Simon O'Callaghan

Simon O'Callaghan
The Commonwealth Scientific and Industrial Research Organisation | CSIRO · Data61

About

27
Publications
10,209
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
565
Citations
Introduction

Publications

Publications (27)
Preprint
Most work in algorithmic fairness to date has focused on discrete outcomes, such as deciding whether to grant someone a loan or not. In these classification settings, group fairness criteria such as independence, separation and sufficiency can be measured directly by comparing rates of outcomes between subpopulations. Many important problems howeve...
Preprint
Algorithmic fairness involves expressing notions such as equity, or reasonable treatment, as quantifiable measures that a machine learning algorithm can optimise. Most work in the literature to date has focused on classification problems where the prediction is categorical, such as accepting or rejecting a loan application. This is in part because...
Preprint
Spatial data about individuals or businesses is often aggregated over polygonal regions to preserve privacy, provide useful insight and support decision making. Given a particular aggregation of data (say into local government areas), the re-aggregation problem is to estimate how that same data would aggregate over a different set of polygonal regi...
Article
Full-text available
Tracing sedimentation through time on existing and vanished seafloor is imperative for constraining long-term eustasy and for calculating volumes of subducted deep-sea sediments that contribute to global geochemical cycles. We present regression algorithms that incorporate the age of the ocean crust and the mean distance to the nearest passive marg...
Conference Paper
Understanding the dynamics of urban environments is crucial for path planning and safe navigation. However, the dynamics might be extremely complex making learning the environment an unfathomable task. Within the methods available for learning dynamic environments, dynamic Gaussian process occupancy maps (DGPOM) are very attractive because they can...
Data
The feature evaluator in this iPython notebook uses probabilistic Gaussian process classification to estimate the significance of individual features in predicting sea floor sediment type. Five real features and one random feature are used in the evaluation. Feature ID Feature Name 1 bathymetry 2 silicate 3 productivity 4 salinity 5 temperature...
Conference Paper
Full-text available
We consider the problem of building continuous occupancy representations in dynamic environments for robotics applications. The problem has hardly been discussed previously due to the complexity of patterns in urban environments, which have both spatial and temporal dependencies. We address the problem as learning a kernel classifier on an efficien...
Article
Full-text available
Geophysical joint inversions seek to exploit the statistical fact that a model that simultaneously satisfies two or more independent data sets is more likely to represent geological ‘reality’ than a model that only satisfies a single data set. Interpreting geophysical data directly rapidly exceeds the capacity of a human as more data are added, so...
Chapter
We present a continuous Bayesian occupancy representation for dynamic environments. Themethod builds on Gaussian processes classifiers and addresses the main limitations of occupancy grids such as the need to discretise the space, strong assumptions of independence between cells, and difficulty to represent occupancy in dynamic environments. We dev...
Data
iPython notebook for Support Vector Machine Classifier A Support Vector Machine (SVM) (Cortes and Vapnik, 1995) is a classifier that attempts to separate classes of data mapped to a space where it can be separated by a hyperplane. The location of the plane in the input space is selected based on a set of training examples, x_i∈ R^d (i=1,2,……,N), l...
Article
Full-text available
Deep-sea sediments represent the largest geological deposit on Earth and provide a record of our planet’s response to conditions at the sea surface from where the bulk of material originates. We use a machine learning method to analyze how the distribution of 14,400 deep-sea sediment sample lithologies is connected to bathymetry and surface oceanog...
Article
Understanding and predicting how influenza propagates is vital to reduce its impact. In this paper we develop a nonparametric model based on Gaussian process (GP) regression to capture the complex spatial and temporal dependencies present in the data. A stochastic variational inference approach was adopted to address scalability. Rather than modeli...
Conference Paper
Full-text available
Understanding and predicting how influenza propagates is vital to reduce its impact. In this paper we develop a nonparametric model based on Gaussian process (GP) regression to capture the complex spatial and temporal dependencies present in the data. A stochastic variational inference approach was adopted to address scalability. Rather than modeli...
Data
This spinning virtual seafloor geology globe is composed of a set of screen captures of an interactive digital globe portraying the distribution of different seafloor sediments available at http://portal.gplates.org/#SEAFLOOR. The globe animation can also be viewed here: https://www.youtube.com/watch?v=GltTEzqiAPw A N-S spinning version of the glob...
Article
Full-text available
Knowing the patterns of distribution of sediments in the global ocean is critical for understanding biogeochemical cycles and how deep-sea deposits respond to environmental change at the sea surface. We present the first digital map of seafloor lithologies based on descriptions of nearly 14,500 samples from original cruise reports, interpolated usi...
Conference Paper
Full-text available
The quantification of uncertainty in exploration for geothermal targets is central to determining the best locations for drilling. We develop algorithms to perform data fusion and joint inversions from a broad range of independent geophysics surveys to infer rock properties such as density, magnetic susceptibility, seismic velocity, resistivity, th...
Conference Paper
We present a novel estimation algorithm for filtering and regression with a number of advantages over existing methods. The algorithm has wide application in robotics as no assumptions are made about the underlying distributions, it can represent non-Gaussian multi-modal posteriors, and learn arbitrary non-linear models from noisy data. Our method...
Article
We propose a machine learning approach to geophysical inversion problems for the exploration of earth resources. Our approach is based on nonparametric Bayesian methods, specifically, Gaussian processes, and provides a full distribution over the predicted geophysical properties whilst enabling the incorporation of data from different modalities. We...
Article
We introduce a new statistical modelling technique for building occupancy maps. The problem of mapping is addressed as a classification task where the robot’s environment is classified into regions of occupancy and free space. This is obtained by employing a modified Gaussian process as a non-parametric Bayesian learning technique to exploit the fa...
Conference Paper
Observing human motion patterns is informative for social robots that share the environment with people. This paper presents a methodology to allow a robot to navigate in a complex environment by observing pedestrian positional traces. A continuous probabilistic function is determined using Gaussian process learning and used to infer the direction...
Conference Paper
We address the problem of building a continuous occupancy representation of the environment with ranging sensors. Observations from such sensors provide two types of information: a line segment or a beam indicating no returns along them (free-space); a point or return at the end of the segment representing an occupied surface. To model these two ty...
Conference Paper
This paper describes a method of incorporating sensor and localisation uncertainty into contextual occupancy maps to provide for robust mapping. This paper builds on a recently proposed application of the Gaussian process (GP) to occupancy mapping. An extension of GPs is employed which incorporates uncertain inputs into the covariance function. In...
Conference Paper
In this paper we introduce a new statistical modeling technique for building occupancy maps. The problem of mapping is addressed as a classification task where the robot's environment is classified into regions of occupancy and unoccupancy. Our model provides both a continuous representation of the robot's surroundings and an associated predictive...
Article
Modelling and predicting the movement of people through different environments offer important benefits to a wide range of applications from estimating intent to security. The study of human motion patterns is also crucial for the develop-ment of social robots that share their surroundings with people. Here we present a methodology to learn sociall...

Network

Cited By