Integrated photonic ultrasound transducers (IPUTs) are compact, high-sensitivity devices that combine mechanical sensing with optical readout using integrated photonics. IPUTs typically consist of optical waveguides integrated on a thin mechanical plate that serves as the acoustic sensing element. In many realizations, this plate is formed from thermally oxidized silicon dioxide layers commonly used in photonic fabrication processes. The oxidation process introduces significant residual compressive stress–typically between 200 MPa and 400 MPa–as the structure cools to room temperature. Such stresses can strongly influence the dynamic response of the plate through their contribution to the geometric stiffness of the structure. In this work, the influence of internal stress on the resonance frequency and receive transfer function (RTF) of IPUTs is investigated. Finite element models incorporating residual stress and geometric nonlinearity are developed and validated against experimental measurements and results reported in the literature. Parametric analysis shows that increasing compressive stress progressively reduces the resonance frequency while enhancing the RTF as the structure approaches the critical buckling condition. Beyond this point, changes in the prestressed equilibrium configuration lead to transitions in the dominant vibration mode, producing abrupt variations in the resonance frequency and RTF. These results highlight the importance of accounting for residual stress in the design and analysis of IPUTs and similar plate-based acoustic sensors to ensure reliable dynamic performance and predictable sensitivity.

Current inverse materials discovery methods face a trade-off between broad exploration of chemical space and control over chemical validity, synthesisability, and target properties. Here, we present the COMPosition Aware Search Strategy (COMPASS), a constrained multi-objective, multi-fidelity search framework for crystalline composition spaces. COMPASS introduces a discrete mixed-site encoding for material families with fixed site stoichiometries and up to two species mixed on each crystallographic site. This encoding preserves chemical identity, allowing empirical chemical rules and property constraints to be evaluated directly during optimisation. COMPASS combines fast composition-only screening with a constrained genetic algorithm, structure-based verification using machine-learning interatomic potentials, and active learning to improve the low-fidelity model. Applied to mixed-site ABX3 perovskites, COMPASS identifies 15,922 computationally promising candidates satisfying chemical, novelty, stability, and band-gap criteria. In the same constrained discovery task, COMPASS achieves an approximately two-orders-of-magnitude higher yield of desired candidates than the tested open-source MatterGen baselines [Zeni et al., Nature, 2025, 639, 624--632]. These results position COMPASS as a framework for chemically well-understood discovery problems where chemical constraints can guide search through large composition spaces.

Research on climate change increasingly highlights the importance of consumer behaviour in reducing greenhouse gas emissions. However, existing studies have mainly focused on either average consumers or the outsized objective environmental impact of high-welfare individuals, offering limited insight into how psychological factors shape pro-environmental behaviour across welfare groups. As a result, the relation between welfare, psychological measures, and pro-environmental behaviour remains insufficiently understood.
This study examines how individuals across welfare groups perceive self-accountability, agency, and efficacy, and how these psychological factors relate to pro-environmental intentions and behaviour across multiple consumption domains. Using a Dutch sample with a wide range of welfare levels, it also analyses how accountability is attributed to different societal actors and how welfare influences psychological predictors of behaviour in domains including fashion consumption, air travel, household heating, meat consumption, and green investment.
The findings show that perceptions of accountability to mitigate GHG emissions are broadly similar across welfare groups, with governments and large corporations seen as most accountable and individuals as least accountable. Welfare had no direct effect on perceived self-accountability, agency, or efficacy. These psychological factors were generally positive predictors of pro-environmental intentions and behaviour across domains. However, their predictive power weakened in specific domains, with the largest intention–behaviour gaps emerging in unbounded, identity-driven domains such as air travel and fashion consumption. Notably, it is also in these same domains that welfare showed significant direct effects on behaviour, with higher-welfare individuals engaging in more environmentally harmful consumption. While psychological measures were largely stable across welfare levels, this pattern suggests that welfare becomes more influential precisely in the domains where psychological factors are least explanatory. Overall, the results indicate that pro-environmental behaviour is shaped by both psychological processes and structural conditions, highlighting the need for integrated approaches that combine behavioural interventions with policies addressing consumption inequalities.
Finally, the findings indicate that climate interventions should focus on reducing the perception–reality gap in environmental impact. This may be achieved by aligning perceptions of “normal” consumption (through improved public information and reduced socio-economic segregation), strengthening carbon literacy (through education, point-of-decision information, and footprint comparison tools), and improving awareness of emission differences across income groups. At the same time, psychological interventions alone are unlikely to be sufficient in domains such as air travel and fashion consumption, where behaviour is strongly shaped by structural conditions. In these domains, policies such as aviation taxes, frequent flyer levies, improved access to low-carbon alternatives, and measures that slow fashion cycles while promoting more pro-environmental and appealing alternatives may be more effective.

Large language models can support automated fact-checking by reasoning over claims and evidence, but single generated explanations can be unreliable for numerical, temporal, or complex claims. Verifier-guided Best-of-N (BoN) decoding improves performance by sampling multiple reasoning traces at inference time and selecting the trace most strongly supported by a verifier. However, exhaustive BoN applies the same fixed inference budget to every claim, making it costly even when a stable prediction emerges early. This thesis proposes a joint generation-and-verification optimization framework for dynamically allocating test-time computation in verifier-guided fact-checking through two methods. Verifier-Guided Adaptive Stopping uses verifier feedback as a confidence signal to stop generation once the prediction appears stable. Surrogate-Guided Selective Verification further uses an online linear surrogate model to estimate trace utility and prioritize which traces are sent for verification. A fixed verifier budget experiment then examines whether surrogate-guided allocation uses limited verifier calls more effectively than generation-order or random selection.
Experiments on QuanTemp and ClaimDecomp show that adaptive stopping reduces inference latency and estimated cost by 42.0% and 43.7%, respectively, while maintaining performance comparable to exhaustive BoN. Surrogate-guided verification reduces verifier calls by 44.5% and 53.7% relative to exhaustive BoN, with additional reductions of 9.4% and 4.0% over adaptive stopping. Under limited verifier budgets, surrogate-guided allocation often outperforms generation-order and random selection at lower budget levels across the evaluated settings, showing that learned trace ordering can improve verification when verifier calls are constrained. These results show that adaptive generation and selective verification preserve much of the benefit of multi-trace reasoning while substantially reducing inference cost, and that fixed-budget surrogate allocation can prioritize useful traces when verifier calls are limited.

This research explores how shrinking rural villages can strengthen their identity through public space, accessibility and multifunctional facilities. The research focuses on the village of Krabbendijke in Zeeland, where processes such as facility decline, increasing mobility dependency and the fragmentation of everyday life influence the social, functional and symbolical identity of the village. As discussed by Christiaanse (2023), shrinking rural regions are increasingly dealing with changing accessibility, disappearing facilities and transformations in everyday village life. Rather than approaching shrinkage only as a demographic or economic issue, this research investigates how these processes affect social interaction, public life and the experience of the village environment.

Through literature research, spatial analysis, fieldwork and research-by-design, the relationship between accessibility, facilities, public space and identity is analysed on multiple scales. The research identifies three interconnected forms of identity within shrinking villages: functional identity, social identity and symbolical identity. These identities are explored through themes such as local habits, historical values, landscape qualities, community spaces, spatial accessibility, functional structures and social interaction potential.

The research demonstrates that facilities function as more than functional services alone. They often act as social anchors, meeting places and symbolical carriers of village identity. Their disappearance therefore affects not only accessibility, but also collective memory, social interaction and public life. At the same time, the research concludes that the car remains an important part of rural life within Zeeland since there are lots of villages with limited accessibility. The challenge therefore lies in balancing accessibility and mobility with more socially active and walkable public spaces.

Through research-by-design, these findings are translated into a spatial vision for Krabbendijke, structured around four overarching strategies focused on multifunctionality, social interaction, intergenerational resilience and landscape identity. These strategies are tested through interventions on smaller scales within the village core, station area and connecting landscape structures. The design proposal illustrates how public space can strengthen the relationship between social, functional and symbolical identity by reconnecting facilities, daily routines, landscape and cultural structures to everyday village life.

Although the spatial outcomes are highly context-specific, the analytical framework and thematic approach developed throughout this research can also contribute to comparable shrinking rural regions within Zeeland and beyond.