arXiv subjects

Ian Stewart

Publications and source records attributed to Ian Stewart.

At least 19 recordsLinked to original sources

arXiv2026

Feasibility of a Flexible, Hybrid Tokamak-Stellarator Experiment using an Axisymmetric Dipole Coil Array

We demonstrate the design of a flexible, university-scale hybrid tokamak-stellarator experiment based on an axisymmetric array of planar HTS dipole coils. Because the coil array has few geometric degrees of freedom, we use single-stage optimization of the coil currents initialized from two-stage solutions to obtain mutually consistent equilibria and coil sets within realistic engineering limits. We find that the field error and coil current thresholds set minimum and maximum coil-plasma distances that confine the boundary to a roughly fixed axisymmetric envelope, within which rotational transform, volume, coil current, and quasi-symmetry (QS) error trade off against one another. Tighter current limits delocalize the non-axisymmetric shaping and raise QS error at fixed transform. From this single coil array we obtain a broad range of equilibria-quasi-axisymmetric vacuum stellarators with $\iota$ up to 0.2, finite-$\beta$ hybrids with realistic profiles reaching on-axis $\iota$ $\approx$ 1 and vacuum transform relevant for MHD stabilization, and strongly shaped tokamaks with elongation $\kappa$ = 1.7 and triangularity $\delta$ = $\pm$0.6, all at peak pointwise coil forces well below the HTS tolerance. We show the same array can additionally correct toroidal field (TF) coil ripple, reducing the number of TF coils required compared to the equivalent tokamak. These results establish the design as a promising platform for hybrid tokamak-stellarator research.

physics.plasm-ph↗

arXiv2026

MULTISEISMO: A Multimodal Seismic Dataset and Model for Cross-Modal Seismic Understanding

The application of generalist multimodal models (GMMs) to specialized scientific domains remains limited due to the scarcity of comprehensive domain-specific datasets that integrate multiple data modalities beyond text and images. In seismology, understanding earthquake phenomena requires the synthesis of timeseries waveform data, geographical imagery, and contextual metadata, a multimodal integration absent in existing seismic datasets. We present MultiSeismo, a large scale structured multimodal seismic dataset, comprising over 16K seismic events spanning 13 years (2010 to 2023) across diverse geographical regions. Each event data integrates waveform recordings from global station networks, intensity maps, population exposure visualizations, and a comprehensive textual description within a standardized JSON format. We additionally develop MISCE, a multimodal instruction set on top of raw data to enable supervised training and evaluation of GMMs on seismic reasoning tasks ranging from basic information retrieval to complex cross modal analysis. We leverage MISCE to finetune an existing multimodal model (Unified IO 2) enhanced with a specialized timeseries encoder, which yields SeisModal, the first domain specific multimodal model for comprehensive seismic analysis. Evaluation of state of the art multimodal models on MultiSeismo reveals significant challenges, particularly with time-series data processing for general purpose models, while demonstrating SeisModal's superior performance on seismic multimodal reasoning tasks. These results prove that MultiSeismo provides a rigorous benchmark for future multimodal research in seismology and validate the success of our domain specific architectural adaptations.

cs.LG↗

arXiv2026

Back to the Barn with LLAMAs: Evolving Pretrained LLM Backbones in Finetuning Vision Language Models

Vision-Language Models (VLMs) have rapidly advanced by leveraging powerful pre-trained Large Language Models (LLMs) as core reasoning backbones. As new and more capable LLMs emerge with improved reasoning, instruction-following, and generalization, there is a pressing need to efficiently update existing VLMs to incorporate these advancements. However, the integration of new LLMs into VLMs, particularly how the evolving LLMs contribute to multimodal reasoning, alignment, and task-specific performance remains underexplored. Addressing this gap is important for VLM development, given the rapid evolution of pretrained LLM backbones. This study presents a controlled and systematic investigation of how changes in the pretrained LLM backbone affect downstream VLM task performance. By having the vision encoder, training data, and post-training algorithm remain same across LLAMA-1, LLAMA-2, and LLAMA-3 based VLMs, we find that newer LLM backbones do not always lead to better VLMs, but the performance depends on the downstream VLM task. For example, in visual question and answering tasks, newer LLM backbones tend to solve different questions rather than just more questions, and our analysis shows this is driven by differences in how the models process information, including better calibrated confidence and more stable internal representations. We also find that some VLM capabilities appear only in the newest LLM generation, while tasks that depend mainly on visual understanding see little benefit from a newer LLM backbone.

cs.AI↗

arXiv2025

Recognizing Distance-Count Matrices is Difficult

Axiomatization of centrality measures often involves proving that something cannot hold by providing a counterexample (i.e., a graph for which that specific centrality index fails to have a given property). In the context of geometric centralities, building such counterexamples requires constructing a graph with specific distance counts between nodes, as expressed by its distance-count matrix. We prove that deciding whether a matrix is the distance-count matrix of a graph is strongly NP-complete. This negative result implies that a brute-force approach to building this kind of counterexample is out of question, and cleverer approaches are required.

cs.SI↗

arXiv2025

Synchronous Propagation of Periodic Signals in Feedforward Networks of Standard Model Neurons

Periodic signals propagating along chains are common in biology, for example in locomotion and peristalsis, and are also of interest for continuum robots. In previous work we constructed such networks as 'feedforward lifts' of a central pattern generator (CPG). When the CPG undergoes periodic oscillations, created by Hopf bifurcation or other mechanisms, it can then transmit periodic signals along one or more feedforward chains in a synchronous or phase-synchronous manner. We proved necessary and sufficient conditions for the stability of these lifted periodic orbits, in several senses. Here we examine the implications of the resulting theory for chains of neurons, using several standard neuron models: FitzHugh-Nagumo, Morris-Lecar, Hindmarsh-Rose, and Hodgkin-Huxley. We compare different notions of transverse stability, and summarize some numerical simulations showing that for all these neuron models the propagating signal can be transversely Floquet stable. Finally we discuss implications for less idealized models.

nlin.CD↗

arXiv2025

Stable Synchronous Propagation in Feedforward Networks for Biped Locomotion

Rhythmic gait patterns in animal locomotion are widely believed to be produced by a central pattern generator (CPG), a network of neurons that drives the muscle groups. In previous papers we have discussed how phase-synchronous signals can propagate along chains of neurons using a feedforward lift of the CPG, given sufficient conditions for stability to synchrony-breaking perturbations, and shown that stable signals are common for four standard neuron models. Here we apply these ideas to biped locomotion using a fifth model: Wilson-Cowan (or rate model) neurons. Feedforward architecture propagates the phase pattern of a CPG along a chain of identical modules. For certain rate models, we give analytic conditions that are sufficient for transverse Liapunov and Floquet stability. We compare different notions of transverse stability, summarise some numerical simulations, and outline an application to signal propagation in a model of biped locomotion.

math.DS↗

arXiv2025

Symmetries of Living Systems: Symmetry Fibrations and Synchronization in Biological Networks

A symmetry is a `change without change'. As simple as it sounds, this concept is the fundamental cornerstone that unifies all branches of theoretical physics. Virtually all physical laws -- ranging from classical mechanics and electrodynamics to relativity, quantum mechanics, and the standard model -- can be expressed in terms of symmetry invariances. In this book, we explore whether the same principle can also explain the emergent laws of biological systems. We introduce a new geometry for biological networks and AI architectures, drawing inspiration from the mystic genius of Grothendieck's fibrations in category theory. We attempt to bridge the gap between physics and biology using symmetries but with a twist. The traditional symmetry groups of physics are global and too rigid to describe biology. Instead, the novel notion of symmetry fibration is local, flexible, and adaptable to evolutionary pressures, providing the right framework for understanding biological complexity. In other words, this more general symmetry invariance is necessary and sufficient to ensure that a given biological network configuration can support a synchronized function. In this book, we review the theoretical progress over the last decades from mathematics, physics, computer science, dynamical systems, and graph theory that has led to the discovery of symmetry fibrations in biological networks. These symmetries act as organizing principles for biological networks. They serve as effective tools for describing the structure of these networks, blending geometry and topology. Fibrations explain how structure dictates function across various biological domains, including the transcriptome, proteome, metabolome, and connectome. Additionally, they facilitate a reduction in the dimensionality of the network, simplifying it into its fundamental building blocks for biological computation.

q-bio.MN↗

arXiv2025

Branching Ratios of Input Trees for Directed Multigraphs

We define the branching ratio of the input tree of a node in a finite directed multigraph, prove that it exists for every node, and show that it is equal to the largest eigenvalue of the adjacency matrix of the induced subgraph determined by all upstream nodes. This real eigenvalue exists by the Perron-Frobenius Theorem for non-negative matrices. We motivate our analysis with simple examples, obtain information about the asymptotics for the limit growth of the input tree, and establish other basic properties of the branching ratio.

math.CO↗

arXiv2024

ThinCurr: An open-source 3D thin-wall eddy current modeling code for the analysis of large-scale systems of conducting structures

In this paper we present a new thin-wall eddy current modeling code, ThinCurr, for studying inductively-coupled currents in 3D conducting structures -- with primary application focused on the interaction between currents flowing in coils, plasma, and conducting structures of magnetically-confined plasma devices. The code utilizes a boundary finite element method on an unstructured, triangular grid to accurately capture device structures. The new code, part of the broader Open FUSION Toolkit, is open-source and designed for ease of use without sacrificing capability and speed through a combination of Python, Fortran, and C/C++ components. Scalability to large models is enabled through use of hierarchical off-diagonal low-rank compression of the inductance matrix, which is otherwise dense. Ease of handling large models of complicated geometry is further supported by automatic determination of supplemental elements through a greedy homology approach. A detailed description of the numerical methods of the code and verification of the implementation of those methods using cross-code comparisons against the VALEN code and Ansys commercial analysis software is shown.

physics.plasm-ph↗

arXiv2024

Surprisingly Fragile: Assessing and Addressing Prompt Instability in Multimodal Foundation Models

Multimodal foundation models (MFMs) such as OFASys show the potential to unlock analysis of complex data such as images, videos, and audio data via text prompts alone. However, their performance may suffer in the face of text input that differs even slightly from their training distribution, which is surprising considering the use of modality-specific data to "ground" the text input. This study demonstrates that prompt instability is a major concern for MFMs, leading to a consistent drop in performance across all modalities, but that instability can be mitigated with additional training with augmented data. We evaluate several methods for grounded prompt perturbation, where we generate perturbations and filter based on similarity to text and/or modality data. After re-training the models on the augmented data, we find improved accuracy and more stable performance on the perturbed test data regardless of perturbation condition, suggesting that the data augmentation strategy helps the models handle domain shifts more effectively. In error analysis, we find consistent patterns of performance improvement across domains, suggesting that retraining on prompt perturbations tends to help general reasoning capabilities in MFMs.

cs.CL↗

arXiv2024

Benchmarking LLMs for Environmental Review and Permitting

The National Environment Policy Act (NEPA) stands as a foundational piece of environmental legislation in the United States, requiring federal agencies to consider the environmental impacts of their proposed actions. The primary mechanism for achieving this is through the preparation of Environmental Assessments (EAs) and, for significant impacts, comprehensive Environmental Impact Statements (EIS). Large Language Model (LLM)s' effectiveness in specialized domains like NEPA remains untested for adoption in federal decision-making processes. To address this gap, we present NEPA Question and Answering Dataset (NEPAQuAD), the first comprehensive benchmark derived from EIS documents, along with a modular and transparent evaluation pipeline, MAPLE, to assess LLM performance on NEPA-focused regulatory reasoning tasks. Our benchmark leverages actual EIS documents to create diverse question types, ranging from factual to complex problem-solving ones. We built a modular and transparent evaluation pipeline to test both closed- and open-source models in zero-shot or context-driven QA benchmarks. We evaluate five state-of-the-art LLMs using our framework to assess both their prior knowledge and their ability to process NEPA-specific information. The experimental results reveal that all the models consistently achieve their highest performance when provided with the gold passage as context. While comparing the other context-driven approaches for each model, Retrieval Augmented Generation (RAG)-based approaches substantially outperform PDF document contexts, indicating that neither model is well suited for long-context question-answering tasks. Our analysis suggests that NEPA-focused regulatory reasoning tasks pose a significant challenge for LLMs, particularly in terms of understanding the complex semantics and effectively processing the lengthy regulatory documents.

cs.CL↗

arXiv2024

Classification of 3-Node Restricted Excitatory-Inhibitory Networks

We classify connected 3-node restricted excitatory-inhibitory networks, extending our previous paper (`Classification of 2-node Excitatory-Inhibitory Networks', Mathematical Biosciences 373 (2024) 109205). We assume that there are two node-types and two arrow-types, excitatory and inhibitory; all excitatory arrows are identical and all inhibitory arrows are identical; and excitatory (resp. inhibitory) nodes can only output excitatory (resp. inhibitory) arrows. The classification is performed under the following two network perspectives: ODE-equivalence and minimality; and valence less or equal to 2. The results of this and the previous work constitute a first step towards analysing dynamics and bifurcations of excitatory-inhibitory networks and have potential applications to biological network models.

math.DS↗

arXiv2024

Generalist Multimodal AI: A Review of Architectures, Challenges and Opportunities

Multimodal models are expected to be a critical component to future advances in artificial intelligence. This field is starting to grow rapidly with a surge of new design elements motivated by the success of foundation models in natural language processing (NLP) and vision. It is widely hoped that further extending the foundation models to multiple modalities (e.g., text, image, video, sensor, time series, graph, etc.) will ultimately lead to generalist multimodal models, i.e. one model across different data modalities and tasks. However, there is little research that systematically analyzes recent multimodal models (particularly the ones that work beyond text and vision) with respect to the underling architecture proposed. Therefore, this work provides a fresh perspective on generalist multimodal models (GMMs) via a novel architecture and training configuration specific taxonomy. This includes factors such as Unifiability, Modularity, and Adaptability that are pertinent and essential to the wide adoption and application of GMMs. The review further highlights key challenges and prospects for the field and guide the researchers into the new advancements.

cs.CL↗

arXiv2024

Homeostasis in Input-Output Networks: Structure, Classification and Applications

Homeostasis is concerned with regulatory mechanisms, present in biological systems, where some specific variable is kept close to a set value as some external disturbance affects the system. Mathematically, the notion of homeostasis can be formalized in terms of an input-output function that maps the parameter representing the external disturbance to the output variable that must be kept within a fairly narrow range. This observation inspired the introduction of the notion of infinitesimal homeostasis, namely, the derivative of the input-output function is zero at an isolated point. This point of view allows for the application of methods from singularity theory to characterize infinitesimal homeostasis points (i.e. critical points of the input-output function). In this paper we review the infinitesimal approach to the study of homeostasis in input-output networks. An input-output network is a network with two distinguished nodes `input' and `output', and the dynamics of the network determines the corresponding input-output function of the system. This class of dynamical systems provides an appropriate framework to study homeostasis and several important biological systems can be formulated in this context. Moreover, this approach, coupled to graph-theoretic ideas from combinatorial matrix theory, provides a systematic way for classifying different types of homeostasis (homeostatic mechanisms) in input-output networks, in terms of the network topology. In turn, this leads to new mathematical concepts, such as, homeostasis subnetworks, homeostasis patterns, homeostasis mode interaction. We illustrate the usefulness of this theory with several biological examples: biochemical networks, chemical reaction networks (CRN), gene regulatory networks (GRN), Intracellular metal ion regulation and so on.

q-bio.MN↗

arXiv2024

Hopf Bifurcation in Asymmetric Ring Networks: Constraints on Phase Shifts

Hopf bifurcation in networks of coupled ODEs creates periodic states in which the relative phases of nodes are well defined near bifurcation. When the network is a fully inhomogeneous nearest-neighbour coupled unidirectional ring, and node spaces are 1-dimensional, we derive constraints on these phase shifts that apply to any ODE that respects the ring topology. We begin with a 3-node ring and generalise the results to any number of nodes. The main point is that such constraints exist even when the only structure present is the network topology. We also prove that the usual nondegeneracy conditions in the classical Hopf Bifurcation Theorem are valid generically for ring networks, by perturbing only coupling terms.

math.DS↗

arXiv2024

Hopf Bifurcation and Phase Patterns in Symmetric Ring Networks

Systems of ODEs coupled with the topology of a closed ring are common models in biology, robotics, electrical engineering, and many other areas of science. When the component systems and couplings are identical, the system has a cyclic symmetry group for unidirectional rings and a dihedral symmetry group for bidirectional rings. Hopf bifurcation in equivariant and network dynamics predicts the generic occurrence of periodic discrete rotating waves whose phase patterns are determined by the symmetry group. We review basic aspects of the theory in some detail and derive general properties of such rings. New results are obtained characterising the first bifurcation for long-range couplings and the direction in which discrete rotating wave states rotate.

math.DS↗

arXiv2024

Classification of 2-node Excitatory-Inhibitory Networks

We classify connected 2-node excitatory-inhibitory networks under various conditions. We assume that, as well as for connections, there are two distinct node-types, excitatory and inhibitory. In our classification we consider four different types of excitatory-inhibitory networks: restricted, partially restricted, unrestricted and completely unrestricted. For each type we give two different classifications. Using results on ODE-equivalence and minimality, we classify the ODE-classes and present a minimal representative for each ODE-class. We also classify all the networks with valence $\le 2$. These classifications are up to renumbering of nodes and the interchange of `excitatory' and `inhibitory' on nodes and arrows.These classifications constitute a first step towards analysing dynamics and bifurcations of excitatory-inhibitory networks. The results have potential applications to biological network models, especially neuronal networks, gene regulatory networks, and synthetic gene networks.

math.DS↗

arXiv2024

Whose wife is it anyway? Assessing bias against same-gender relationships in machine translation

Machine translation often suffers from biased data and algorithms that can lead to unacceptable errors in system output. While bias in gender norms has been investigated, less is known about whether MT systems encode bias about social relationships, e.g., "the lawyer kissed her wife." We investigate the degree of bias against same-gender relationships in MT systems, using generated template sentences drawn from several noun-gender languages (e.g., Spanish) and comprised of popular occupation nouns. We find that three popular MT services consistently fail to accurately translate sentences concerning relationships between entities of the same gender. The error rate varies considerably based on the context, and same-gender sentences referencing high female-representation occupations are translated with lower accuracy. We provide this work as a case study in the evaluation of intrinsic bias in NLP systems with respect to social relationships.

cs.CL↗