arXiv subjects

Cedric Lim

Publications and source records attributed to Cedric Lim.

5 recordsLinked to original sources

arXiv2026

Declarative Skills for AI Agents in Knowledge-Grounded Tool-Use Workflows

We study orchestration mechanisms for tool-using AI agents in realistic customer-service workflows over an unstructured knowledge base. We argue that declarative agents -- AI agents equipped with natural-language skill files appended to the system prompt -- are an effective orchestration paradigm. Concretely, we compare (i) a DeclarativeAgent that reads three domain-specific skill files at inference time and decides its own control flow, (ii) an ImperativeAgent based on a programmatic state machine with explicit phases, and (iii) an unscaffolded baseline agent modeled after the $\tau$-Knowledge benchmark agent. Our ImperativeAgent is motivated by externalised-control inference as in Recursive Language Models and graph-based orchestration frameworks. We formalise the three agents as policy classes within a decentralised partially-observable Markov decision process and analyse their information-theoretic and structural properties; we then test the predicted differences empirically on five language models and two retrieval regimes. Our results show that retrieval quality is a dominant bottleneck for AI agents: when evidence is incomplete or skewed, all agents degrade substantially, and skill files cannot recover lost performance. Under high-quality retrieval, however, declarative skills consistently improve accuracy on procedural tasks and reduce orchestration errors, while the imperative state machine's brittleness does not reliably improve task success or compliance.

cs.AI↗

arXiv2026

Plasmonic Photocatalysis Enables Selective Oxidative Coupling of Methane with Nitrous Oxide under Ambient Conditions

Methane (CH4) and nitrous oxide (N2O) are potent greenhouse gases that represent substantial chemical energy. Conversion of these abundant waste gases to high-value chemicals typically requires high temperatures up to 1000 C, producing substantial CO2 emissions and limited selectivity toward desirable multi-carbon products. Here we demonstrate a plasmonic photocatalyst that enables CH4 and N2O conversion under ambient conditions to form C2 and C3 hydrocarbons. By systematically tuning AuPd alloys on TiO2, we identify an optimal composition (AuPd0.05) where Au enhances light harvesting and Pd enables selective C-H activation and C-C coupling. Under visible-light illumination, this catalyst produces C2H4, C2H6, C3H6, and C3H8 with ~80% selectivity while suppressing CO2 formation. In-situ spectroscopy and hot-carrier calculations show that plasmon-generated carriers redistribute interfacial hydroxyl intermediates, shifting the hydrophilic center to suppress overoxidation. Ab-initio calculations further reveal the reduction in C-C coupling barriers from 2.7 eV to 0.7 eV under illumination. Our work illustrates how engineering interfacial electronic and adsorbate dynamics enables selective multicarbon formation.

cond-mat.mtrl-sci↗

arXiv2026

In situ Gas-Cell Electron Microscopy Reveals Pressure-Selected Restructuring Pathways in AuRu Ammonia Catalysts

Bimetallic catalysts provide new routes toward sustainable ammonia synthesis, but the nanoscale structural dynamics under reaction-relevant conditions remain poorly understood. Here, we combine in situ gas-cell and multimodal electron microscopy to determine how temperature, gas pressure, and chemistry select among distinct restructuring pathways in AuRu nanocrystal catalysts. Initially, the AuRu nanocrystals form polycrystalline face-centered cubic (FCC) alloys with Au/Ru intermixing. Elevated temperature ($\geq 350~^\circ$C) induces intraparticle phase segregation into distinct Au-rich (FCC) and Ru-rich hexagonal close-packed (HCP) domains that exhibit localized plasmonic modes. Atmospheric-pressure 3:1 H$_2$:N$_2$ gas unlocks a distinct restructuring regime absent at lower pressures, characterized by pronounced faceting and nanovoid formation. Systematic gas variation identifies hydrogen as the dominant driver. Density functional theory-trained machine-learning interatomic potentials and grand-canonical Monte Carlo simulations reveal that H-Ru interactions enhance the Au/Ru diffusivity mismatch, promoting vacancy accumulation and nanovoid formation. Together, these results show that, rather than simply accelerating the thermally driven phase segregation observed at lower pressures, atmospheric-pressure H$_2$:N$_2$ gas redirects restructuring toward faceting and nanovoid formation through a gas-mediated Kirkendall-type mechanism.

cond-mat.mtrl-sci↗

arXiv2025

Missing Wedge Inpainting and Joint Alignment in Electron Tomography through Implicit Neural Representations

Electron tomography is a powerful tool for understanding the morphology of materials in three dimensions, but conventional reconstruction algorithms typically suffer from missing-wedge artifacts and data misalignment imposed by experimental constraints. Recently proposed supervised machine-learning-enabled reconstruction methods to address these challenges rely on training data and are therefore difficult to generalize across materials systems. We propose a fully self-supervised implicit neural representation (INR) approach using a neural network as a regularizer. Our approach enables fast inline alignment through pose optimization, missing wedge inpainting, and denoising of low dose datasets via model regularization using only a single dataset. We apply our method to simulated and experimental data and show that it produces high-quality tomograms from diverse and information limited datasets. Our results show that INR-based self-supervised reconstructions offer high fidelity reconstructions with minimal user input and preprocessing, and can be readily applied to a wide variety of materials samples and experimental parameters.

eess.IV↗

arXiv2024

Defining classical and quantum chaos through adiabatic transformations

We propose a formalism which defines chaos in both quantum and classical systems in an equivalent manner by means of \textit{adiabatic transformations}. The complexity of adiabatic transformations which preserve classical time-averaged trajectories (quantum eigenstates) in response to Hamiltonian deformations serves as a measure of chaos. This complexity is quantified by the (properly regularized) fidelity susceptibility. Physically this measure quantifies long time instabilities of physical observables due to small changes in the Hamiltonian of the system. Our exposition clearly showcases the common structures underlying quantum and classical chaos and allows us to distinguish integrable, chaotic but non-thermalizing, and ergodic/mixing regimes. We apply the fidelity susceptibility to a model of two coupled spins and demonstrate that it successfully predicts the universal onset of chaos, both for finite spin $S$ and in the classical limit $S\to\infty$. Interestingly, we find that finite $S$ effects are anomalously large close to integrability.

cond-mat.stat-mech↗