David Murphy

Vu Dinh Xuan, Hao Vo, David Murphy et al.

The scarcity of data depicting dangerous situations presents a major obstacle to training AI systems for safety-critical applications, such as construction safety, where ethical and logistical barriers hinder real-world data collection. This creates an urgent need for an end-to-end framework to generate synthetic data that can bridge this gap. While existing methods can produce synthetic scenes, they often lack the semantic depth required for scene simulations, limiting their effectiveness. To address this, we propose a novel multi-agent framework that employs an iterative, in-the-loop collaboration between two agents: an Evaluator Agent, acting as an LLM-based judge to enforce semantic consistency and safety-specific constraints, and an Editor Agent, which generates and refines scenes based on this guidance. Powered by LLM's capabilities to reasoning and common-sense knowledge, this collaborative design produces synthetic images tailored to safety-critical scenarios. Our experiments suggest this design can generate useful scenes based on realistic specifications that address the shortcomings of prior approaches, balancing safety requirements with visual semantics. This iterative process holds promise for delivering robust, aesthetically sound simulations, offering a potential solution to the data scarcity challenge in multimedia safety applications.

Gareth W. Young, David Murphy, Jeffrey Weeter

We present the findings of a pilot-study that analysed the role of haptic feedback in a musical context. To examine the role of haptics in Digital Musical Instrument (DMI) design an experiment was formulated to measure the users' perception of device usability across four separate feedback stages: fully haptic (force and tactile combined), constant force only, vibrotactile only, and no feedback. The study was piloted over extended periods with the intention of exploring the application and integration of DMIs in real-world musical contexts. Applying a music orientated analysis of this type enabled the investigative process to not only take place over a comprehensive period, but allowed for the exploration of DMI integration in everyday compositional practices. As with any investigation that involves creativity, it was important that the participants did not feel rushed or restricted. That is, they were given sufficient time to explore and assess the different feedback types without constraint. This provided an accurate and representational set of qualitative data for validating the participants' experience with the different feedback types they were presented with.

David Murphy, Thomas S. Paula, Wagston Staehler et al.

In the future we can expect that artificial intelligent agents, once deployed, will be required to learn continually from their experience during their operational lifetime. Such agents will also need to communicate with humans and other agents regarding the content of their experience, in the context of passing along their learnings, for the purpose of explaining their actions in specific circumstances or simply to relate more naturally to humans concerning experiences the agent acquires that are not necessarily related to their assigned tasks. We argue that to support these goals, an agent would benefit from an episodic memory; that is, a memory that encodes the agent's experience in such a way that the agent can relive the experience, communicate about it and use its past experience, inclusive of the agents own past actions, to learn more effective models and policies. In this short paper, we propose one potential approach to provide an AI agent with such capabilities. We draw upon the ever-growing body of work examining the function and operation of the Medial Temporal Lobe (MTL) in mammals to guide us in adding an episodic memory capability to an AI agent composed of artificial neural networks (ANNs). Based on that, we highlight important aspects to be considered in the memory organization and we propose an architecture combining ANNs and standard Computer Science techniques for supporting storage and retrieval of episodic memories. Despite being initial work, we hope this short paper can spark discussions around the creation of intelligent agents with memory or, at least, provide a different point of view on the subject.

David Murphy, Conor Higgins

This paper presents an experiment to assess the feasibility of using secondary input data as a method of determining user engagement in immersive virtual reality (VR). The work investigates whether secondary data (biosignals) acquired from users are useful as a method of detecting levels of concentration, stress, relaxation etc. in immersive environments, and if they could be used to create an affective feedback loop in immersive VR environments, including educational contexts. A VR Experience was developed in the Unity game engine, with three different levels, each designed to expose the user in one of three different states (relaxation, concentration, stress). While in the VR Experience users physiological responses were measured using ECG and EEG sensors. After the experience users completed questionnaires to establish their perceived state during the levels, and to established the usability of the system. Next a comparison between the reported levels of emotion and the measured signals is presented, which show a strong correspondence between the two measures indicating that biosignals are a useful indicator of emotional state while in VR. Finally we make some recommendations on the practicalities of using biosensors, and design considerations for their incorporation in to a VR system, with particular focus on their integration in to task-based training and educational virtual environments.