Disentangling neural network representations for improved generalization

Despite the increasingly broad perceptual capabilities of neural networks, applying them to new tasks requires significant engineering effort in data collection and model design. Generally, inductive biases can make this process easier by leveraging knowledge about the world to guide neural network design. One such inductive bias is disentanglment, which can help preven neural networks from learning representations that capture spurious patterns that do not generalize past the training data, and instead encourage them to capture factors of variation that explain the data generally. In this thesis we identify three kinds of disentanglement, implement a strategy for enforcing disentanglement in each case, and show that more general representations result. These perspectives treat disentanglement as statistical independence of features in image classification, language compositionality in goal driven dialog, and latent intention priors in visual dialog. By increasing the generality of neural networks through disentanglement we hope to reduce the effort required to apply neural networks to new tasks and highlight the role of inductive biases like disentanglement in neural network design.

Level thesis:degree_level

Doctoral

Department dc:contributor.department

Interactive Computing

Grantor dc:publisher

Georgia Institute of Technology

Year dc:date.issued

2020

Author dc:creator

• Cogswell, Michael Andrew

Advisor dc:contributor.advisor

• Batra, Dhruv

Committee members dc:contributor.committeemember

• Parikh, Devi
• Hays, James
• Goel, Ashok
• Lee, Stefan

• Deep learning
• Disentanglement
• Compositionality
• Representation learning
• Visual dialog
• Language emergence

Language dc:language.iso

en_US

Handle dc:identifier.uri

http://hdl.handle.net/1853/62813

OAI identifier oai:identifier

oai:repository.gatech.edu:1853/62813