Deep transfer learning: a novel glucose prediction framework for new subjects with type 2 diabetes

Hybrid CNN-LSTM for Predicting Diabetes: A Review

BACKGROUND

Diabetes is a common and deadly chronic disease caused by high blood glucose levels that can cause heart problems, neurological damage, and other illnesses. Through the early detection of diabetes, patients can live healthier lives. Many machine learning and deep learning techniques have been applied for noninvasive diabetes prediction. The results of some studies have shown that the CNN-LSTM method, a combination of CNN and LSTM, has good performance for predicting diabetes compared to other deep learning methods.

METHOD

This paper reviews CNN-LSTM-based studies for diabetes prediction. In the CNNLSTM model, the CNN includes convolution and max pooling layers and is applied for feature extraction. The output of the max-pooling layer was fed into the LSTM layer for classification.

DISCUSSION

The CNN-LSTM model performed well in extracting hidden features and correlations between physiological variables. Thus, it can be used to predict diabetes. The CNNLSTM model, like other deep neural network architectures, faces challenges such as training on large datasets and biological factors. Using large datasets can further improve the accuracy of detection.

CONCLUSION

The CNN-LSTM model is a promising method for diabetes prediction, and compared with other deep-learning models, it is a reliable method.

Adaptive Personalized Prior-Knowledge-Informed Model Predictive Control for Type 1 Diabetes

This work considers the problem of adaptive prior-informed model predictive control (MPC) formulations that explicitly incorporate prior knowledge in the model development and is robust to missing data in the output measurements. The proposed prediction model is based on a latent variables model to extract glycemic dynamics from highly-correlated data and incorporates prior knowledge of exponential stability to improve the prediction ability. Missing data structures are formulated to enable model predictions when output measurements are missing for short periods of time. Based on the latent variables model, the MPC strategy and adaptive rules are developed to automatically tune the aggressiveness of the MPC. The adaptive prior-knowledge-informed MPC is evaluated with computer simulations for the control of blood glucose concentrations in people with Type 1 diabetes (T1D) using simulated virtual patients. Due to the variability among people with T1D, the hyperparameters of the prior-knowledge-informed model are personalized to individual subjects. The percentage of time spent in the target range is 76.48% when there are no missing data and 76.52% when there are missing data episodes lasting up to 30 mins (6 samples). Incorporating the adaptive rules further improves the percentage of time in target range to 84.58% and 84.88% for cases with no missing data and missing data, respectively. The proposed adaptive prior-informed MPC formulation provides robust, effective, and safe regulation of glucose concentration in T1D despite disturbances and missing measurements.

A cross-domain fruit classification method based on lightweight attention networks and unsupervised domain adaptation

Image-based fruit classification offers many useful applications in industrial production and daily life, such as self-checkout in the supermarket, automatic fruit sorting and dietary guidance. However, fruit classification task will have different data distributions due to different application scenarios. One feasible solution to solve this problem is to use domain adaptation that adapts knowledge from the original training data (source domain) to the new testing data (target domain). In this paper, we propose a novel deep learning-based unsupervised domain adaptation method for cross-domain fruit classification. A hybrid attention module is proposed and added to MobileNet V3 to construct the HAM-MobileNet that can suppress the impact of complex backgrounds and extract more discriminative features. A hybrid loss function combining subdomain alignment and implicit distribution metrics is used to reduce domain discrepancy during model training and improve model classification performance. Two fruit classification datasets covering several domains are established to simulate common industrial and daily life application scenarios. We validate the proposed method on our constructed grape classification dataset and general fruit classification dataset. The experimental results show that the proposed method achieves an average accuracy of 95.0% and 93.2% on the two datasets, respectively. The classification model after domain adaptation can well overcome the domain discrepancy brought by different fruit classification scenarios. Meanwhile, the proposed datasets and method can serve as a benchmark for future cross-domain fruit classification research.

Generative knowledge-based transfer learning for few-shot health condition estimation

In the field of high-end manufacturing, it is valuable to study few-shot health condition estimation. Although transfer learning and other methods have effectively improved the ability of few-shot learning, they still cannot solve the lack of prior knowledge. In this paper, by combining data enhancement, knowledge reasoning, and transfer learning, a generative knowledge-based transfer learning model is proposed to achieve few-shot health condition estimation. First, with the effectiveness of data enhancement on machine learning, a novel batch monotonic generative adversarial network (BM-GAN) is designed for few-shot health condition data generation, which can solve the problem of insufficient data and generate simulated training data. Second, a generative knowledge-based transfer learning model is proposed with the performance advantages of the belief rule base (BRB) method on few-shot learning, which combines expert knowledge and simulated training data to obtain a generalized BRB model and then fine-tunes the generalized model with real data to obtain a dedicated BRB model. Third, through uniform sampling of NASA lithium battery data and simulating few-shot conditions, the generative transfer-belief rule base (GT-BRB) method proposed in this paper is verified to be feasible for few-shot health condition estimation and improves the estimation accuracy of the BRB method by approximately 17.3%.

Cyberbullying detection using deep transfer learning

Social networking platforms like Facebook, Twitter, and others have numerous advantages, but they have many dark sides also. One of the issues on these social platforms is cyberbullying. The impact of cyberbullying is immeasurable on the life of victims as it’s very subjective to how the person would tackle this. The message may be a bully for victims, but it may be normal for others. The ambiguities in cyberbullying messages create a big challenge to find the bully content. Some research has been reported to address this issue with textual posts. However, image-based cyberbullying detection is received less attention. This research aims to develop a model that helps to prevent image-based cyberbullying issues on social platforms. The deep learning-based convolutional neural network is initially used for model development. Later, transfer learning models are utilized in this research. The experimental outcomes of various settings of the hyper-parameters confirmed that the transfer learning-based model is the better choice for this problem. The proposed model achieved a satisfactory accuracy of 89% for the best case, indicating that the system detects most cyberbullying posts.