Institution profile

Trinity College Dublin

Academic institutioneurope · ie
Official website
Research library291linked papers
Opportunities0open roles
Selected work

Representative Papers

Finite Neural Networks as Mixtures of Gaussian Processes: From Provable Error Bounds to Prior Selection

Jul 26, 2024arXiv.org

Existing finite-width deep neural networks lack analytically tractable Gaussian process (GP) approximations with provable error bounds. Method: We propose the first Gaussian Process Mixture (GPM) approximation framework with certified error bounds: leveraging Wasserstein distance to model output distributions layer-wise, it achieves ε-accurate approximation of arbitrary non-i.i.d. parameterized networks over finite input sets. The method integrates optimal transport theory with hierarchical probabilistic modeling, yielding differentiable error bounds that guide network parameter optimization toward user-specified prior distributions. Results: Experiments demonstrate that GPM enables controllable-accuracy approximation on both regression and classification tasks, while simultaneously supporting principled uncertainty quantification and Bayesian prior design—bridging finite-width neural networks and rigorous GP inference with guaranteed approximation quality.

4 citationsRead paper

Dynamic Model Routing and Cascading for Efficient LLM Inference: A Survey

Feb 23, 2026

Static deployment of large language models (LLMs) struggles to dynamically select the optimal model based on query complexity and domain, leading to an imbalance between performance and cost. This work proposes a three-dimensional framework—centered on decision timing, information sources, and computation mechanisms—to systematically analyze dynamic routing and cascading strategies across multiple LLMs. It introduces the first taxonomy of routing paradigms spanning independently trained LLMs and integrates diverse technical approaches, including query difficulty estimation, human preference modeling, uncertainty quantification, reinforcement learning, and multimodal fusion. Experimental results demonstrate that well-designed routing systems can surpass the strongest individual model, achieving a superior trade-off between inference efficiency and performance, while also highlighting critical challenges such as generalization.

2 citationsRead paper

Mapping the Regulatory Learning Space for the EU AI Act

Feb 27, 2025arXiv.org

Facing challenges posed by rapid AI technological evolution, regulatory uncertainty, and difficulties in cross-level coordination under the EU AI Act, this study introduces the “AI Regulatory Learning Space” — the first systematic theoretical framework bridging the gap between technical regulation and sectoral enforcement. Methodologically, it integrates RegTech modeling, multi-stakeholder collaborative learning, policy pathway analysis, and adaptive open-data governance. Contributions include: (1) a novel, mapping-capable, and operationally deployable regulatory learning space tool; (2) a dynamic implementation and adaptive governance roadmap for EU Member States; and (3) a paradigm shift in AI governance—from principle-based approaches toward reproducible, accountable, and standardized risk quantification—thereby strengthening the empirical foundations of fairness, transparency, and cross-rights coordination.

1 citations1 influentialRead paper

Agentic Design Patterns: A System-Theoretic Framework

Jan 27, 2026

Current agent system designs often lack grounding in systems theory, resulting in ad hoc architectures prone to hallucination and reasoning flaws that undermine reliability. This work addresses this gap by introducing systems theory into agent architecture design for the first time, proposing a structured framework composed of five core functional subsystems. Building on this foundation, the authors abstract twelve reusable and clearly categorized agent design patterns. Through the reconstruction and validation of representative frameworks such as ReAct, the proposed approach effectively rectifies inherent architectural deficiencies, significantly enhancing modularity, interpretability, and reliability. This contribution establishes a standardized language and a structured development paradigm for agent engineering, offering a principled foundation for future research and practice.

1 citationsRead paper

Multi-Partner Project: COIN-3D -- Collaborative Innovation in 3D VLSI Reliability

Jan 20, 2026

This work addresses the reliability challenges in 3D chip stacking arising from process scaling—such as the transition from FinFET to GAAFET—and heterogeneous chiplet integration. In collaboration with leading European research institutions, the project presents the first open-source reliability analysis toolchain tailored for 2.5D/3D VLSI systems. The toolchain integrates multi-scale modeling across physical and system levels, leveraging advanced algorithms within an open-source EDA framework to enable comprehensive reliability assessment in heterogeneous chiplet integration scenarios. Its modular and extensible architecture significantly enhances the capability to predict and optimize reliability during the design phase of 3D integrated systems.

1 citationsRead paper
Recent publications

Latest Papers