Vinny Cahill

I undertake fundamental research into novel machine learning (ML) algorithms and their application to cyber-physical systems. My expertise sits at the intersection of deep reinforcement learning (RL), model-based RL, sequence modelling, and multi-agent systems. In the context of NeuroAI, I am particularly interested in how recent results in neuroscience can inform the development of new ML algorithms. Current projects are addressing continual RL to allow agents to learn policies for new tasks on an on-going basis without impacting performance on previously learned tasks, and causal modelling to capture the underlying dynamics of an environment in order to overcome the impact of unobserved phenomena that might negatively affect the ability to learn a new task. I have significant expertise in the application of emerging technologies to support sustainable mobility in general and Intelligent Transportation Systems in particular. I’m particularly interested in making travel more predictable by optimizing travel-time reliability for travellers and freight. If you have an idea for an ICT-based project addressing these areas, I’d be happy to discuss supervising it with you. Otherwise, I am proposing the following projects, a subset of which will be allocated this year, and all of which require that you are already comfortable developing complex software systems.

Continual Model-based Reinforcement Learning – TAKEN

Vinny Cahill (email | all projects)

Model-based reinforcement learning (MBRL) is emerging as a powerful paradigm for building intelligent agents capable of operating in the physical world. Most existing MBRL algorithms assume that the environment is stationary, i.e., that its behaviour doesn’t change so that the learnt world model and the strategies than depend on it remain optimal over time. But … Read more

Mitigating Bus Bunching using Reinforcement Learning

Vinny Cahill (email | all projects)

An ideal bus route would offer highly predictable inter-bus arrival times to travellers, in particular, overcoming the well-known ‘bus bunching’ phenomenon – “I waited ages then three came together!”. There are many sources of variability in journey times such as traffic conditions, the impact of traffic lights, and interference between buses themselves, e.g.,  queueing at … Read more

Slot-based Motorway Driving for Connected and Automated Vehicles

Vinny Cahill (email | all projects)

The Research Ireland funded ClearWay1 project is investigating new models of road management is which each vehicle is allocated a so-called ‘slot’ in which to travel for the duration of its journey. Adherence to travelling in its allocated slot ensures congestion-free travel from source to destination for each vehicle (in the absence of unexpected events). … Read more

ML-powered Parking Assistance System

Vinny Cahill (email | all projects)

Parking in busy locations eg at large-scale events (such as concerts and mass-participation sporting events) or in shopping centres at busy times (eg during ‘Christmas shopping’) often results in traffic congestion in surrounding areas (see, for example, [1], for an extreme example!). This may arise due to vehicles searching for, entering, and leaving parking spaces … Read more

Swarm Intelligence for Travel Time Reliability

Vinny Cahill (email | all projects)

Poor travel-time reliability, meaning that travel times for the same journey are highly variable and unpredictable, gives rise to similar negative impacts on the environment and the economy as does traffic congestion. Moreover, being able to offer a high degree of travel-time reliability will facilitate the uptake of sustainable road transportation including future public, shared, … Read more

Improving Road Transportation Resilience using LLMs and Agentic AI – TAKEN

Vinny Cahill (email | all projects)

Mobility is the lifeblood of human connectedness and economic progress, and roads are a critical component of the transportation system for both people and goods. Unfortunately, roads are potentially vulnerable to medium- to long-term disruptions (ie, lasting hours to days) due to extreme weather, infrastructure failures, accidents, maintenance, and unexpected demand all resulting in high … Read more