Just Pick a Sign: Optimizing Deep Multitask Models with Gradient Sign Dropout

By Zhao Chen, Jiquan Ngiam, Yanping Huang, Thang Luong, Henrik Kretzschmar, Yuning Chai, Dragomir Anguelov. Neural Information Processing Systems, 2020

300 citations recorded by the source corpus.

The vast majority of deep models use multiple gradient signals, typically corresponding to a sum of multiple loss terms, to update a shared set of trainable weights. However, these multiple updates can impede optimal training by pulling the model in conflicting directions. We present Gradient Sign Dropout (GradDrop), a probabilistic masking procedure which samples gradients at an activation layer based on their level of consistency. GradDrop is implemented as a simple deep layer that can be used in any deep net and synergizes with other gradient balancing approaches. We show that GradDrop outperforms the state-of-the-art multiloss methods within traditional multitask and transfer learning settings, and we discuss how GradDrop reveals links between optimal multiloss training and gradient stochasticity.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/de7c79528c1cdad734b363d92b35ee189f1e3ef5

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Contextualizing Multiple Tasks via Learning to Decompose (2021, background): “In standard multi-task learning, the correspondence between an example with one of the multiple tasks is explicit [24, 42, 8].”
Conflict-Averse Gradient Descent for Multi-task Learning (2021, background): “More recently, GradDrop [4] randomly drops out task gradients based on how much they conflict.”
Efficiently Identifying Task Groupings for Multi-Task Learning (2021, background): “Another set of methods seek to mitigate inter-task conflict by manipulating the direction of task gradients rather than simply their magnitude [48, 57, 53, 11, 50, 36].”
Efficient Diffusion Training via Min-SNR Weighting Strategy (2023, background): “To address this issue, some previous works [58, 54, 8] try to modulate the gradient to prevent conﬂicts.”