Iterative Temporal Learning and Prediction with the Sparse Online Echo State Gaussian Process

Paper accepted at IJCNN 2012; Never been to Brisbane!

Summary: In this work, we contribute the online echo state gaussian process (OESGP), a novel Bayesian-based online method that is capable of iteratively learning complex temporal dynamics and producing predictive distributions (instead of point predictions). Our method can be seen as a combination of the echo state network with a sparse approximation of Gaussian processes (GPs). Extensive experiments on the one-step prediction task on well-known benchmark problems show that OESGP produced statistically superior results to current online ESNs and state-of-the-art regression methods. In addition, we characterise the benefits (and drawbacks) associated with the considered online methods, specifically with regards to the trade-off between computational cost and accuracy. For a high-dimensional action recognition task, we demonstrate that OESGP produces high accuracies comparable to a recently published graphical model, while being fast enough for real-time interactive scenarios.

Download PDF

Involving Young Children in the Design of a Safe, Smart Paediatric Wheelchair

Independent mobility is crucial for a growing child and its loss can severely impact cognitive, emotional and social development. Unfortunately, powered wheelchair provision for young children has been difficult due to safety concerns. But powered mobility need not be unsafe. Risks can be reduced through the use of robotic technology (e.g., obstacle avoidance) and we present a prototype safe smart paediatric wheelchair: the Assistive Robot Transport for Youngsters (ARTY). A core aspect of our work is that we aim to bring ARTY to the field and we discuss the challenges faced when trying to involve children in the development/testing of medical technology. We discuss one preliminary experiment designed as a “Hide-and-Seek” game as a short case study.

Note: This paper was submitted to the ACM/IEEE HRI-2011 Pioneers Workshop, Lausanne 2011.


Multi-Reward Policies for Medical Applications: Anthrax Attacks and Smart Wheelchairs

Medical decisions are often difficult; they involve uncertain information, multiple-objectives and debatable outcomes. In this work, we discuss the application of the multi-reward partially-observable Markov decision process (MR-POMDP) and NSGA2-LS, a hybridised multi-objective evolutionary solver, to two problems in the medical domain: anthrax re- sponse and smart-wheelchair control. For the first problem, we use a discrete model and analyse the trade-offs between the best solutions (in the form of finite-state controllers) found by our evolutionary algorithm. For the second, we contribute an extension of our method to the continuous space and optimising recurrent neural networks (RNNs) for use on medical robots such as smart wheelchairs.

Download | ACM Digital Library Link

Weighted complex network analysis of travel routes on the Singapore public transportation system

The structure and properties of public transportation networks have great implications for urban planning, public policies and infectious disease control. We contribute a complex weighted network analysis of travel routes on the Singapore rail and bus transportation systems. We study the two networks using both topological and dynamical analyses. Our results provide additional evidence that a dynamical study adds to the information gained by traditional topological analysis, providing a richer view of complex weighted networks. For example, while initial topological measures showed that the rail network is almost fully connected, dynamical measures highlighted hub nodes that experience disproportionately large traffic. The dynamical assortativity of the bus networks also differed from its topological counterpart. In addition, inspection of the weighted eigenvector centralities highlighted a significant difference in traffic flows for both networks during weekdays and weekends, suggesting the importance of adding a temporal perspective missing from many previous studies.

Download | Physica A Link

Evolving policies for multi-reward partially observable markov decision processes (MR-POMDPs)

Plans and decisions in many real-world scenarios are made under uncertainty and to satisfy multiple, possibly conflicting, objectives. In this work, we contribute the multi-reward partially-observable Markov decision process (MR-POMDP) as a general modelling framework. To solve MR-POMDPs, we present two hybrid (memetic) multi-objective evolutionary algorithms that generate non-dominated sets of policies (in the form of stochastic finite state controllers). Performance comparisons between the methods on multi-objective problems in robotics (with 2, 3 and 5 objectives), web-advertising (with 3, 4 and 5 objectives) and infectious disease control (with 3 objectives), revealed that memetic variants outperformed their original counterparts. We anticipate that the MR-POMDP along with multi-objective evolutionary solvers will prove useful in a variety of theoretical and real-world applications.

Download | ACM Digital Library Link

Steak Night with Friends

Had an awesome steak night with Kyu Hwa, Sing and Miguel. Fresh north highland fillet and sirloin with Miguel’s famous secret sauce.

Miguel’s no-longer-secret sauce recipe:

Dice 1 onion, 2 tomatoes and a hand-full of coriander.

Mix together with (the best) olive oil and freshly squeezed lemon juice (about 1 lemon).

Add salt to taste. Enjoy!