Based on the architecture of the single-task model, the multi-task model contains two main parts: shared layers for learning general hidden features from all data and task-specific layers for learning specific weights for different tasks [34]. Here, we have two different tasks: binary classification and regression. The input, feature extraction and concatenation parts are similar to those of the single-task model. The loss functions for different tasks are defined as: binary cross-entropy for classification (Loss1) and the MSE with L2 regularization for regression (Loss2).

where N, M and f(), and g() correspond to the samples and models for the classification and regression tasks; xi and yi correspond to the input and labels, respectively. ||w|| is an L2 regularization term; and λ ≥ 0 is used to adjust the relationship between the empirical risk and regularization term of the regression task.

Note: The content above has been extracted from a research article, so it may not display correctly.



Q&A
Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.



We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.