Seizure prediction algorithms will be developed using the one-of-a-kind ultra-long-term human intracranial EEG dataset obtained from the Neurovista Corporation clinical trial of their Seizure Advisory System, or data from other implantable or wearable devices. This involves consideration of both feature-based machine learning or data science approaches and neural mass parameter estimation approaches to classify the EEG and predict seizures. Recent approaches focus on critical slowing as a marker for seizure susceptability and the influence of brain rhythms.
Research projects in Information Technology
Displaying 11 - 20 of 99 projects.
Safe Neuro-symbolic Automated Decision Making with Mathematical Optimisation
Planning is the reasoning side of acting in Artificial Intelligence. Planning automates the selection and the organisation of actions to reach desired states of the world as best as possible. For many real-world planning problems however, it is difficult to obtain a transition model that governs state evolution with complex dynamics.
Safe Continuous-time Automated Decision Making with Mathematical Optimisation
SCIPPlan is a mathematical optimisation based automated planner for domains with i) mixed (i.e., real and/or discrete valued) state and action spaces, ii) nonlinear state transitions that are functions of time, and iii) general reward functions. SCIPPlan iteratively i) finds violated constraints (i.e., zero-crossings) by simulating the state transitions, and ii) adds the violated constraints back to its underlying optimization model, until a valid plan is found.
End-to-End Prediction and Optimisation for Neuro-Symbolic Artificial Intelligence
Optimisation methods, such as mixed integer linear programming, have been very successful at decision-making for more than 50 years. Optimisation algorithms support basically every industry behind the scenes and the simplex algorithm is one of the top 10 most influential algorithms. Major success stories include rostering nurses in hospitals, managing chains of organ transplants, planning production levels for manufacturing, routing delivery trucks for transport, scheduling power stations and electricity grids, to name just a few.
Branch-and-Cut-and-Price Algorithms for Computing Cost-Effective and Time-Efficient Delivery Routes for Trucks and Drones
Transport and logistics businesses today use a large fleet of trucks and vans to deliver packages widely across a city. Deciding which package should be loaded on to which vehicle and deciding which package should be prioritised are surprisingly difficult computational tasks. State-of-the-art high-performance algorithms are used to calculate routes for the vehicles in order to minimise costs and maximise efficiency.
Measures of Simplicity in Optimisation-based Machine Learning
The notion of simplicity can occur in many contexts. Sometimes simplicity can be explicitly sought so that the number of variables to be considered is manageable, and sometimes simplicity can arise as a consequence of other desiderata. In the context of machine learning, many approaches propose to learn simple models since judicious simplicity can lead to good predictions [3,1]. These approaches can be used to learn simple models (e.g., decision trees [6], decision graphs [4, 5], etc.) or improve more complex models (e.g., neural networks [2]).
Machine learning based kinetic modeling on the thermal decomposition of plastic waste via pyrolysis
This project aims to develop a machine learning based kinetic model for an accurate prediction on the product yield and quality from the pyrolysis of plastic waste. The primary outcome of the Project is the development of a robust model that is effective in simulating the entire pyrolysis process at a relatively low computing cost, whereas its results will be sufficiently accurate to predict the composition and yields of the products.
[Malaysia] - Integration of heterogeneous biomedical data for robust and interpretable prediction
Many machine learning (ML) approaches have been applied to biomedical data but without substantial applications due to the poor interpretability of models. Although ML approaches have shown promising results in building prediction models, they are typically data-centric, lack context, and work best for specific feature types. Interpretability is the ability of an ML model to identify the causal relationships among variables.
Creating a turnkey solution to classify, predict and simulate behaviour from videos of rodents
Rodent behavioural testing is the study of the neural mechanisms underlying emotions [1]. It is used in the study of almost all mental conditions, including PTSD [2], OCD [3] and autism [4]. For example, to measure anxiety, researchers may place a rodent in a large tub, record a top-down video and measure the time spent near the safety of walls [2]. These videos also contain rich information about behavioural patterns, but scoring this manually is time consuming.