AdaMT-Net: An Adaptive Weight Learning Based Multi-Task Learning Model For Scene Understanding.
Ankit JhaAwanish KumarBiplab BanerjeeSubhasis ChaudhuriPublished in: CVPR Workshops (2020)
Keyphrases
- multi task learning
- learning models
- scene understanding
- learning problems
- prior knowledge
- multi task
- learning tasks
- multiple tasks
- similarity measure
- reinforcement learning
- active learning
- high order
- learning algorithm
- multitask learning
- loss function
- probabilistic model
- unsupervised learning
- vision system
- dimensionality reduction
- object detection
- high resolution
- learning process
- computer vision
- real time