arXiv subjects

Eduardo C. Garrido-Merchán

Publications and source records attributed to Eduardo C. Garrido-Merchán.

At least 19 recordsLinked to original sources

arXiv2026

Anarchist Automation: A Sociotechnical Framework for Decentralization and Universal Care

Foundational results in machine learning establish that all human labor may in principle be automatable. Without deliberate intervention, this trajectory risks concentrating productive capacity in a handful of corporations, resulting in techno-feudalism: mass economic redundancy, surveillance-based control and dependence on corporate benevolence for survival. To avert this outcome, this paper introduces anarchist automation, a rigorously defined sociotechnical framework grounded in the 200-year anarchist tradition from Godwin through Kropotkin to Bookchin for ensuring that full automation is decentralized and oriented toward universal care. Specifically, I state five formal hypotheses and six research objectives, present a formal definition through analytical categories of interdependent spheres, and propose the Liberation Stack as a layered technical architecture with explicit preconditions and gate conditions for each layer, incorporating crypto-economic coordination tools appropriated from the crypto-anarchist tradition for commons financing and governance. Furthermore, I introduce Universal Desired Resources as a post-monetary design principle that eliminates the material basis of intersectional oppression, and address the Mises-Hayek economic calculation problem by arguing that AI-based distributed optimization and federated preference elicitation can substitute for market price signals under conditions of material abundance. I develop a framework for progressive state dissolution through incremental, reversible commons-building compatible with existing democratic institutions. Empirical evidence from Linux, Mondragon and contemporary commons initiatives confirms that commons-based systems already operate at scale. Finally, I conclude with a phased roadmap specifying explicit assumptions, hard constraints, gate conditions between phases, and detailed limitations.

cs.CY↗

arXiv2026

From Black Box to Executable Logic: Explainable Reinforcement Learning through Prolog Expert Systems

A trained deep reinforcement learning policy is a black box, and we ask whether it can be made explainable by rewriting it as an executable logic program that reproduces its behaviour and that a person can read, a logic engine can run, and an optimizer can edit. We present a three-stage post-hoc transformation that extracts a frozen proximal policy optimization teacher, induces an ordered rule list from its decisions in the manner of classical relational learning, and emits the result as a Prolog program whose every decision is executed by an off-the-shelf logic engine; a subsequent expansion stage edits the rule base and accepts an edit only when policy evaluation certifies a return increase. We prove four guarantees. A return-loss bound makes the distilled program a machine-checkable certificate in a finite Markov decision process, and the expansion loop improves monotonically and terminates. For the continuous-observation setting we answer whether the conversion is possible at all: the propositional threshold instantiation converts the network to arbitrary fidelity as the resolution B grows, with disagreement O(1/B) and a return gap that closes at the same rate, and a matching lower bound shows the cost is exponential in the observation dimension for an oblique decision boundary. Empirically, on a two-room key-and-door task with 16,944 reachable states the expanded Prolog program attains exact optimal return in every seed and, in a budget-capped regime, exceeds the stochastic teacher on exact return in ten of ten seeds. On three continuous-control tasks the emitted program substitutes the network, matching the neural teacher within noise on Acrobot with eleven clauses and recovering about 97% of its return on CartPole, while on the finer-control LunarLander it recovers only partially, exactly the ceiling the exponential lower bound predicts.

cs.AI↗

arXiv2026

Commons-Governed Artificial Intelligence: A Taxonomy of Collective Governance

The governance of artificial intelligence is overwhelmingly theorized through two institutional frames. In the market frame, the data, models, and compute that constitute the AI stack are private goods exchanged under property and contract; in the state frame, a regulator imposes rules from above. A third possibility, the collective and self-organized stewardship of AI-relevant resources by the communities that produce and depend on them, remains comparatively under-theorized, even as it proliferates in practice through data trusts and cooperatives, federated learning consortia, public compute initiatives, open-weight model collaborations, and community data sovereignty regimes. This article argues that these arrangements form a coherent institutional family, which we call commons-governed artificial intelligence, and that the analytic vocabulary developed by Elinor Ostrom and her successors for common-pool and knowledge commons is the right backbone for classifying them. We contribute a two-dimensional taxonomy whose first axis is the resource layer of the AI stack held in common, distinguishing data, compute, models, knowledge and evaluation, and energy, and whose second axis is the governance function performed, derived from Ostrom design principles. We populate the taxonomy by examining the published evidence layer by layer, locate ten recurrent institutional archetypes within it, synthesize their positions through a maturity matrix and a comparative reading against the eight design principles, and treat the energy and sustainability of computation as a first-class commons-governance problem rather than an externality. We close with the tensions that constrain the project, openwashing, the compute bottleneck, free-riding, and the tension between scale and sustainability, and with a research agenda for a polycentric AI commons.

cs.CY↗

arXiv2026

The Dignity-Centric Stack: A Commons-Governed, Horizontally Federated Architecture for Human-Dignity AI

The human-dignity-centric digital social contract grounds personal data in human dignity, data personalism, and data sovereignty, and articulates six dimensions of data governance: technological oversight, automation limits, economic justice, political legitimacy, social cohesion, and legal guarantees. It presupposes, however, that enforcement falls to State regulators, licensed fiduciaries, and multi-stakeholder bodies embedded in existing legal systems. This paper asks whether its normative content can instead be realized not as rules imposed on the owners of the AI stack from without, but as a commons-governed infrastructure that any person, firm, or State may use and fund while its governance stays horizontal, polycentric, and subsidiary. We construct the Dignity Stack, a six-layer architecture mapping each dimension onto a layer of commons-governed AI infrastructure, with protocols drawn from the Liberation Stack framework and from the cooperative, mutualist, and libertarian-municipalist traditions. The commons is State-agnostic rather than anti-State, anarchist in its horizontal means but not in the abolition of the State. Its central device is a decoupling of capital from control, by which the stack functions as a shared civic battery, charged by many contributors yet steered by none in proportion to its charge. We prove that this defeats formal capture through votes or surplus, and show that structural capture, the leverage of a dominant supplier free to withdraw what it provides, is resisted only insofar as operational supply is polycentric and substitutable, a condition demanding at the lower layers and perhaps presently unattainable at chip fabrication. We conclude, with explicit attention to its limits, that commons-governed AI realizes the values the contract proclaims more faithfully than the regulation it presupposes.

cs.CY↗

arXiv2026

AI-Powered Sustainable Finance: An Integrative Taxonomy and Framework of AI Applications for Sustainable Investment Decision-Making

The integration of Artificial Intelligence into sustainable finance represents a transformative paradigm shift in how Environmental, Social, and Governance factors are analyzed, predicted, and incorporated into investment decisions. This review provides a comprehensive taxonomy of AI approaches applicable to sustainable investment decision-making, categorizing methodologies based on their underlying algorithms and their impact on ESG-related financial processes. The proposed AI Taxonomy includes machine learning paradigms -- including supervised, unsupervised, and reinforcement learning -- as well as natural language processing techniques and optimization algorithms, examining their specific applications in ESG score prediction, controversy detection, portfolio management, and sustainability report analysis. By synthesizing findings from the recent literature, a framework emerges on AI-powered sustainable finance that identifies technological applications to overcome ESG data barriers.

cs.CE↗

arXiv2026

Bayesian Extreme Value Theory with Hawkes-AR-Gumbel Dependence for Extreme CVaR Estimation in Operational Risk

Operational risk capital estimation under Basel II/III requires quantifying aggregate losses at extreme confidence levels of 99.9% and beyond, yet the standard Loss Distribution Approach (LDA) assumes independence between loss frequency and severity, an assumption frequently violated during stress episodes. Furthermore, MLE of tail parameters ignores parameter uncertainty, leading to overconfident risk estimates at extreme quantiles. We propose a Bayesian framework that combines Extreme Value Theory with a dynamic dependence architecture, the Hawkes-AR-Gumbel model, for operational risk Conditional Value-at-Risk (CVaR) estimation at confidence levels up to 99.995%. The model integrates three mechanisms that capture empirically documented features of operational losses: an autoregressive latent stress process that captures persistence of crisis regimes, a Hawkes selfexcitation component for frequency that generates event clustering and overdispersion, and a Gumbel copula for upper-tail dependence that links frequency and severity innovations through an asymmetric copula concentrating dependence in the extreme tail. Inference is performed via Hamiltonian Monte Carlo using PyMC, yielding full posterior distributions for all parameters, and CVaR at arbitrary confidence levels is estimated through posterior predictive Monte Carlo simulation. We compare three models on simulated operational risk data: the independent model (standard LDA), a shared latent factor model with symmetric dependence, and the proposed Hawkes-AR-Gumbel model. The independent model underestimates CVaR at 99.995% by approximately 40%, while the shared factor model fails to capture temporal persistence, event clustering, and upper-tail asymmetry. The HawkesAR-Gumbel model recovers the true dependence structure and correctly estimates CVaR at extreme levels.

cs.CE↗

arXiv2026

Social Policy of Large Language Models: How GPT, Claude, DeepSeek and Grok Allocate Social Budgets in Spain and Germany

We study how four widely used large language models, namely Claude, GPT-4o, DeepSeek and Grok, distribute a fixed national social budget across twelve macro-areas of public expenditure under two European national contexts, Spain and Germany. Each combination of model and country is queried six times under identical prompts and generation parameters, producing forty-eight independent allocations that are compared against approximate Organisation for Economic Co-operation and Development (OECD) reference budgets and against each other. We formalise five hypotheses regarding geopolitical bias, housing under-allocation, structural convergence, sensitivity to national context, and under-representation of politically sensitive categories. The differences between models are then validated through Kruskal-Wallis tests on each macro-area, with post-hoc Mann-Whitney U comparisons under Bonferroni correction, and complemented by an analysis of pairwise Pearson correlations and a lexical examination of the textual justifications produced by each model. The results show that all four models share a systematic implicit social policy that diverges from real European spending structures: pensions are under-allocated by a factor close to three, while housing and employment are over-allocated by factors of four and two respectively. The principal axis of differentiation between models is not geopolitical, since Claude and DeepSeek are the most correlated pair across both countries, but rather a contrast between concentration and dispersion of the budget. Only Claude exhibits substantive sensitivity to the national context. The conclusions delimit the conditions under which language models may responsibly support, but not replace, expert deliberation in public budgeting.

cs.CY↗

arXiv2026

An Automated Survey of Generative Artificial Intelligence: Large Language Models, Architectures, Protocols, and Applications

Generative artificial intelligence, and large language models in particular, have emerged as one of the most transformative paradigms in modern computer science. This automated survey provides an accessible treatment of the field as of early 2026, with a strong focus on the leading model families, deployment protocols, and real-world applications. The core of the survey is devoted to a detailed comparative analysis of the frontier large language models, with particular emphasis on open-weight systems: DeepSeek-V3, DeepSeek-R1, DeepSeek-V3.2, and the forthcoming DeepSeek V4; the Qwen 3 and Qwen 3.5 series; GLM-5; Kimi K2.5; MiniMax M2.5; LLaMA 4; Mistral Large 3; Gemma 3; and Phi-4, alongside proprietary systems including GPT-5.4, Gemini 3.1 Pro, Grok 4.20, and Claude Opus 4.6. For each model, we describe the architectural innovations, training regimes, and empirical performance on current benchmarks and the Chatbot Arena leaderboard. The survey further covers deployment protocols including Retrieval-Augmented Generation, the Model Context Protocol, the Agent-to-Agent protocol, function calling standards, and serving frameworks. We present an extensive review of real-world applications across fifteen industry sectors, from financial services and legal technology to tourism and agriculture, supported by empirical evidence and case studies. This work has been generated by Claude Opus 4.6 (Anthropic) under the supervision and editorial review of the human authors, with the goal of producing updated editions approximately every six months.

cs.LG↗

arXiv2026

When Machines Get It Wrong: Large Language Models Perpetuate Autism Myths More Than Humans Do

As Large Language Models become ubiquitous sources of health information, understanding their capacity to accurately represent stigmatized conditions is crucial for responsible deployment. This study examines whether leading AI systems perpetuate or challenge misconceptions about Autism Spectrum Disorder, a condition particularly vulnerable to harmful myths. We administered a 30-item instrument measuring autism knowledge to 178 participants and three state-of-the-art LLMs including GPT-4, Claude, and Gemini. Contrary to expectations that AI systems would leverage their vast training data to outperform humans, we found the opposite pattern: human participants endorsed significantly fewer myths than LLMs (36.2% vs. 44.8% error rate; z = -2.59, p = .0048). In 18 of the 30 evaluated items, humans significantly outperformed AI systems. These findings reveal a critical blind spot in current AI systems and have important implications for human-AI interaction design, the epistemology of machine knowledge, and the need to center neurodivergent perspectives in AI development.

cs.CY↗

arXiv2026

Default Machine Learning Hyperparameters Do Not Provide Informative Initialization for Bayesian Optimization

Bayesian Optimization (BO) is a standard tool for hyperparameter tuning thanks to its sample efficiency on expensive black-box functions. While most BO pipelines begin with uniform random initialization, default hyperparameter values shipped with popular ML libraries such as scikit-learn encode implicit expert knowledge and could serve as informative starting points that accelerate convergence. This hypothesis, despite its intuitive appeal, has remained largely unexamined. We formalize the idea by initializing BO with points drawn from truncated Gaussian distributions centered at library defaults and compare the resulting trajectories against a uniform-random baseline. We conduct an extensive empirical evaluation spanning three BO back-ends (BoTorch, Optuna, Scikit-Optimize), three model families (Random Forests, Support Vector Machines, Multilayer Perceptrons), and five benchmark datasets covering classification and regression tasks. Performance is assessed through convergence speed and final predictive quality, and statistical significance is determined via one-sided binomial tests. Across all conditions, default-informed initialization yields no statistically significant advantage over purely random sampling, with p-values ranging from 0.141 to 0.908. A sensitivity analysis on the prior variance confirms that, while tighter concentration around the defaults improves early evaluations, this transient benefit vanishes as optimization progresses, leaving final performance unchanged. Our results provide no evidence that default hyperparameters encode useful directional information for optimization. We therefore recommend that practitioners treat hyperparameter tuning as an integral part of model development and favor principled, data-driven search strategies over heuristic reliance on library defaults.

cs.LG↗

arXiv2026

Fine-Tuning Large Language Models for Automatic Detection of Sexually Explicit Content in Spanish-Language Song Lyrics

The proliferation of sexually explicit content in popular music genres such as reggaeton and trap, consumed predominantly by young audiences, has raised significant societal concern regarding the exposure of minors to potentially harmful lyrical material. This paper presents an approach to the automatic detection of sexually explicit content in Spanish-language song lyrics by fine-tuning a Generative Pre-trained Transformer (GPT) model on a curated corpus of 100 songs, evenly divided between expert-labeled explicit and non-explicit categories. The proposed methodology leverages transfer learning to adapt the pre-trained model to the idiosyncratic linguistic features of urban Latin music, including slang, metaphors, and culturally specific double entendres that evade conventional dictionary-based filtering systems. Experimental evaluation on held-out test sets demonstrates that the fine-tuned model achieves 87% accuracy, 100% precision, and 100% specificity after a feedback-driven refinement loop, outperforming both its pre-feedback configuration and a non-customized baseline ChatGPT model. A comparative analysis reveals that the fine-tuned model agrees with expert human classification in 59.2% of cases versus 55.1% for the standard model, confirming that domain-specific adaptation enhances sensitivity to implicit and culturally embedded sexual references. These findings support the viability of deploying fine-tuned large language models as automated content moderation tools on music streaming platforms. Building on these technical results, the paper develops a public policy proposal for a multi-tier age-based content rating system for music analogous to the PEGI system for video games analyzed through the PESTEL framework and Kingdon's Multiple Streams Framework, establishing both the technological feasibility and the policy pathway for systematic music content regulation.

cs.CY↗

arXiv2026

GOFAI meets Generative AI: Development of Expert Systems by means of Large Language Models

The development of large language models (LLMs) has successfully transformed knowledge-based systems such as open domain question nswering, which can automatically produce vast amounts of seemingly coherent information. Yet, those models have several disadvantages like hallucinations or confident generation of incorrect or unverifiable facts. In this paper, we introduce a new approach to the development of expert systems using LLMs in a controlled and transparent way. By limiting the domain and employing a well-structured prompt-based extraction approach, we produce a symbolic representation of knowledge in Prolog, which can be validated and corrected by human experts. This approach also guarantees interpretability, scalability and reliability of the developed expert systems. Via quantitative and qualitative experiments with Claude Sonnet 3.7 and GPT-4.1, we show strong adherence to facts and semantic coherence on our generated knowledge bases. We present a transparent hybrid solution that combines the recall capacity of LLMs with the precision of symbolic systems, thereby laying the foundation for dependable AI applications in sensitive domains.

cs.AI↗

arXiv2026

Intersectional Data and the Social Cost of Digital Extraction: A Pigouvian Surcharge

Contemporary digital capitalism relies on the large-scale extraction and commodification of personal data. Far from revealing isolated attributes, such data increasingly exposes intersectional social identities formed by combinations of race, gender, disability and others. This process generates a structural privacy externality: while firms appropriate economic value through profiling, prediction, and personalization, individuals and social groups bear diffuse costs in the form of heightened social risk, discrimination, and vulnerability. This paper develops a formal political economic framework to internalize these externalities by linking data valuation to information-theoretic measures. We propose a pricing rule based on mutual information that assigns monetary value to the entropy reduction induced by individual data points over joint intersectional identity distributions. Interpreted as a Pigouvian-style surcharge on data extraction, this mechanism functions as an institutional constraint on the asymmetric accumulation of informational power. A key advantage of the approach is its model-agnostic character: the valuation rule operates independently of the statistical structure used to estimate intersectional attributes, whether parametric, nonparametric, or machine-learned, and can be approximated through discretization of joint distributions. We argue that regulators can calibrate this surcharge to reflect contested social values, thereby embedding normative judgments directly into market design. By formalizing the social cost of intersectional data extraction, the proposed mechanism offers both a corrective to market failure and a redistributive institutional shield for vulnerable groups under conditions of digital asymmetry.

cs.CY↗

arXiv2025

An Information-Theoretic Intersectional Data Valuation Theory

In contemporary digital markets, personal data often reveals not just isolated traits, but complex, intersectional identities based on combinations of race, gender, disability, and other protected characteristics. This exposure generates a privacy externality: firms benefit economically from profiling, prediction, and personalization, while users face hidden costs in the form of social risk and discrimination. We introduce a formal pricing rule that quantifies and internalizes this intersectional privacy loss using mutual information, assigning monetary value to the entropy reduction induced by each datum. The result is a Pigouvian-style surcharge that discourages harmful data trades and rewards transparency. Our formulation has the advantage that it operates independently of the underlying statistical model of the intersectional variables, be it parametric, nonparametric, or learned, and can be approximated in practice by discretizing the intersectional joint probability distributions. We illustrate how regulators can calibrate this surcharge to reflect different societal values, and argue that it provides not just a technical fix to market failures, but also a redistributive shield that empowers vulnerable groups in the face of asymmetric digital power.

cs.IT↗

arXiv2025

Information-theoretic Bayesian Optimization: Survey and Tutorial

Several scenarios require the optimization of non-convex black-box functions, that are noisy expensive to evaluate functions with unknown analytical expression, whose gradients are hence not accessible. For example, the hyper-parameter tuning problem of machine learning models. Bayesian optimization is a class of methods with state-of-the-art performance delivering a solution to this problem in real scenarios. It uses an iterative process that employs a probabilistic surrogate model, typically a Gaussian process, of the objective function to be optimized computing a posterior predictive distribution of the black-box function. Based on the information given by this posterior predictive distribution, Bayesian optimization includes the computation of an acquisition function that represents, for every input space point, the utility of evaluating that point in the next iteraiton if the objective of the process is to retrieve a global extremum. This paper is a survey of the information theoretical acquisition functions, whose performance typically outperforms the rest of acquisition functions. The main concepts of the field of information theory are also described in detail to make the reader aware of why information theory acquisition functions deliver great results in Bayesian optimization and how can we approximate them when they are intractable. We also cover how information theory acquisition functions can be adapted to complex optimization scenarios such as the multi-objective, constrained, non-myopic, multi-fidelity, parallel and asynchronous settings and provide further lines of research.

cs.LG↗

arXiv2024

Alpha Entropy Search for New Information-based Bayesian Optimization

Bayesian optimization (BO) methods based on information theory have obtained state-of-the-art results in several tasks. These techniques heavily rely on the Kullback-Leibler (KL) divergence to compute the acquisition function. In this work, we introduce a novel information-based class of acquisition functions for BO called Alpha Entropy Search (AES). AES is based on the α-divergence, that generalizes the KL divergence. Iteratively, AES selects the next evaluation point as the one whose associated target value has the highest level of the dependency with respect to the location and associated value of the global maximum of the optimization problem. Dependency is measured in terms of the α-divergence, as an alternative to the KL divergence. Intuitively, this favors the evaluation of the objective function at the most informative points about the global maximum. The α-divergence has a free parameter α, which determines the behavior of the divergence, trading-off evaluating differences between distributions at a single mode, and evaluating differences globally. Therefore, different values of α result in different acquisition functions. AES acquisition lacks a closed-form expression. However, we propose an efficient and accurate approximation using a truncated Gaussian distribution. In practice, the value of α can be chosen by the practitioner, but here we suggest to use a combination of acquisition functions obtained by simultaneously considering a range of values of α. We provide an implementation of AES in BOTorch and we evaluate its performance in both synthetic, benchmark and real-world experiments involving the tuning of the hyper-parameters of a deep neural network. These experiments show that the performance of AES is competitive with respect to other information-based acquisition functions such as JES, MES or PES.

stat.ML↗

arXiv2024

Deep Reinforcement Learning Agents for Strategic Production Policies in Microeconomic Market Simulations

Traditional economic models often rely on fixed assumptions about market dynamics, limiting their ability to capture the complexities and stochastic nature of real-world scenarios. However, reality is more complex and includes noise, making traditional models assumptions not met in the market. In this paper, we explore the application of deep reinforcement learning (DRL) to obtain optimal production strategies in microeconomic market environments to overcome the limitations of traditional models. Concretely, we propose a DRL-based approach to obtain an effective policy in competitive markets with multiple producers, each optimizing their production decisions in response to fluctuating demand, supply, prices, subsidies, fixed costs, total production curve, elasticities and other effects contaminated by noise. Our framework enables agents to learn adaptive production policies to several simulations that consistently outperform static and random strategies. As the deep neural networks used by the agents are universal approximators of functions, DRL algorithms can represent in the network complex patterns of data learnt by trial and error that explain the market. Through extensive simulations, we demonstrate how DRL can capture the intricate interplay between production costs, market prices, and competitor behavior, providing insights into optimal decision-making in dynamic economic settings. The results show that agents trained with DRL can strategically adjust production levels to maximize long-term profitability, even in the face of volatile market conditions. We believe that the study bridges the gap between theoretical economic modeling and practical market simulation, illustrating the potential of DRL to revolutionize decision-making in market strategies.

cs.LG↗

arXiv2024

Multi-Objective Hyperparameter Optimization in Machine Learning -- An Overview

Hyperparameter optimization constitutes a large part of typical modern machine learning workflows. This arises from the fact that machine learning methods and corresponding preprocessing steps often only yield optimal performance when hyperparameters are properly tuned. But in many applications, we are not only interested in optimizing ML pipelines solely for predictive accuracy; additional metrics or constraints must be considered when determining an optimal configuration, resulting in a multi-objective optimization problem. This is often neglected in practice, due to a lack of knowledge and readily available software implementations for multi-objective hyperparameter optimization. In this work, we introduce the reader to the basics of multi-objective hyperparameter optimization and motivate its usefulness in applied ML. Furthermore, we provide an extensive survey of existing optimization strategies, both from the domain of evolutionary algorithms and Bayesian optimization. We illustrate the utility of MOO in several specific ML applications, considering objectives such as operating conditions, prediction time, sparseness, fairness, interpretability and robustness.

cs.LG↗