Health data are distributed across organisations, while care, research, and artificial intelligence (AI) increasingly depend on using data across settings and populations. Health systems have often responded by concentrating data in shared environments. This works where common custody is lawful, trusted, and practical, but not where data, models, or services cannot or should not move under common control. Distributed access, analysis, model training, and exchange are already possible, but technical interoperability alone does not establish how independently governed organisations collaborate. Collaboration also requires agreement on purpose, participants and resources, authority and permissions, responsibilities and risks, governance, and incentives, which we call the relational configuration. When collaborations change, parts of it must be established again. The internet provides a precedent for making coordination more reusable. Before common protocols such as TCP/IP, capable but independently controlled networks remained disconnected; the “Islands of Networking” problem. Standardising only what had to cross their boundaries allowed them to interoperate while preserving local autonomy. Health faces an analogous “Islands of Data and AI” problem, with the additional requirement that participation across organisational boundaries must be governed. Federated Computing offers an architectural framework for governing participation across independently controlled organisations under sovereignty constraints. Relevant elements of an agreed relational configuration can be represented as verifiable capabilities and evaluated through admission when an operation crosses a boundary of independent authority producing evidence for accountability. Our Open Health demonstrator provides evidence of feasibility, not of health-system scalability or value. The organising reference for coordination should follow its purpose for whole-person and cross-sector care, the person can be the common reference for identifying relevant resources, while authority remains grounded in legitimate arrangements. We define the federated data advantage as the collective capacity created when complementary distributed resources can be used together under governed conditions without requiring common ownership. Whether this advantage can be realised in practice requires evidence that reusable coordination reduces the collaboration effort, remains governable across institutional boundaries, and creates value for participants and health systems.

A major opportunity for quantitative research lies in federated systems that turn distributed, institutionally governed data into reliable and verifiable research findings. As quantitative research moves beyond individual price-based signal models toward multimodal, multi-model, and alternative-data research, the opportunity is to coordinate analytics across organisational, contractual, and jurisdictional boundaries while preserving data locality, institutional autonomy, and regulatory compliance. This paper takes a federated systems view of this opportunity. Building on advances in predictive modelling and federated machine learning, it focuses on the broader architecture required for institution-grade quantitative research: who may request a task, which data and code are authorised, what derived output may leave a node, and how the resulting artefact can be reproduced and challenged. We synthesise literature on financial data assets, AI methods including agentic systems, multimodal and multi-model research, alternative data, data-collaboration architectures, trusted research environments, federated computing, governance, and machine-learning operations, and compare centralised, trusted, and federated designs as complementary responses to different operating conditions. To make the argument concrete, the paper presents an exemplar reference architecture that separates local collection, policy-aware orchestration, authorised execution, validation and aggregation, and research-facing delivery. Where valuable analysis must cross boundaries that raw data cannot, federated computing provides a foundational pattern for scalable, collaborative, and regulationaware quantitative research-moving computation to data, and elevating governance and evidence management to first-class system responsibilities.

This paper recommends how professional individuals and companies can respond to overwhelming technology innovation and "future proof" themselves. The greatest challenge for any business or individual is keeping up to date with the current technology innovation revolution (cf. "innovate or die"). Continually adapting, creating and exploring new ideas, methods, or products to remain relevant and successful, or risk becoming obsolete. The term, "creative destruction" is the process where new innovations and technologies replace old ones, leading to the decline and obsolescence of established industries, businesses, and jobs. Leading to whole industry sectors threatened, such as the European car industry, and a significant number of university graduates unable to find "entry-level" employment due to AI automation.

This paper presents a taxonomy and review of machine learning areas relevant to the Asset Allocator application domain. Asset Allocators include endowment funds, pension funds, and sovereign wealth funds, which typically have longer time horizons than other investor types and often socially important purposes. Allocators often fall under the radar in the financial data science literature, despite the large quantity of assets under management globally. Applying algorithms at the Allocator level has been a challenge historically due to limited datasets, long feedback loops, and a complex variety of underlying investment characteristics. The data challenge faced by Allocators is arguably one of the most complex problems within asset management. However, the recent period has seen growing interest in applying data science methods across all areas of the asset management industry, including within this domain. This paper discusses applications of machine learning and computational statistics algorithms to support investment analysis, portfolio management, and workflow productivity enhancement use cases. We outline key challenges and further examine private equity investing, from which there are a number of transferable learnings.

The objective of this paper is to provide a methodology for applying the DeTEcT framework to modelling token economies, to formalise the configuration of the simulation environment, and to introduce an event analysis framework. A token economy is an economic system that has a unique mechanism for controlling its monetary supply, and a medium, in the form of a token or currency, for the valuation of goods and services, the settlement of transactions, and the storage of value. We show the key decisions that must be made when modelling an economy with the DeTEcT framework and showcase some numerical methods that can be used in conjunction with the framework to perform economic simulations. We also propose a framework for analysing and measuring the impacts of events on an economy, while also developing a procedure to measure the significance of these impacts. Throughout the paper, we use Bitcoin as a case study to demonstrate how to apply the frameworks and tools we proposed here. We show how a model of Bitcoin token economy can be set up, and how to measure the impacts of Bitcoin's endogenous policies (i.e., BIPs) on the wealth distribution of its economic participants.

Embodied AI systems operate in the physical world, where failures can cause irreversible harm, yet safety research remains siloed across robotics, autonomous driving, and foundation-model communities. We survey 163 papers (2015-2026) through a dual taxonomy that cross-references 10 safety aspects with 8 embodied system types. Combining topic modeling, co-occurrence analysis, and a novel gap-score metric that quantifies under-explored areas relative to expected research coverage, we systematically map where effort concentrates and where critical blind spots persist. Safe reinforcement learning in simulation and alignment of vision-language-action (VLA) models are well studied, whereas alignment in navigation, robustness in multi-agent systems, and safe reinforcement learning for VLA models remain substantially neglected. Risk-weighted scoring, bootstrap confidence intervals, and keyword-sensitivity tests confirm the stability of these findings. We distill the results into a quantitative roadmap of research priorities to close the most consequential safety gaps in embodied AI. A comprehensive list of papers is available here: https://github.com/kleyt0n/awesome-safety-embodied-ai

The digital substrate - data, algorithms, infrastructure, platforms, applications - is being governed without adequate conceptual foundations. The ability and legitimacy required to govern this substrate, and to govern with it, are simultaneously misaligned, contested, and structurally absent. We introduce digital statecraft as the organising concept for this emerging field, arguing that 'digital' reconstitutes the statecraft question rather than merely extending its domain. The concept operates on two dimensions - statecraft over digital systems, concerning the authority and capacity of the state in relation to the digital substrate itself, and statecraft with digital systems, concerning the deployment of algorithmic tools as instruments of governing authority. And it rests on two foundational requirements, technical coherence and legitimate authority, that are genuinely in tension. We derive ten principles of digital statecraft from these foundations, each naming a condition whose absence produces an identifiable and structural governance failure: public interest first, human-machine complementarity, governability by design, systemic coherence, hybrid institutions, adaptive governance, human centricity and civic agency, accountable and traceable authority, judgment across time, and the non-delegable core. This article takes the state as the starting point, the institutional form that developed historically in response to the problem of effective and legitimate public governance, and the only current candidate for which the full set of legitimacy conditions is institutionally available. But the digital statecraft programme holds open a deeper question than just whether states can reform themselves: governing well in the algorithmic age may require rethinking the boundaries, scale, and affiliative basis of statehood itself.

The manipulation of the personality traits of large language models (LLMs) has emerged as a key area of research. Methods like prompt-based In-Context Knowledge Editing (IKE) and gradient-based Model Editor Networks (MEND) have been explored but show irregularity and variability; IKE depends on the prompt, leading to variability and sensitivity, while MEND yields inconsistent and gibberish outputs. To address this, we employed Opinion QA Based Parameter-Efficient Fine-Tuning (PEFT), specifically Quantized Low-Rank Adaptation (QLoRA), to manipulate the Big Five personality traits: Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. After PEFT, models such as Mistral-7B-Instruct and LLaMA-2-7B-chat began generating emojis, even though no emojis were present in the PEFT data. For instance, LLaMA-2-7B-chat generated emojis in 99.5 instances, while Mistral-7B-Instruct did so in 92.5 instances. ICL Explainability analysis indicated that the LLMs used emojis intentionally to express these traits. Mechanistic Interpretability analysis showed that this latent behaviour of LLMs could be traced to specific neurons that became activated or amplified after PEFT. This paper provides a number of novel contributions. First, introducing an Opinion QA dataset for PEFT-driven personality manipulation; second, developing metric models to benchmark LLM personality traits; third, demonstrating PEFT's superiority over IKE in personality manipulation; and finally, analysing and validating emoji usage through explainability methods such as Mechanistic Interpretability and In-context learning Explainability methods.

This paper proposes a framework for categorizing economic policies in a form of a tree taxonomy. The purpose of this approach is to construct an exhaustive and standardized list of actions that a governing authority has access to and can change to control an economy. This is advantageous from two perspectives: by having an exhaustive list of tools, it becomes easier to construct "complete" models (i.e., models that take in all empirical data and aim to simulate economic dynamics) of an economy and understand what the assumptions of these models are; and by knowing all available actions, economic strategies can be devised that target specific economic performance metrics with an exhaustive list of policies.

This paper presents a theoretical extension of the Decentralized Token Economy Theory (DeTEcT) framework proposed by Sadykhov et al. (Front. Blockchain, 2023, 6, 1298330), where a formal analysis framework was introduced for modelling wealth distribution in token economies. DeTEcT is a framework for analysing economic activity, simulating macroeconomic scenarios, and algorithmically setting policies in token economies. This paper proposes four ways of parametrizing the framework, where dynamic vs. static parametrization is considered along with the probabilistic vs. non-probabilistic parameters. Using these parametrization techniques, we demonstrate that by adding restrictions to the framework, it is possible to derive the existing wealth distribution models from DeTEcT. In addition to exploring parametrization techniques, this paper explores how money supply in the DeTEcT framework can be transformed to become dynamic and how this change will affect the dynamics of wealth distribution. The motivation for studying dynamic money supply is that it enables DeTEcT to be applied to modelling token economies without maximum supply (i.e., Ethereum) and it adds constraints to the framework in the form of symmetries.

As artificial intelligence transforms public sector operations, governments struggle to integrate technological innovations into coherent systems for effective service delivery. This paper introduces the Algorithmic State Architecture (ASA), a novel four-layer framework conceptualising how Digital Public Infrastructure, Data-for-Policy, Algorithmic Government/Governance, and GovTech interact as an integrated system in AI-enabled states. Unlike approaches that treat these as parallel developments, ASA positions them as interdependent layers with specific enabling relationships and feedback mechanisms. Through comparative analysis of implementations in Estonia, Singapore, India, and the UK, we demonstrate how foundational digital infrastructure enables systematic data collection, which powers algorithmic decision-making processes, ultimately manifesting in user-facing services. Our analysis reveals that successful implementations require balanced development across all layers, with particular attention to integration mechanisms between them. The framework contributes to both theory and practice by bridging previously disconnected domains of digital government research, identifying critical dependencies that influence implementation success, and providing a structured approach for analysing the maturity and development pathways of AI-enabled government systems.

As large language models transition to agentic systems, current safety evaluation frameworks face critical gaps in assessing deployment-specific risks. We introduce AgentSeer, an observability-based evaluation framework that decomposes agentic executions into granular action and component graphs, enabling systematic agentic-situational assessment. Through cross-model validation on GPT-OSS-20B and Gemini-2.0-flash using HarmBench single turn and iterative refinement attacks, we demonstrate fundamental differences between model-level and agentic-level vulnerability profiles. Model-level evaluation reveals baseline differences: GPT-OSS-20B (39.47

As the industry increasingly adopts agentic AI systems, understanding their unique vulnerabilities becomes critical. Prior research suggests that security flaws at the model level do not fully capture the risks present in agentic deployments, where models interact with tools and external environments. This paper investigates this gap by conducting a comparative red teaming analysis of GPT-OSS-20B, a 20-billion parameter open-source model. Using our observability framework AgentSeer to deconstruct agentic systems into granular actions and components, we apply iterative red teaming attacks with harmful objectives from HarmBench at two distinct levels: the standalone model and the model operating within an agentic loop. Our evaluation reveals fundamental differences between model level and agentic level vulnerability profiles. Critically, we discover the existence of agentic-only vulnerabilities, attack vectors that emerge exclusively within agentic execution contexts while remaining inert against standalone models. Agentic level iterative attacks successfully compromise objectives that completely failed at the model level, with tool-calling contexts showing 24% higher vulnerability than non-tool contexts. Conversely, certain model-specific exploits work exclusively at the model level and fail when transferred to agentic contexts, demonstrating that standalone model vulnerabilities do not always generalize to deployed systems.

We study the optimal Market Making problem in a Limit Order Book (LOB) market simulated using a high-fidelity, mutually exciting Hawkes process. Departing from traditional Brownian-driven mid-price models, our setup captures key microstructural properties such as queue dynamics, inter-arrival clustering, and endogenous price impact. Recognizing the realistic constraint that market makers cannot update strategies at every LOB event, we formulate the control problem within an impulse control framework, where interventions occur discretely via limit, cancel, or market orders. This leads to a high-dimensional, non-local Hamilton-Jacobi-Bellman Quasi-Variational Inequality (HJB-QVI), whose solution is analytically intractable and computationally expensive due to the curse of dimensionality. To address this, we propose a novel Reinforcement Learning (RL) approximation inspired by auxiliary control formulations. Using a two-network PPO-based architecture with self-imitation learning, we demonstrate strong empirical performance with limited training, achieving Sharpe ratios above 30 in a realistic simulated LOB. In addition to that, we solve the HJB-QVI using a deep learning method inspired by Sirignano and Spiliopoulos 2018 and compare the performance with the RL agent. Our findings highlight the promise of combining impulse control theory with modern deep RL to tackle optimal execution problems in jump-driven microstructural markets.

Large language models increasingly rely on synthetic data due to human-written content scarcity, yet recursive training on model-generated outputs leads to model collapse, a degenerative process threatening factual reliability. We define knowledge collapse as a distinct three-stage phenomenon where factual accuracy deteriorates while surface fluency persists, creating "confidently wrong" outputs that pose critical risks in accuracy-dependent domains. Through controlled experiments with recursive synthetic training, we demonstrate that collapse trajectory and timing depend critically on instruction format, distinguishing instruction-following collapse from traditional model collapse through its conditional, prompt-dependent nature. We propose domain-specific synthetic training as a targeted mitigation strategy that achieves substantial improvements in collapse resistance while maintaining computational efficiency. Our evaluation framework combines model-centric indicators with task-centric metrics to detect distinct degradation phases, enabling reproducible assessment of epistemic deterioration across different language models. These findings provide both theoretical insights into collapse dynamics and practical guidance for sustainable AI training in knowledge-intensive applications where accuracy is paramount.

Large Language Model (LLM) alignment conventionally relies on supervised fine-tuning or reinforcement learning based alignment frameworks. These methods typically require labeled or preference datasets and involve updating model weights to align the LLM with the training objective or reward model. Meanwhile, in social sciences such as cross-cultural studies, factor analysis is widely used to uncover underlying dimensions or latent variables that explain observed patterns in survey data. The non-differentiable nature of these measurements deriving from survey data renders the former alignment methods infeasible for alignment with cultural dimensions. To overcome this, we propose a parameter efficient strategy that combines soft prompt tuning, which freezes the model parameters while modifying the input prompt embeddings, with Differential Evolution (DE), a black-box optimization method for cases where a differentiable objective is unattainable. This strategy ensures alignment consistency without the need for preference data or model parameter updates, significantly enhancing efficiency and mitigating overfitting. Our method demonstrates significant improvements in LLama-3-8B-Instruct's cultural dimensions across multiple regions, outperforming both the Naive LLM and the In-context Learning (ICL) baseline, and effectively bridges computational models with human cultural nuances.