About
PhD Candidate at RMIT University & the ARC Centre for Excellence in Automated Decision-Making and Society (ADM+S) | ADM+S Australian Search Experience 2.0 | Sessional Tutor for Psychology
Computational Models of Cognition | Resource-Rational Analysis | Sequential Latent-state Learning | Human-Computer Interaction
Polarization | Misinformation | Belief-updating | Fact-checking
The goal of my research is to understand how people learn and revise beliefs when time, attention, and computation are limited, and to use this understanding to design sequences of information that optimize learning under those constraints. I compare cases when learners begin with similar versus varied prior beliefs, asking how these starting points shape the ordering of items that yields the greatest learning for the least effort.
My current case study is political messages and news items, especially content that can mislead or provoke strong emotion. Using large-scale online experiments, I manipulate sequence features and measure when people shift their level of attention from intuitive judgments to focused deliberation, and how beliefs change in response.
To do so, we model how exposure history builds and tunes cached associations, or habits that tune compact, structure‑aware policies that can be executed rapidly without reconstructing the full causal model on every trial. The guiding assumption is that such models are candidates for capturing human behaviour because they allow us to think about judgments in terms of the optimal solutions to the problems the human mind faces, which in the case of online environments, is hypothetically to determine the cues which are most conducive to quick judgment. If attention and computation are limited, people want to attain a strategy that functions well across environments and so store coarse predictive categories, or, defaults used when the context is ambiguous. Crucially, these coarse categories often track stereotypes about the speaker, their motivations and what political or social groups they belong to, biasing decision-making towards these cues.
Algorithmic feeds now structure much of human information access, necessitating an investigation that formalizes these environments and the coarse categories used to navigate them. Different feeds expose people to different statistics, including those that frequently surprise to those that consistently affirm. These regimes modulate attention and memory and, in ways that we do not yet fully understand, increase cognitive load and degrade efficiency. We specify rules for these dynamics, with the aim of embedding learner models into information systems to enable adaptive placement of corrections or explanations when they can most effectively disrupt habitual responding and encourage planning.
Australasian Society for Philosophy and Psychology 2025 Conference (2025)
I want to put down some loosely constellated reflections on the conference, mainly the questions it raised that I rarely have to address in my work. Rather than answering those questions directly, I want to have a guess at why they might not come up that often for me. These reflections were prompted in particular by Dr Cammi Murrup‑Stewart’s talk, in which she argued that research relationships always negotiate and delimit what knowledge can be discovered, and called for a more deliberate recognition of the hidden assumptions underlying relationships we have long taken as simply given. She spoke about how deepening our awareness of our relationships with our teams and the people involved in our work allows us to deepen the knowledge we can uncover; and, conversely, how deepening our awareness of the assumptions in our knowledge can open up new, potentially closer modes of relating. Here, I want to focus on this second point, by offering an up‑front attempt at a conscious reflection on the assumptions that underpin my work.
I am an Anglo‑Australian, born to Anglo‑Australian parents who both dedicated their careers to education, one as a school principal and the other as a teacher. They were, respectively, the son of a clerk and a high‑school teacher, and the daughter of a meat carver and a truck driver. Both sides of my family emigrated to Australia after the First World War, from Ireland, Scotland, and France. I therefore come from a Western background, in which my heritage and my education are tensely interlinked.
Against this background, I reflect on what I think is possible in the world and realise that my own position in relation to knowledge is shaped by a certain framework. This framework centres on efficacy in the pursuit of contingent rewards, and on an assumption that no reward motivates human behaviour apart from those causally derived from genetics or from what we learn from the environment. Both are taken to structure a set of states accessed only via prior states, and by modelling those states one aims to obtain rewards at both the individual and collective level.
But I wonder, is there a type of motivation that could not be captured by that? Obviously, because I am only looking for empirical evidence, we naturally have to limit ourselves to something we can agree is true, and perhaps it takes something physically in our environment to garner enough consensus that we can call it empirical. A fleeting, deeply personal moment in imagination is not enough to get us there. This process of consensus is all too often engaged without concern for the possibility that, by structuring these states into causes represented by much‑debated theoretical stand‑ins, we may absent‑mindedly treat those stand‑ins as if they preserved the very complexity for which they were originally coined.
Initially, I took my experiments to be tools for inferring stable properties of the mind. Thinking more carefully about this however, especially in light of Dr Murrup‑Stewart’s talk, I now see that I may be relating to the mind in precisely the way the framework itself encourages, as something that takes on the structure built into experiments. The tasks used to capture psychological concepts may be as much about the mind’s ability to take on the structure of the task as about any pre‑existing mental property. The data I collect can therefore be attributed as much to the task as to the mind. I have little way of knowing whether the task engages an enduring property, or something constructed ad hoc for this particular setting. All I can say, then, is that I can model the environment computationally, and assume that the mind takes up that model in the form of what I describe as mapped items and outcomes, or associations. In other words, associations may be as much a product of the experimental environment as of anything intrinsic to the mind. I hesitate to attribute exactly that structure to “the mind itself”, whether as a model that precedes exposure to the environment, or as one that arises from it. For now, I treat such models as sufficient only for that purpose, and I want to be explicit that this is a constraint of my approach.
In this sense, experimental environments construct a relationship to the mind that resembles the research relationships Dr Murrup‑Stewart described. They negotiate and delimit what can appear as the mind in its outcome. The mind, embedded in the experimental environment, is penetrated by the patterning of the scenarios it is asked to behold. I can therefore reframe cognitive inferences as describing a structure or subjectivity that is technically owned by the environment, and only contingently by the participant. But at the same time, I really am tempted to talk about an “interior” environment that can sequence percepts in such a way that a subject emerges independently of the task design, the environment encountered, and the genes encoded. To approach that, I would again reach for computational tools to describe how patterns might be extracted from sequences. But how could I do so? It is partly for this reason that I leave out step three of Marr’s process in my own work and replace it with something that keeps the focus on this environment–mind relationship rather than on a fully specified internal representation.
What this brings into focus is a genuine problem for me: if my methods and relationships help constitute what I call “the mind”, how can I claim to be inferring anything beyond those methods and relationships? Causality, in my current framing, is largely described in terms of genetics and environment, and I risk treating all patterns as if they belong only to those domains. But causality cannot be assessed separately from questions about which causes ought to matter. The spaces occupied by morals, relationships, and even ancestral connections also enact causes from a position with as much reality as those imposed by genetics or our environment, and, as such, we are required to attribute the patterns we extract to an “I”, a “you”, and a “them”. Achieving that, valuing the rewards of contingencies not merely because they marginally relate to some intended outcome, is something my current viewpoint is not well equipped to do. It is incredibly difficult to understand the limits of such things, perhaps because they are the very stuff from which I have been constituted since birth. There may be causes that have always existed in eternity, much like mathematical truths, and yet, because they have not appeared in time as structured sequences, we can say nothing about them. From within my framework, behaviour becomes the activity of something I can only treat as an outcome of something else. The more I reflect on this, the more I see that my approach may be unable, on its own, to answer the question of what the mind is apart from the experimental and relational structures through which it is encountered. Rather than resolving that tension here, I want to name it as a limit within which my work currently operates, and to leave open the question of whether there is any way to infer “the mind” without such structures at all.
I presented a poster (see it here)
ADM+S hackathon: navigating the “wicked problems” of search (2025)
We just finished the ADM+S hackathon having won the second day, or the solution phase. The hackathon really felt like a top example of what I imagine they wanted to come from ADMs+ when they started it in the first place. Our team was a mix of qualitative and quantitative researchers, from computer science, media/comms, AI ethics, and myself in psychology, and I believe our solution really reflected everyone's taste.
I started with a lot of preconceived ideas, mainly inspired by one of my favourite papers, "The Two Settings of Kind and Wicked Learning Environments," which I really recommend you read. My inspiration was initially centered on the idea that an environment can be "wicked" in the sense that it is structured in such a way that the probabilistic relationships one has learned previously no longer apply. This negatively impacts inference, because people are acting on models of the environment without knowing that it is a model separate from the one in which they now take action. Funnily enough, I really appreciate my team taking me away from simply creating something based on what I already know, in environments I am already used to. What I learned previously didn't apply here, and hopefully I'm smarter for it. I'll attach the related paper once it's finished.
Below is an image of me trying to scare the other teams by holding a book filled with equations I don't really understand.
Using a psychological method to increase performance of LLM for relevance judgement and ranking tasks (2025)
Summary and links coming soon.
When order shapes belief: sequence design for bounded agents (2025)
People rarely encounter information in isolation. Instead, they encounter it through algorithmic feeds that embed items within sequences and contexts that shape their salience, credibility, and integration with prior beliefs. We argue that human learning from recommender systems and the ways it may differ, can be understood by considering credibility as a computational problem of inference over time. The learner must identify reliable cues and sources, allocate attention, and update beliefs under constraints, which constitutes a computational problem that the human learner is optimized to detect and exploit. We claim that this problem acquires a rough solution from three fundamental control systems that the brain uses to maximize functioning across two desirable but competing performance dimensions: speed-accuracy, stability-flexibility, and generalization-multitasking. If improvements on one predictably degrades the other, then sequences that push the system toward speed, stability, or divided attention should yield predictable biases in belief updating. These trade-offs shape belief updating under sequenced exposure and serve to explain common findings in the misinformation and digital persuasion literature.
Clickbait nudges social media away from accuracy and toward partisan bias (2025)
Prior work shows that accuracy prompts reliably shift attention towards more careful decision-making, reducing belief in misinformation. Variation in headline format, beyond its content, may likewise act as a prompt that changes which evaluative dimension is applied during veracity assessment. We formalize this question using a rational inattention framework in which people selectively allocate limited evidence accumulation across multiple evaluative dimensions. However, this account predicts either that sensational (“clickbait”) cues may make evidence utilization more difficult and encourage reliance on polarized beliefs, or by signalling uncertainty, they may elicit stricter decision criteria and reduce false endorsements. In total, 82 participants from the U.S. judged the veracity of 64 headlines, varying in political congruence (congruent/incongruent) and truth (true/false), and were randomly assigned either to view clickbait or neutral formats. Results showed that both classification accuracy and response time exhibited a significant three-way interaction among headline style, congruence, and truth. For the clickbait condition, judgements aligned more strongly with political congruence, leading to more correct judgements for aligned headlines and fewer for conflicting ones. This pattern weakened under neutral formatting. A hierarchical drift-diffusion model indicated that clickbait changed how evidence was accumulated, without making people more cautious or pre-biased at the start. Interpreted through rational inattention, these results suggest that clickbait makes it less worthwhile to accumulate evidence on the accuracy dimension. When accuracy evidence becomes noisier, partisanship is selected as an alternative, and judgments increasingly follow political congruence rather than truth.
Making accuracy easier: credibility perceptions as a latent cause assignment problem (2025)
The challenge of identifying reliable content amid widespread falsehoods requires a capacity for well-calibrated credibility judgments. We propose that humans address this challenge using statistical learning-the same cognitive process that helps us acquire language and recognize patterns. However, this adaptive mechanism may backfire in algorithmically curated information environments, leading to increased polarization and susceptibility to misinformation. We first outline a latent-cause model of credibility perception, which broadly describes the impact of structure learning, allowing people to categorize information based on recurring patterns (source identity, emotional tone, in-group signals) and then use these categories as shortcuts when evaluating new information. While this process is cognitively efficient, it can make users more vulnerable to misinformation that matches their learned patterns and more resistant to corrections that violate them. We present a Bayesian framework linking environmental statistics to belief formation, propose three testable predictions, and outline experimental protocols for studying these phenomena in controlled settings.
Confidence as inference over time: the effects of sequences on misinformation susceptibility (2025)
Overview and preregistration link coming soon.
Offgrid online (2025)
Like everyone without a vested interest in it's success, I'm pretty skeptical of A.I's ability to contribute positively, or even to help people contribute creatively. I think its hard to sell a second techbubble to people who are in the midst of throwing, or failing to throw, their addiction to the first. But I recently got a free year of Gemini under a student account and wanted to play around with what it could do. People reminisce on the pumping up of the first tech bubble, or the early days of the web, and see it as a place underestricted by the tollways and corporate border patrols that have since been set between ourselves and our intentions, before our names takes a detour between my phone and the printer, bouncing across the desks of the worlds top five richest companies. So I think, is this a time we will look back on similarly, before it inevitably also turns into a watered down slot machine of advertisements as our web experience now seems to be. That early time of the internet is looked on as a time of extreme creativity, partially due to survivorship bias maybe, and partially because it was before they really figured out how to monetize the damn thing. We're in the same spot now with AI, I think, most likely why I now find myself now with free access. They're hoping either I come up with whatever will make them trillions of dollars through use of the current tech, or become sufficiently addicted to what they have now that the whole innovation thing loses its real point. But anyway, this is what I made while playing around. Its buggy, and merely a stand-in for what I wanted to create. But it is fully AI coded. I keep mentioning the first tech bubble, among the already mentioned reasons, because if there is a lesson to be learnt from it, is the idea that resources do not equate to choice. The way that the internet was sold to us was that people would sample from ideas they never previously would have, creating a richer, more nuanced worldview, but the fact is that we don't do that. We were instead enabled to be even more like we would have been in the first place. I suspect we will see the same thing in A.I. People will be the most themselves to an extent they have ever been.
Offgrid Online is a browser extension that visualises the environmental cost of your computer use and the data centres that power it. Your activity is converted into research‑based CO₂ estimates and applied in real time to a simulated micro‑forest. You can restore the forest by making small donations to partnered environmental charities.
The tool watches for a few specific types of activity from your browser and turns them into estimated emissions. For each category, AI interaction, media streaming, page loads, and general browsing, two pieces of information are combined. First, an estimate of the computer's total electricity use while active, and second, an estimate of the load on data centres and networks. And yes... an estimation is made for what the tool itself is also expected to generate.
ADM+S Symposium — Automated social services: Building inclusive digital futures (2025)
It was really exciting to be able to hear everyone's thoughts on the progress of Australian Search Experience 2.0 team. It looks like the project is evolving into something very comprehensive. We spoke a lot about how people people appear to gain instrumental rewards at the same time as fulfilling their semantic or procedural information needs. This is a topic I am really interested in!
Report on the 3rd Workshop on NeuroPhysiological Approaches for Interactive Information Retrieval (NeuroPhysIIR 2025) at SIGIR CHIIR 2025
The International Workshop on NeuroPhysiological Approaches for Interactive Information Retrieval (NeuroPhysIIR’25), co-located with ACM SIGIR CHIIR 2025 in Naarm/Melbourne, Australia, included 19 participants who discussed 12 statements addressing open challenges in neurophysiological interactive IR. The report summarizes the statements presented and the discussions held at the full-day workshop.
The basic idea of what I promoted in the attached report is that the speed, density, and volatility of an interface is known have predictable impacts on how people make decisions. I put forward task demands as a construct for computer science capturing how the architecture of decision-making couples with the architecture of the task. Users pursue goals, but the choice they perceive and take are constrained not only by the available options, but additionally by the cognitive resources they can afford in the moment. In other words, the architecture of decision-making couples with the architecture of the task. This report bridges split-second chioces and longer-horizon learning by inverting the user's generative model of the task, and provides measurable constructs, such as priors, learning rates, and uncertainty from behaviour to reveal when interfaces change behaviour. The payoff is an adaptive design that detects overload and nudges users back toward accurate, less biased belief updating.
Papyrus (2025)
The scholarly record is increasingly fragmented: more than four million peer-reviewed papers now appear each year, yet mismatched metadata, discipline-specific jargon, and paywalls hinder efficient discovery. When the volume of information overwhelms a finite capacity, researchers must either spend months undertaking cross-disciplinary syntheses or retreat into an ever-narrower niche, fueling professional fatigue and the well-documented reasoning errors that accompany cognitive overload.
We re-frame this crisis as a tractable machine-learning problem. Transformer-based language models embed full-text articles in a shared semantic space, while citation graphs add complementary topological cues. Together these representations can automatically surface convergent findings, flag near-replications, and spotlight neglected intersections. Each of these tasks currently overwhelm manual review. Layering an adaptive recommender that infers a reader's evolving goals yields the benefit of passive discovery of high-utility papers as well as an explicit push toward serendipity, both measurable with standard ranked-retrieval and diversity metrics.
Papyrus is a desktop and mobile platform that atomises every article into fine-grained Units of Knowledge, or minimal, context-free propositions expressed in subject-predicate-object form. Each Unit inherits its parent article's embedding, enabling intra-paper summarisation and inter-paper synthesis through a two-stage pipeline (predicate-argument extraction followed by cosine-similarity clustering). Granularity adapts to expertise: novices see high-level explanatory Units, while domain experts receive technical Units that reveal citation pathways and unresolved contradictions.
Semester 1: Psychology of Everyday Thinking (1st year subject) — 2025
Began tutoring the course Psychology of Everyday Thinking for the first time, at the City Campus of RMIT University.
Dataset Sentiment Classifier for Experimental Stimuli (2025)
A new protocol is emerging in misinformation research that aims to improve the rigor of stimulus design. But pretesting statements for emotional or categorical content is often costly and places an undue burden on participants, who have to repeatedly evaluate large sets of stimuli simply to verify that they meet predefined criteria.
As a preliminary solution, machine learning offers a promising route for approximating human-like text perceptions. I've developed a simple tool on Kaggle that allows researchers to upload a dataset of statements and receive outputs categorizing each one by emotional sentiment and its intensity. Essentially, it serves as a user-friendly interface for psychological researchers without a technical background, providing access to a sentiment classifier without requiring extensive coding.
Warning: this method has not yet been validated against established psychological measures at the time of writing. Perhaps a future PhD project could take this further? Someone, pick it up!
Post-SWIRL’25 mini-conference (following the 4th SWIRL) — 2025
Brief recap and links coming soon.
Leveraging the serial position effect to enhance auditory recall for voice-only information retrieval (2024)
Prototype description and evaluation plan coming soon.
The role of conflict and alignment in shaping responses to polarizing factchecks (2024)
People often continue to endorse claims after fact-checks (the continued influence effect, CIE). Dominant accounts attribute CIE to a failure to detect conflict between prior beliefs and corrective evidence, or because, once they do notice, they respond in a reactive partisan-consistent way. In a 2×2 (Congruence × Truth) online experiment with N=200 Australian adults, we measured participant evaluations at two levels: rapid associations (Brief Implicit Association Test; BIAT) and normative judgments (self-report). We predicted that larger implicit–explicit discrepancies (IED) would magnify partisan-congruent processing if an identity-protective form of dissonance dominated. Contrary to prediction, while perceptions of statements were clearly shaped by partisanship and fact-checking robustly predicted belief updating, IED did not reliably moderate the degree of belief updating. Instead, higher overall source liking, indicated by alignment across measures, broadly dampened corrective shifts irrespective of congruence, suggesting that coherent intuitions about both preferred and nonpreferred sources set the weight of correction, not conflict between them. This pattern is consistent with conflict being registered at the statement–evidence level while source-level discrepancy remained inert. We propose that to increase item-level accuracy, corrections should generally suppress identity coactivation and keep sources peripheral so that judgments anchor themselves in statement-evidence coherence.
Semester 2: Cognitive Psychology (2nd year subject) — 2024
Began tutoring the course Cognitive Psychology for the first time, at the City Campus of RMIT University.
Semester 1: Philosophy & Methodology of Psychology (3rd year subject) — 2024
Began tutoring the course Philosophy & Methodology of Psychology for the first time, at the City Campus of RMIT University.
PhD Proposal — Impact of Information Environments on the Perception of Credibility (2024)
Relevance and credibility are shaped by the information environments people inhabit and the noise they encounter. Information environments range from feeds that consistently surprise the user to those that consistently affirm them, each altering attention and memory systems in distinct ways. Previous research has shown that noise, defined as inconsistent or distracting information, can disrupt attention, increase cognitive load, and reduce efficiency.
As algorithms increasingly become the main source of information globally, it is essential to investigate how the structure these algorithms impose on information affects the way individuals maintain and revise their beliefs, accounting for the varying presence of noise in these feeds. Understanding how these varying environments impact perception is key to unraveling the cognitive and neural mechanisms involved in human-information interaction.
The primary aim of this topic is to explore the relationship between different information environments and their impact on perceptions of credibility. Specifically, it seeks to investigate how various sequences affect information processing and beliefs regarding the specific information being sequenced. This is to examine the potentially disruptive effects of noise during credibility judgements and maintenance of states of cognitive control.
Previous research has revealed adaptive strategies employed by the meta-cognitive system in noisy environments. But as of yet, it is unknown whether these strategies are also utilized in algorithmically sequenced news information, potentially explaining key findings in studies of misinformation and the susceptibility of people to polarized narratives when such narratives confirm existing beliefs. These investigations seek to contribute to the body of literature tasked with developing strategies to moderate credibility perception in diverse information settings.