova $ ^{*} $ , Seokwhan Kang, Young Im Cho $ ^{**} $ Department of Computer Engineering, Gachon University, Sujeong-gu, Seongnam-si 461-701, Gyeonggi-do, South Korea This paper proposes a novel framework called RL-CancerNet, which integrates reinforcement learning (RL) with convolutional neural networks (CNNs) to improve the accuracy of cervical cancer diagnosis. The framework uses a modified EfficientNetV2 model with novel r-by-object supporter blocks to enhance the model's ability to learn critical feature information. The model was trained and tested on two publicly available datasets, SipaKMeD and Herlev, to assess its performance and compare it with existing methods. The results show that RL-CancerNet achieves a total recognition accuracy of 99.32%, which is a significant improvement over the baseline EfficientNetV2 model with 98.53% accuracy. The model also demonstrates high precision, recall, and F1 score, indicating its effectiveness in classifying cervical cancer cells. The study highlights the potential of RL-enhanced CNNs in improving the accuracy and reliability of cervical cancer diagnosis. The proposed method addresses the challenge of class imbalance in the datasets, which is a critical factor affecting the model's ability to accurately distinguish between normal and abnormal cervical cells. The results show that the model's specificity has been improved to 99.35%, indicating its reliability in distinguishing between normal and abnormal samples. The study concludes that RL-CancerNet is a promising approach for improving the accuracy and reliability of cervical cancer diagnosis.