S
Svetha Venkatesh
Researcher at Deakin University
Publications - 864
Citations - 20118
Svetha Venkatesh is an academic researcher from Deakin University. The author has contributed to research in topics: Bayesian optimization & Computer science. The author has an hindex of 60, co-authored 828 publications receiving 16441 citations. Previous affiliations of Svetha Venkatesh include Australian National University & National University of Singapore.
Papers
More filters
Posted Content
Semantic Host-free Trojan Attack
TL;DR: Zhang et al. as discussed by the authors proposed a host-free Trojan attack with triggers that are fixed in the semantic space but not necessarily in the pixel space, which makes their attack more practical to be applied in the real-world and harder to defend against.
Journal Article
ALT-MAS: A Data-Efficient Framework for Active Testing of Machine Learning Algorithms
TL;DR: In this article, the authors propose a framework to estimate the metrics of interest for a model-under-test using Bayesian neural network and devise an entropy-based sampling strategy to sample the data point such that the proposed framework can give accurate estimations for the metrics.
Posted Content
Stable Bayesian Optimisation via Direct Stability Quantification.
TL;DR: This approach uses multiple gradient Gaussian Process models to estimate the probability that worst-case output variation for specified input perturbation exceeded the desired maxima, and these probabilities are used to guide the optimisation process toward solutions satisfying the authors' stability criteria.
Posted Content
Boosted Markov Networks for Activity Recognition
TL;DR: In this article, a boosted Markov network with hidden variables is proposed to combine the learning capacity of boosting and the rich modeling semantics of Markov networks for video-based activity recognition.
Posted Content
Cumulative Restricted Boltzmann Machines for Ordinal Matrix Data Analysis
TL;DR: It is shown that the model is able to capture latent opinion profile of citizens around the world, and is competitive against state-of-art collaborative filtering techniques on large-scale public datasets.