T
Trevor Darrell
Researcher at University of California, Berkeley
Publications - 734
Citations - 222973
Trevor Darrell is an academic researcher from University of California, Berkeley. The author has contributed to research in topics: Computer science & Object detection. The author has an hindex of 148, co-authored 678 publications receiving 181113 citations. Previous affiliations of Trevor Darrell include Massachusetts Institute of Technology & Boston University.
Papers
More filters
Proceedings ArticleDOI
Discriminative Gaussian process latent variable model for classification
Raquel Urtasun,Trevor Darrell +1 more
TL;DR: This work introduces a method for Gaussian Process Classification using latent variable models trained with discriminative priors over the latent space, which can learn a discrim inative latent space from a small training set.
Posted Content
Generalized Zero- and Few-Shot Learning via Aligned Variational Autoencoders
TL;DR: In this paper, a shared latent space of image features and class embeddings is learned by modality-specific aligned variational autoencoders, on which they train a softmax classifier.
Proceedings Article
Reinforcement Learning from Imperfect Demonstrations
TL;DR: This work proposes a unified reinforcement learning algorithm, Normalized Actor-Critic (NAC), that effectively normalizes the Q-function, reducing theQ-values of actions unseen in the demonstration data, making NAC robust to suboptimal demonstration data.
Proceedings ArticleDOI
Pyramid based depth from focus
Trevor Darrell,K. Wohn +1 more
TL;DR: A method is presented for depth recovery through the analysis of scene sharpness across changing focus position, modeling a defocused image as the application of a low pass-filter on a properly focused image of the same scene.
Posted Content
Algorithmic Framework for Model-based Deep Reinforcement Learning with Theoretical Guarantees
TL;DR: A novel algorithmic framework for designing and analyzing model-based RL algorithms with theoretical guarantees is introduced and a meta-algorithm with a theoretical guarantee of monotone improvement to a local maximum of the expected reward is designed.