Controlling Directions Orthogonal to a Classifier
We propose to identify directions invariant to a given classifier so that these directions can be controlled in tasks such as style transfer. While orthogonal decomposition is directly identifiable when the given classifier is linear, we formally define a notion of orthogonality in the non-linear case. We also provide a surprisingly simple method for constructing the orthogonal classifier (a classifier utilizing directions other than those of the given classifier). Empirically, we present three use cases where controlling orthogonal variation is important: style transfer, domain adaptation, and fairness. The orthogonal classifier enables desired style transfer when domains vary in multiple aspects, improves domain adaptation with label shifts and mitigates the unfairness as a predictor. The code is available at http://github.com/Newbeeer/orthogonal_classifier
Code (1)
Tasks
Domain AdaptationFairnessStyle TransferSimilar Papers 제목 키워드 기반
The Multiverse Loss for Robust Transfer Learning
Deep learning techniques are renowned for supporting effective transfer learning. However, as we demonstrate, the transferred representations support only a few modes of separation and much of its dimensionality is unuti…
Transfer LearningGASS: Geometry-Aware Spherical Sampling for Disentangled Diversity Enhancement in Text-to-Image Generation
Despite high semantic alignment, modern text-to-image (T2I) generative models still struggle to synthesize diverse images from a given prompt. In this work, we enhance the T2I diversity through a geometric lens. Unlike m…
Text-to-Image GenerationA unified strategy for implementing curiosity and empowerment driven reinforcement learning
Although there are many approaches to implement intrinsically motivated artificial agents, the combined usage of multiple intrinsic drives remains still a relatively unexplored research area. Specifically, we hypothesize…
reinforcement-learningReinforcement LearningReinforcement Learning (RL)The Cylindrical Representation Hypothesis for Language Model Steering
Steering is a widely used technique for controlling large language models, yet its effects are often unstable and hard to predict. Existing theoretical accounts are largely based on the Linear Representation Hypothesis (…
What Role Do Intelligent Reflecting Surfaces Play in Multi-Antenna Non-Orthogonal Multiple Access?
Massive multiple-input multiple-output (MIMO) and non-orthogonal multiple access (NOMA) are two key techniques for enabling massive connectivity in future wireless networks. A massive MIMO-NOMA system can deliver remarka…
Fairness