Abstract:
To improve performance, multi-task learning combines related tasks. If we can identify related tasks, we can often identify unrelated tasks. Most researchers focused on exploiting correlations between interrelated tasks and ignored unrelated tasks that could provide valuable prior knowledge for multi-task learning.
This paper introduces a hierarchical learning method for multi-task metrics using prior knowledge about related and unrelated tasks. First, a visual tree organizes many image categories from coarse to fine. Over the visual tree, a multi-task metric classifier is learned for each node by exploiting both related and unrelated tasks.
The learning tasks for training classifiers for sibling child nodes under the same parent node are interrelated, while the others are unrelated.
To control inter-level error propagation, the parent node’s node-specific metric is propagated to its siblings. Our hierarchical metric learning algorithm outperforms other state-of-the-art algorithms in experiments.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
