Showing 1–50 of 287 results for author: Machado, M

Systematic pathway comparison on the powerset of rule-based biochemical systems

Authors: Anne-Susann Abel, Sissel Banke, Erika M. Herrera Machado, Jakob Lykke Andersen, Peter Dittrich, Rolf Fagerberg, Daniel Merkle

Abstract: Computational pathway design often focuses on evaluating selected pathways or optimizing fluxes in a fixed network, but gives less direct access to the combinatorial question of which other enzyme subsets of the network can support productive alternative pathways. A structured computational analysis of these networks can act as a valuable pre-step to the pathway design process. We present here a s… ▽ More Computational pathway design often focuses on evaluating selected pathways or optimizing fluxes in a fixed network, but gives less direct access to the combinatorial question of which other enzyme subsets of the network can support productive alternative pathways. A structured computational analysis of these networks can act as a valuable pre-step to the pathway design process. We present here a systematic approach for exploring biochemical pathway alternatives across enzyme subsets, using a computational methodology based on a rule-based modeling of the enzymes: for a biochemical system with enzyme set $S$, we evaluate all subsets $s \subseteq S$ by generating chemical reaction spaces, searching for integer-hyperflow pathways from prescribed inputs to target products, and organizing feasible subsets by set inclusion. This yields an inclusion-ordered landscape of pathway feasibility and carbon efficiency. We apply the approach to the non-oxidative pentose phosphate pathway, to non-oxidative glycolysis, and to glycolysis. Across these systems, feasible subsets occupy only a moderate fraction of all enzyme subsets, but the structure of this feasible region differs strongly between the systems. Larger enzyme sets do not consistently improve carbon efficiency when every enzyme in the tested subset is required to participate in the pathway. Instead, performance depends on specific enzyme combinations. The resulting subset landscapes are valuable means for identifying essential enzymes, candidate redundancies, and small high-performing enzyme subsets. By making the enzyme-subset landscape itself the object of analysis, the approach addresses the gap between detailed evaluation of individual candidate pathways and early-stage design decisions about which enzyme combinations are worth investigating at all. △ Less

Submitted 24 August, 2026; originally announced August 2026.

High Precision Fundamental Physics Experiments at JLab with Spin-transparent Storage Rings of Low-energy Polarized Electron Beams

Authors: Vladimir Khachatryan, Riad Suleiman, Alberto Accardi, Carlos Ayerbe Gayoso, Marco Battaglieri, Devesh Bhattarai, Silviu Covrig Dusa, Yaroslav Derbenev, Forrest Friesen, Joseph Grames, Paul Guèye, Tyler J. Hague, Ching Him Leung, Wenliang, Li, Magno V. T. Machado, Preet Mann, Vasiliy Morozov, Son Nguyen, Michael Nycz, Howard Oh, Udit Raha, Mario Reig, Marco Schreck, Nathaniel Sherrill , et al. (5 additional authors not shown)

Abstract: A breakthrough in fundamental physics experiments measuring particle spin precession may happen if spin-transparent storage rings become adopted tools for such experiments. We present a new design of highly specialized table-sized storage rings, which use low-energy polarized electron beams and Mott polarimetry. Based on the spin transparency ansatz, the spin precession stemming from the magnetic… ▽ More A breakthrough in fundamental physics experiments measuring particle spin precession may happen if spin-transparent storage rings become adopted tools for such experiments. We present a new design of highly specialized table-sized storage rings, which use low-energy polarized electron beams and Mott polarimetry. Based on the spin transparency ansatz, the spin precession stemming from the magnetic dipole moment is canceled at any beam energy after an electron's turn along the periodic orbit in the ring. Meanwhile, a spin precession induced by the fundamental physics of interest, e.g., the electron's permanent electric dipole moment (EDM) and/or ultralight-dark-matter-mediated forces such as axions, will accumulate. However, capitalizing on such types of rings is not only desirable for measurements of EDMs and axion searches relevant to $CP$ violation and matter-antimatter asymmetry in the Universe, but may also find very promising applications in quantum computing. △ Less

Submitted 9 August, 2026; originally announced August 2026.

Comments: 30 pages, 5 figures

Report number: JLab LOI12-26-003

HASTE: A Platform for Rapid Post-Disaster Building Damage Assessment

Authors: Caleb Robinson, Anthony Ortiz, Simone Fobi Nsutezo, Cameron Birge, Meygha Machado, Marcelo Duarte, Joaquin Rivero Rodriguez, Anthony Cintron Roman, Kevin White, Inbal Becker-Reshef, Juan M. Lavista Ferres

Abstract: When a large disaster strikes, responders need a map of which buildings are damaged within hours. The models that do well on public benchmarks assume matched before-and-after imagery and a training set drawn from similar past events, and neither is usually available for a new disaster in its first day. We present HASTE (High-speed Assessment and Satellite Tracking for Emergencies), a no-code web p… ▽ More When a large disaster strikes, responders need a map of which buildings are damaged within hours. The models that do well on public benchmarks assume matched before-and-after imagery and a training set drawn from similar past events, and neither is usually available for a new disaster in its first day. We present HASTE (High-speed Assessment and Satellite Tracking for Emergencies), a no-code web platform that lets analysts who are not machine learning engineers produce per-building damage maps from post-disaster satellite imagery. HASTE implements two methods that share one interface. The first requires the user to label polygons over the post-disaster scene, trains a small semantic segmentation model on that single scene, runs it over the whole image, and joins the per-pixel output to existing building footprints. The second embeds every footprint with a pretrained vision model, requires the user to label a handful of buildings, and fits a logistic regression in the browser that scores the rest of the scene in seconds. We describe the platform, both methods, and the engineering that supports them. We also report preliminary experiments on xBD showing that foundation-model embeddings pooled over footprints separate damaged from intact buildings using post-disaster imagery alone, matching a fully supervised ResNet-50 baseline with a twentieth of its labels. HASTE and its predecessors have supported more than thirty real-world disaster responses since 2023, spanning earthquakes, hurricanes, cyclones, floods, wildfires, and tornadoes, delivering results to humanitarian partners within hours to days of imagery becoming available. We close with the directions we think are most promising, including vision-language assessment, active learning, and damage models for roads and other infrastructure. HASTE is open source at https://github.com/microsoft/haste. △ Less

Submitted 13 July, 2026; originally announced July 2026.

Accelerating Q-learning through Efficient Value-Sharing across Actions

Authors: Prabhat Nagarajan, Brett Daley, Martha White, Marlos C. Machado

Abstract: Action values are foundational to many control algorithms such as Q-learning. Therefore, efficient action-value learning is central to reinforcement learning (RL). However, learning them can be slow, requiring many updates to move values from their initialization, typically near zero, to their true values, which may be far from zero. Moreover, action-value learning algorithms typically update each… ▽ More Action values are foundational to many control algorithms such as Q-learning. Therefore, efficient action-value learning is central to reinforcement learning (RL). However, learning them can be slow, requiring many updates to move values from their initialization, typically near zero, to their true values, which may be far from zero. Moreover, action-value learning algorithms typically update each state-action pair independently, without learning a value that is common to all actions within a state. In this paper, we address these inefficiencies by introducing the mean-expansion layer, which accelerates action-value learning by sharing values across actions within a state and by changing the problem from directly learning potentially large action-values to learning a lower-norm representation of them. In deep RL, this layer can be applied as a parameter-free addition to Q-network architectures without altering the underlying algorithm. Applied to deep Q-networks and implicit quantile networks, it improves aggregate performance across 57 Atari 2600 games while increasing action gaps and dramatically reducing value overestimation. △ Less

Submitted 6 August, 2026; v1 submitted 29 June, 2026; originally announced June 2026.

Comments: ICML 2026 (Spotlight); Adaptive and Learning Agents workshop 2026 (Best paper runner-up)

Project SPARROW and the Future of Conservation Technology

Authors: Juan M. Lavista Ferres, Carl Chalmers, Bruno Demuro Segundo, Zhongqi Miao, Andres Hernandez Celis, Federico Alves Torres, Isai Daniel Chacon Silva, Anthony Cintron Roman, Allen Kim, Meygha Machado, Luana Marotti, Amy Michaels, Daniela Ruiz Lopez, Catherine Romero, Rahul Dodhia, Inbal Becker-Reshef, Pablo Arbelaez

Abstract: Global biodiversity is declining at unprecedented rates, yet the tools available to monitor and protect ecosystems remain limited by constraints in power, connectivity, and accessibility. We present SPARROW, a hardware and software open-source platform that integrates solar energy, edge artificial intelligence, and satellite communication to enable continuous, autonomous biodiversity monitoring in… ▽ More Global biodiversity is declining at unprecedented rates, yet the tools available to monitor and protect ecosystems remain limited by constraints in power, connectivity, and accessibility. We present SPARROW, a hardware and software open-source platform that integrates solar energy, edge artificial intelligence, and satellite communication to enable continuous, autonomous biodiversity monitoring in remote environments. Each SPARROW node combines a low-power Graphics Processing Unit (GPU) with modular visual, acoustic, and environmental sensors, performing on-device deep learning inference and transmitting summarized results through Low-Earth-Orbit (LEO) satellite or Global System for Mobile Communications (GSM) networks. We deployed SPARROW across tropical, temperate, and montane ecosystems in Colombia, Peru, Tanzania, and the United States, where it sustained 24/7 operation under variable environmental conditions and collected more than two million images and acoustic recordings in the first 190 days. The system demonstrated robust real-time classification and adaptive power management, achieving full autonomy without on-site human intervention. By integrating renewable energy, on-edge AI, and open-source design, SPARROW lowers the technical and financial barriers to ecological monitoring and establishes a scalable foundation for a distributed, intelligent network of sensors, an emerging "Internet of Living Things" for planetary biodiversity monitoring. △ Less

Submitted 26 May, 2026; originally announced June 2026.

Topology of Percolation Clusters: Central Limit Theorems beyond the Lattice

Authors: Luciano H. L. de Araújo, Daniel Miranda Machado, Cristian F. Coletti

Abstract: We prove central limit theorems (CLTs) for topological functionals of Bernoulli bond percolation on infinite graphs beyond the Euclidean lattice $\mathbb{Z}^{d}$. For quasi-transitive graphs of subexponential growth, we show that the number $K_{r}$ of open clusters intersecting the metric ball $B_{r}$ satisfies a CLT as $r\to\infty$. For amenable Cayley graphs, we prove a general CLT for stationar… ▽ More We prove central limit theorems (CLTs) for topological functionals of Bernoulli bond percolation on infinite graphs beyond the Euclidean lattice $\mathbb{Z}^{d}$. For quasi-transitive graphs of subexponential growth, we show that the number $K_{r}$ of open clusters intersecting the metric ball $B_{r}$ satisfies a CLT as $r\to\infty$. For amenable Cayley graphs, we prove a general CLT for stationary percolation functionals along Folner sequences under sequential stabilization and a finite-moment assumption, provided the group admits a left-orderable finite-index subgroup. This applies in particular to groups of polynomial growth. As an application, we obtain CLTs for Betti numbers of graph-generated random simplicial complexes, including clique and neighbor complexes. The proofs combine invariant edge orderings, martingale decompositions, and stabilization estimates for single-edge perturbations. △ Less

Submitted 8 April, 2026; originally announced April 2026.

MSC Class: 60K35 (Primary) 60F05; 05C80; 20F65; 60D05; 55N35 (Secondary)

Quantitative Stability for Minkowski's problem

Authors: Károly Böröczky, João Miguel Machado, João P. G. Ramos

Abstract: We derive quantitative stability results for Minkowski bodies, as well as their counterparts, the $L_p$-Minkowski bodies in the range $1 \le p \neq n$. We prove that, for every pair of probability measures $μ,ν$ satisfying a quantitative form of the classical dispersion assumptions yielding existence of such bodies, we have a control of the form \[ \inf_{x\in \mathbb{R}^n}\mathrm{d_H}(E_μ, x + E… ▽ More We derive quantitative stability results for Minkowski bodies, as well as their counterparts, the $L_p$-Minkowski bodies in the range $1 \le p \neq n$. We prove that, for every pair of probability measures $μ,ν$ satisfying a quantitative form of the classical dispersion assumptions yielding existence of such bodies, we have a control of the form \[ \inf_{x\in \mathbb{R}^n}\mathrm{d_H}(E_μ, x + E_ν) \le C \mathrm{d_C}(μ,ν)^{\frac{1}{n-1}}, \quad α(E_μ, E_ν)^2 \le C \mathrm{d_C}(μ,ν)^{1 + \frac{1}{n-1}}, \] where $\mathrm{d_H}$ denotes the Hausdorff distance, $α$ denotes the Fraenkel asymmetry and $\mathrm{d_C}$ is the dual-convex distance of probability measures on the sphere. Our arguments are based on a variational problem whose optimizers are Minkowski bodies, for which we can obtain strong-concavity properties with the quantitative Brunn-Minkowski and isoperimetric inequalities. While the exponent in the Hausdorff distance is sharp, the exponent in the Fraenkel asymmetry is optimal in dimension $2$. △ Less

Submitted 13 May, 2026; v1 submitted 18 March, 2026; originally announced March 2026.

MSC Class: 52A21; 52A40; 49Q10

Wasserstein Gradient Flows of semi-discret energies: evolution of urban areas anduniform quantization

Authors: Joao Miguel Machado

Abstract: We study the Wasserstein gradient flow of semi-discrete energies in the space of probability measures, that is functionals depending on two measures-one being an absolutely continuous density and the other an atomic measure. These energies appear naturally in the field of urban planning. This is done via the celebrated JKO scheme, for which we prove convergence to a limiting system composed of a p… ▽ More We study the Wasserstein gradient flow of semi-discrete energies in the space of probability measures, that is functionals depending on two measures-one being an absolutely continuous density and the other an atomic measure. These energies appear naturally in the field of urban planning. This is done via the celebrated JKO scheme, for which we prove convergence to a limiting system composed of a parabolic PDE with singular advection coupled with an ODE, also presenting singular dynamics. This is first done under more general assumptions using classical tools, and in a second moment convergence is proven to hold in $L^2_tH^1_x$ for the cases of linear and Porous-Medium type diffusions. We then pass to the study of some qualitative properties of this system, such as the convergence of the atoms towards the baricenters of their corresponding Laguerre cells. We finish this work with extensive numerical simulations that aid in formulating conjectures for the qualitative behavior of this system; in the case of linear diffusion, for instance, we observe a dynamic crystallization phenomenon. △ Less

Submitted 4 March, 2026; originally announced March 2026.

Sharkovskiis theorem under small random perturbations

Authors: Isabella Alvarenga, Daniel Miranda Machado

Abstract: We establish a Sharkovskii-type theorem for a class of discrete random dynamical systems via the random Conley index. Using the continuation property of the Conley index, we extend classical forcing results to random systems obtained from small random perturbations of one-dimensional maps. In contrast to earlier measure-theoretic results, which are typically subject to an inherent period-doubling… ▽ More We establish a Sharkovskii-type theorem for a class of discrete random dynamical systems via the random Conley index. Using the continuation property of the Conley index, we extend classical forcing results to random systems obtained from small random perturbations of one-dimensional maps. In contrast to earlier measure-theoretic results, which are typically subject to an inherent period-doubling ambiguity (realizing period $n$ or $2n$), our topological approach allows us to detect random periodic points and orbits with precise minimal periods. This yields realisation results for arbitrary finite tails of the Sharkovskii ordering. These results are illustrated by constructing random periodic orbits for perturbed versions of the tent map and the logistic map. △ Less

Submitted 12 February, 2026; originally announced February 2026.

MSC Class: 37H10; 37E05; 37E15; 47H04; 47H40

The Laplacian Keyboard: Beyond the Linear Span

Authors: Siddarth Chandrasekar, Marlos C. Machado

Abstract: Across scientific disciplines, Laplacian eigenvectors serve as a fundamental basis for simplifying complex systems, from signal processing to quantum mechanics. In reinforcement learning (RL), they similarly form a basis over the state space, enabling reward functions to be approximated by projection onto a small set of eigenvectors. This projection makes zero-shot control possible, but it also im… ▽ More Across scientific disciplines, Laplacian eigenvectors serve as a fundamental basis for simplifying complex systems, from signal processing to quantum mechanics. In reinforcement learning (RL), they similarly form a basis over the state space, enabling reward functions to be approximated by projection onto a small set of eigenvectors. This projection makes zero-shot control possible, but it also imposes a fundamental limitation: the induced policies are only as expressive as the linear span of the chosen eigenvectors. We introduce the Laplacian Keyboard (LK), a hierarchical framework that goes beyond this linear span. LK constructs a task-agnostic library of behaviors from these eigenvectors, forming a behavior basis guaranteed to contain the optimal policy for any reward within the linear span. A meta-policy learns to stitch these behaviors dynamically, enabling efficient learning of policies outside the original linear constraints. We establish theoretical bounds on zero-shot approximation error and demonstrate empirically that LK improves over the zero-shot solution while achieving better sample efficiency compared to standard RL methods. △ Less

Submitted 16 May, 2026; v1 submitted 7 February, 2026; originally announced February 2026.

Comments: 31 pages, 17 figures

Laplacian Representations for Decision-Time Planning

Authors: Dikshant Shehmar, Matthew Schlegel, Matthew E. Taylor, Marlos C. Machado

Abstract: Planning with a learned model remains a key challenge in model-based reinforcement learning (RL). In decision-time planning, state representations are critical as they must support local cost computation while preserving long-horizon structure. In this paper, we show that the Laplacian representation provides an effective latent space for planning by capturing state-space distances at multiple tim… ▽ More Planning with a learned model remains a key challenge in model-based reinforcement learning (RL). In decision-time planning, state representations are critical as they must support local cost computation while preserving long-horizon structure. In this paper, we show that the Laplacian representation provides an effective latent space for planning by capturing state-space distances at multiple time scales. This representation preserves meaningful distances and naturally decomposes long-horizon problems into subgoals, also mitigating the compounding errors that arise over long prediction horizons. Building on these properties, we introduce ALPS, a hierarchical planning algorithm, and demonstrate that it outperforms commonly used baselines on a selection of offline goal-conditioned RL tasks from OGBench, a benchmark previously dominated by model-free methods. △ Less

Submitted 2 June, 2026; v1 submitted 4 February, 2026; originally announced February 2026.

Comments: Accepted at ICML 2026

Journal ref: Proceedings of the 43rd International Conference on Machine Learning (ICML 2026)

DROGO: Default Representation Objective via Graph Optimization in Reinforcement Learning

Authors: Hon Tik Tse, Marlos C. Machado

Abstract: In computational reinforcement learning, the default representation (DR) and its principal eigenvector have been shown to be effective for a wide variety of applications, including reward shaping, count-based exploration, option discovery, and transfer. However, in prior investigations, the eigenvectors of the DR were computed by first approximating the DR matrix, and then performing an eigendecom… ▽ More In computational reinforcement learning, the default representation (DR) and its principal eigenvector have been shown to be effective for a wide variety of applications, including reward shaping, count-based exploration, option discovery, and transfer. However, in prior investigations, the eigenvectors of the DR were computed by first approximating the DR matrix, and then performing an eigendecomposition. This procedure is computationally expensive and does not scale to high-dimensional spaces. In this paper, we derive an objective for directly approximating the principal eigenvector of the DR with a neural network. We empirically demonstrate the effectiveness of the objective in a number of environments, and apply the learned eigenvectors for reward shaping. △ Less

Submitted 30 January, 2026; originally announced February 2026.

Universal computation is intrinsic to language model decoding

Authors: Alex Lewandowski, Marlos C. Machado, Dale Schuurmans

Abstract: Language models now provide an interface to express and often solve general problems in natural language, yet their ultimate computational capabilities remain a major topic of scientific debate. Unlike a formal computer, a language model is trained to autoregressively predict successive elements in human-generated text. We prove that chaining a language model's autoregressive output is sufficient… ▽ More Language models now provide an interface to express and often solve general problems in natural language, yet their ultimate computational capabilities remain a major topic of scientific debate. Unlike a formal computer, a language model is trained to autoregressively predict successive elements in human-generated text. We prove that chaining a language model's autoregressive output is sufficient to perform universal computation. That is, a language model can simulate the execution of any algorithm on any input. The challenge of eliciting desired computational behaviour can thus be reframed in terms of programmability: the ease of finding a suitable prompt. Strikingly, we demonstrate that even randomly initialized language models are capable of universal computation before training. This implies that training does not give rise to computational expressiveness -- rather, it improves programmability, enabling a natural language interface for accessing these intrinsic capabilities. △ Less

Submitted 10 February, 2026; v1 submitted 12 January, 2026; originally announced January 2026.

Comments: Minor formatting corrections

Symmetry Breaking, Hysteresis, and Convergence to the Mean Voter in two-party Spatial Competition

Authors: Daniel Miranda Machado, Roberto Venegeroles

Abstract: Classical spatial models of two-party competition typically predict convergence to the median voter, yet real-world party systems often exhibit persistent and asymmetric polarization. We develop a spatial model of two-party competition in which voters evaluate parties through general satisfaction functions, and a width parameter $q$ captures how tolerant they are of ideological distance. This para… ▽ More Classical spatial models of two-party competition typically predict convergence to the median voter, yet real-world party systems often exhibit persistent and asymmetric polarization. We develop a spatial model of two-party competition in which voters evaluate parties through general satisfaction functions, and a width parameter $q$ captures how tolerant they are of ideological distance. This parameter governs the balance between centripetal and centrifugal incentives and acts as the bifurcation parameter governing equilibrium configurations. Under mild regularity assumptions, we characterize Nash equilibria through center-distance coordinates, which separate the endogenous political center from polarization. When the voter density is symmetric, the reduced equilibrium condition exhibits a generic supercritical pitchfork bifurcation at a critical value $q_{c}$. Above $q_{c}$, the unique stable equilibrium features convergence to the center, recovering the classical median voter result, whereas below it two symmetric polarized equilibria arise. Asymmetry in the voter distribution unfolds the pitchfork, producing drift in the endogenous center and asymmetric polarized equilibria. The resulting equilibrium diagram has an S-shaped geometry that generates hysteresis, allowing polarization to persist even after tolerance returns to levels that would support convergence in a symmetric environment. In the high-tolerance regime, we show that the unique non-polarized equilibrium converges to the mean of the voter distribution, while the median is recovered only under symmetry. Hence, unlike the Hotelling--Downs model, where convergence to the median is universal, the median voter appears here as an asymptotic benchmark rather than a robust predictor. △ Less

Submitted 11 January, 2026; originally announced January 2026.

Comments: 28 pages, 8 figure

The World Is Bigger! A Computationally-Embedded Perspective on the Big World Hypothesis

Authors: Alex Lewandowski, Adtiya A. Ramesh, Edan Meyer, Dale Schuurmans, Marlos C. Machado

Abstract: Continual learning is often motivated by the idea, known as the big world hypothesis, that "the world is bigger" than the agent. Recent problem formulations capture this idea by explicitly constraining an agent relative to the environment. These constraints lead to solutions in which the agent continually adapts to best use its limited capacity, rather than converging to a fixed solution. However,… ▽ More Continual learning is often motivated by the idea, known as the big world hypothesis, that "the world is bigger" than the agent. Recent problem formulations capture this idea by explicitly constraining an agent relative to the environment. These constraints lead to solutions in which the agent continually adapts to best use its limited capacity, rather than converging to a fixed solution. However, explicit constraints can be ad hoc, difficult to incorporate, and may limit the effectiveness of scaling up the agent's capacity. In this paper, we characterize a problem setting in which an agent, regardless of its capacity, is constrained by being embedded in the environment. In particular, we introduce a computationally-embedded perspective that represents an embedded agent as an automaton simulated within a universal (formal) computer. Such an automaton is always constrained; we prove that it is equivalent to an agent that interacts with a partially observable Markov decision process over a countably infinite state-space. We propose an objective for this setting, which we call interactivity, that measures an agent's ability to continually adapt its behaviour by learning new predictions. We then develop a model-based reinforcement learning algorithm for interactivity-seeking, and use it to construct a synthetic problem to evaluate continual learning capability. Our results show that deep nonlinear networks struggle to sustain interactivity, whereas deep linear networks sustain higher interactivity as capacity increases. △ Less

Submitted 29 December, 2025; originally announced December 2025.

Comments: NeurIPS 2025 (spotlight)

Quantitative stability for the Brascamp-Lieb inequality and moment measures

Authors: João Miguel Machado, João P. G. Ramos

Abstract: We develop a quantitative stability theory for moment measures based on a new sharp uniform stability principle for the Brascamp-Lieb variance inequality in terms of the $L^1$-distance. Our results yield structural stability estimates for solutions of the moment-measure problem that are uniform over a natural class of convex functions, thereby addressing several questions that have been open in th… ▽ More We develop a quantitative stability theory for moment measures based on a new sharp uniform stability principle for the Brascamp-Lieb variance inequality in terms of the $L^1$-distance. Our results yield structural stability estimates for solutions of the moment-measure problem that are uniform over a natural class of convex functions, thereby addressing several questions that have been open in this direction. A key novelty of our approach is that the Brascamp-Lieb stability bound is not only sharp in its stability exponent, but also uniform across a broad class of convex potentials. This uniformity is absent from previous results in the literature and, beyond its intrinsic mathematical interest, it is the mechanism that allows stability of the Brascamp-Lieb inequality to transfer to nonlinear variational problems such as the moment-measure problem. We moreover show that the $L^1-$nature of this stability estimate is sharp, in the sense that such a uniform estimate cannot hold in any $L^p$-metric, with $p>1$. △ Less

Submitted 28 July, 2026; v1 submitted 27 November, 2025; originally announced November 2025.

MSC Class: 39B82; 26D10; 49Q22

The Maki-Thompson Model with Spontaneous Stifling on Symmetric Networks

Authors: Nancy Lopes Garcia, Denis Araujo Luiz, Daniel Miranda Machado

Abstract: We investigate rumor spreading in a generalized Maki-Thompson model with spontaneous stifling, evolving on quasi-transitive networks. Individuals are either ignorants, spreaders, or stiflers; spreaders stop by contact with other spreaders or stiflers or after an independent random waiting time sampled from a given distribution, modeling a spontaneous loss of interest. The topology of the underlyin… ▽ More We investigate rumor spreading in a generalized Maki-Thompson model with spontaneous stifling, evolving on quasi-transitive networks. Individuals are either ignorants, spreaders, or stiflers; spreaders stop by contact with other spreaders or stiflers or after an independent random waiting time sampled from a given distribution, modeling a spontaneous loss of interest. The topology of the underlying population network is incorporated by modeling it as a broad class of symmetric networks, whose vertices are partitioned into finitely many orbit types. This yields a unified framework for homogeneous and heterogeneous networks. For sequences of finite quasi-transitive graphs, and for infinite quasi-transitive graphs with subexponential growth, we establish a Functional Law of Large Numbers and a Functional Central Limit Theorem for the densities of each vertex type for the three states. The mean-field limit is described by a system of nonlinear integral equations, while fluctuations are asymptotically Gaussian and governed by a system of stochastic integral equations with explicit covariance. Our results show how the topology and the law of spontaneous stifling jointly shape the speed and variability of rumor outbreaks. As a special case, our model reduces to the classical Maki-Thompson model when spontaneous stifling is absent. △ Less

Submitted 24 November, 2025; originally announced November 2025.

Comments: 32 pages, 20 images

MSC Class: 60G55; 60F17

Energy Efficiency in Network Slicing: Survey and Taxonomy

Authors: Adnei Willian Donatti, Marcia Cristina Machado, Marvin Alexander Lopez Martinez, Sabino Rogério S. Antunes, Eli Carlos Figueiredo Souza, Sand Correa, Tiago Ferreto, José Augusto Suruagy, Joberto S. B. Martins, Tereza Cristina Carvalho

Abstract: Network Slicing (NS) is a fundamental feature of 5G, 6G, and future mobile networks, enabling logically isolated virtual networks over shared infrastructure. As data demand increases and services diversify, ensuring Energy Efficiency (EE) in NS is vital (not only for operational cost savings but also to reduce the Information and Communication Technology (ICT) sector's environmental footprint). Th… ▽ More Network Slicing (NS) is a fundamental feature of 5G, 6G, and future mobile networks, enabling logically isolated virtual networks over shared infrastructure. As data demand increases and services diversify, ensuring Energy Efficiency (EE) in NS is vital (not only for operational cost savings but also to reduce the Information and Communication Technology (ICT) sector's environmental footprint). This survey addresses the need for a comprehensive and holistic perspective on energy-efficient NS by reviewing and classifying recent strategies across the NS life cycle. Our contributions are threefold: (i) a thorough review of state-of-the-art techniques aimed at reducing energy consumption in NS; (ii) a novel taxonomy that organizes strategies into infrastructure, path/route, and slice operation levels; and (iii) the identification of open challenges and research directions, with a focus on systemic, cross-layer, and AI-driven approaches. By consolidating insights from recent developments, our work bridges existing gaps in the literature, offering a structured foundation for researchers and practitioners to design, evaluate, and improve energy-efficient network slicing systems. △ Less

Submitted 2 November, 2025; originally announced November 2025.

Comments: 20 pages, 6 figures, 5 tables

Journal ref: IEEE Access, 2025

A Sensitivity Analysis Methodology for Rule-Based Stochastic Chemical Systems

Authors: Erika M. Herrera Machado, Jakob L. Andersen, Rolf Fagerberg, Daniel Merkle

Abstract: In this study, we introduce a sensitivity analysis methodology for stochastic systems in chemistry, where dynamics are often governed by random processes. Our approach is based on gradient estimation via finite differences, averaging simulation outcomes, and analyzing variability under intrinsic noise. We characterize gradient uncertainty as an angular range within which all plausible gradient dir… ▽ More In this study, we introduce a sensitivity analysis methodology for stochastic systems in chemistry, where dynamics are often governed by random processes. Our approach is based on gradient estimation via finite differences, averaging simulation outcomes, and analyzing variability under intrinsic noise. We characterize gradient uncertainty as an angular range within which all plausible gradient directions are expected to lie. A key feature of our approach is that this uncertainty measure adaptively guides the number of simulations performed for each nominal-perturbation pair of points in order to minimize unnecessary computations while maintaining robustness. Systematically exploring a range of parameter values across the parameter space, rather than focusing on a single value, allows us to identify not only sensitive parameters but also regions of parameter space associated with different levels of sensitivity. These results are visualized through vector field plots to offer an intuitive representation of local sensitivity across parameter space. Additionally, global sensitivity coefficients over sampled points in the parameter space are computed to capture overall trends. Flexibility regarding the choice of output observable measures is another key feature of our method: while traditional sensitivity analyses often focus on species concentrations, our framework allows for the definition of a large range of problem-specific observables. This makes it broadly applicable in diverse chemical and biochemical scenarios. We demonstrate our approach on two systems: classical Michaelis-Menten kinetics and a rule-based model of the formose reaction, using the cheminformatics software MØD for Gillespie-based stochastic simulations. △ Less

Submitted 9 January, 2026; v1 submitted 22 September, 2025; originally announced September 2025.

From Nash to Cournot--Nash equilibria via $Γ$-convergence

Authors: João Miguel Machado, Guilherme Mazanti, Laurent Pfeiffer

Abstract: This work addresses the issue of the convergence of an $N$-player game towards a limit model involving a continuum of players, as the number of agents $N$ goes to infinity. More precisely, we investigate the convergence of Nash equilibria to a Cournot--Nash equilibrium of the limit model. When the cost function of the players is the first variation of some potential function, equilibria can be cha… ▽ More This work addresses the issue of the convergence of an $N$-player game towards a limit model involving a continuum of players, as the number of agents $N$ goes to infinity. More precisely, we investigate the convergence of Nash equilibria to a Cournot--Nash equilibrium of the limit model. When the cost function of the players is the first variation of some potential function, equilibria can be characterized by a stationarity condition, satisfied in particular by the minimizers of the potential. We demonstrate such a characterization under low regularity assumptions. Then we focus on the case where the players interact in a pairwise fashion; in this case we show that the original sequence of $N$-player games also admit a potential structure and prove that their corresponding potential functions converge in the sense of $Γ$-convergence to the potential function of the limit game. △ Less

Submitted 2 April, 2026; v1 submitted 2 September, 2025; originally announced September 2025.

Rule-Based Gillespie Simulation of Chemical Systems

Authors: Erika M. Herrera Machado, Jakob L. Andersen, Rolf Fagerberg, Christoph Flamm, Daniel Merkle, Peter F. Stadler

Abstract: The MØD computational framework implements rule-based generative chemistries as explicit transformations of graphs representing chemical structural formulae. Here, we expand MØD by a stochastic simulation module that simulates the time evolution of species concentrations using Gillespie's well-known stochastic simulation algorithm (SSA). This module distinguishes itself among competing implementat… ▽ More The MØD computational framework implements rule-based generative chemistries as explicit transformations of graphs representing chemical structural formulae. Here, we expand MØD by a stochastic simulation module that simulates the time evolution of species concentrations using Gillespie's well-known stochastic simulation algorithm (SSA). This module distinguishes itself among competing implementations of rule-based stochastic simulation engines by its flexible network expansion mechanism and its functionality for defining custom reaction rate functions. It enables direct sampling from actual reactions instead of rules. We present methodology and implementation details followed by examples which demonstrate the capabilities of the stochastic simulation engine. △ Less

Submitted 29 September, 2025; v1 submitted 1 September, 2025; originally announced September 2025.

Comments: 20 pages, 7 figures

Evaluating the ratio of the exclusive vector meson photoproduction to inclusive hadron/jet production cross section in ultraperipheral heavy ion collisions

Authors: Joao Vitor C. Lovato, Edgar Huayra, Magno V. T. Machado

Abstract: Using the QCD color dipole picture to study exclusive vector meson and inclusive jet/open meson photoproduction, we calculate the ratio of elastic meson production to inclusive hadron production cross sections for ultraperipheral heavy ion collisions. Predictions are evaluated for run 4 of the Large Hadron Collider in proton-nucleus ($pA$) and nucleus-nucleus ($AA$) collisions. The dependencies of… ▽ More Using the QCD color dipole picture to study exclusive vector meson and inclusive jet/open meson photoproduction, we calculate the ratio of elastic meson production to inclusive hadron production cross sections for ultraperipheral heavy ion collisions. Predictions are evaluated for run 4 of the Large Hadron Collider in proton-nucleus ($pA$) and nucleus-nucleus ($AA$) collisions. The dependencies of the ratio on jet/hadron transverse momentum and atomic number are investigated. The double ratio $R_{\mathrm{UPC}}$ for $AA$ over $pA$ collisions is also computed, which has been previously proposed as a new observable probing parton saturation physics. △ Less

Submitted 4 May, 2026; v1 submitted 25 August, 2025; originally announced August 2025.

Comments: 29 pages, 14 figures, and 3 tables

An Analysis of Action-Value Temporal-Difference Methods That Learn State Values

Authors: Brett Daley, Prabhat Nagarajan, Martha White, Marlos C. Machado

Abstract: The hallmark feature of temporal-difference (TD) learning is bootstrapping: using value predictions to generate new value predictions. The vast majority of TD methods for control learn a policy by bootstrapping from a single action-value function (e.g., Q-learning and Sarsa). Significantly less attention has been given to methods that bootstrap from two asymmetric value functions: i.e., methods th… ▽ More The hallmark feature of temporal-difference (TD) learning is bootstrapping: using value predictions to generate new value predictions. The vast majority of TD methods for control learn a policy by bootstrapping from a single action-value function (e.g., Q-learning and Sarsa). Significantly less attention has been given to methods that bootstrap from two asymmetric value functions: i.e., methods that learn state values as an intermediate step in learning action values. Existing algorithms in this vein can be categorized as either QV-learning or AV-learning. Though these algorithms have been investigated to some degree in prior work, it remains unclear if and when it is advantageous to learn two value functions instead of just one -- and whether such approaches are theoretically sound in general. In this paper, we analyze these algorithmic families in terms of convergence and sample efficiency. We find that while both families are more efficient than Expected Sarsa in the prediction setting, only AV-learning methods offer any major benefit over Q-learning in the control setting. Finally, we introduce a new AV-learning algorithm called Regularized Dueling Q-learning (RDQ), which significantly outperforms Dueling DQN in the MinAtar benchmark. △ Less

Submitted 4 September, 2025; v1 submitted 13 July, 2025; originally announced July 2025.

Comments: Published in RLC/RLJ 2025. Camera-ready version

Investigating QCD Dynamical Entropy in high-energy nuclear collisions

Authors: G. S. Ramos, L. S. Moriggi, M. V. T. Machado

Abstract: In this work, the concept of QCD dynamical entropy is extended to heavy ion systems. This notion of entropy can be understood as a relative entropy and can also be used to estimate the initial entropy density in ultra-relativistic heavy ion collisions. The key quantity used to calculate this entropy is the nuclear unintegrated gluon distribution (nUGD), which provides a transverse momentum probabi… ▽ More In this work, the concept of QCD dynamical entropy is extended to heavy ion systems. This notion of entropy can be understood as a relative entropy and can also be used to estimate the initial entropy density in ultra-relativistic heavy ion collisions. The key quantity used to calculate this entropy is the nuclear unintegrated gluon distribution (nUGD), which provides a transverse momentum probability density. In the numerical analysis, both the geometric scaling phenomenon and the Glauber-Gribov approach have been used to evaluate realistic models for the nUGD. It is shown that the normalization procedure and the geometric scaling property make the dynamical entropy almost independent of the nucleus mass number $A$. Results are presented for the dynamical entropy density, $dS_D/dy$, in terms of the rapidity. △ Less

Submitted 12 July, 2025; originally announced July 2025.

Comments: 10 pages, 5 figures. Version to be published in Physics Letters B

A Study of Value-Aware Eigenoptions

Authors: Harshil Kotamreddy, Marlos C. Machado

Abstract: Options, which impose an inductive bias toward temporal and hierarchical structure, offer a powerful framework for reinforcement learning (RL). While effective in sequential decision-making, they are often handcrafted rather than learned. Among approaches for discovering options, eigenoptions have shown strong performance in exploration, but their role in credit assignment remains underexplored. I… ▽ More Options, which impose an inductive bias toward temporal and hierarchical structure, offer a powerful framework for reinforcement learning (RL). While effective in sequential decision-making, they are often handcrafted rather than learned. Among approaches for discovering options, eigenoptions have shown strong performance in exploration, but their role in credit assignment remains underexplored. In this paper, we investigate whether eigenoptions can accelerate credit assignment in model-free RL, evaluating them in tabular and pixel-based gridworlds. We find that pre-specified eigenoptions aid not only exploration but also credit assignment, whereas online discovery can bias the agent's experience too strongly and hinder learning. In the context of deep RL, we also propose a method for learning option-values under non-linear function approximation, highlighting the impact of termination conditions on performance. Our findings reveal both the promise and complexity of using eigenoptions, and options more broadly, to simultaneously support credit assignment and exploration in reinforcement learning. △ Less

Submitted 11 July, 2025; originally announced July 2025.

Comments: Presented at the RLC Workshop on Inductive Biases in Reinforcement Learning 2025

Deep Reinforcement Learning with Gradient Eligibility Traces

Authors: Esraa Elelimy, Brett Daley, Andrew Patterson, Marlos C. Machado, Adam White, Martha White

Abstract: Achieving fast and stable off-policy learning in deep reinforcement learning (RL) is challenging. Most existing methods rely on semi-gradient temporal-difference (TD) methods for their simplicity and efficiency, but are consequently susceptible to divergence. While more principled approaches like Gradient TD (GTD) methods have strong convergence guarantees, they have rarely been used in deep RL. R… ▽ More Achieving fast and stable off-policy learning in deep reinforcement learning (RL) is challenging. Most existing methods rely on semi-gradient temporal-difference (TD) methods for their simplicity and efficiency, but are consequently susceptible to divergence. While more principled approaches like Gradient TD (GTD) methods have strong convergence guarantees, they have rarely been used in deep RL. Recent work introduced the generalized Projected Bellman Error ($\overline{\text{PBE}}$), enabling GTD methods to work efficiently with nonlinear function approximation. However, this work is limited to one-step methods, which are slow at credit assignment and require a large number of samples. In this paper, we extend the generalized $\overline{\text{PBE}}$ objective to support multistep credit assignment based on the $λ$-return and derive three gradient-based methods that optimize this new objective. We provide both a forward-view formulation compatible with experience replay and a backward-view formulation compatible with streaming algorithms. Finally, we evaluate the proposed algorithms and show that they outperform both PPO and StreamQ in MuJoCo and MinAtar environments, respectively. Code available at https://github.com/esraaelelimy/gtd\_algos △ Less

Submitted 18 September, 2025; v1 submitted 11 July, 2025; originally announced July 2025.

Journal ref: Reinforcement Learning Journal, 2025

Deep Double Q-learning

Authors: Prabhat Nagarajan, Martha White, Marlos C. Machado

Abstract: Double Q-learning is a classical control algorithm that mitigates the maximization bias of Q-learning. To do so, it explicitly trains two independent action-value functions and uses them to decouple action-selection and action-evaluation when computing bootstrap targets. Double DQN adapts target bootstrap decoupling to deep reinforcement learning (RL), but explicitly trains only a single action-va… ▽ More Double Q-learning is a classical control algorithm that mitigates the maximization bias of Q-learning. To do so, it explicitly trains two independent action-value functions and uses them to decouple action-selection and action-evaluation when computing bootstrap targets. Double DQN adapts target bootstrap decoupling to deep reinforcement learning (RL), but explicitly trains only a single action-value function and does not fully decouple its estimators. Consequently, the two estimators remain correlated, and overestimation persists. In this paper, we introduce Deep Double Q-learning (DDQL), a deep RL algorithm that explicitly trains two Q-functions through Double Q-learning. DDQL stabilizes training through a combination of techniques, including lower replay ratios, longer target network update intervals, and shared layers. Across 57 Atari 2600 games, DDQL improves aggregate performance over Double DQN, outperforming it on 47 games while further reducing overestimation. In addition, we study key design choices when adapting Double Q-learning to deep RL, including the network architecture, replay ratio, and minibatch sampling strategies. △ Less

Submitted 14 May, 2026; v1 submitted 30 June, 2025; originally announced July 2025.

Comments: 44 pages

Discovering Temporal Structure: An Overview of Hierarchical Reinforcement Learning

Authors: Martin Klissarov, Akhil Bagaria, Ziyan Luo, George Konidaris, Doina Precup, Marlos C. Machado

Abstract: Developing agents capable of exploring, planning and learning in complex open-ended environments is a grand challenge in artificial intelligence (AI). Hierarchical reinforcement learning (HRL) offers a promising solution to this challenge by discovering and exploiting the temporal structure within a stream of experience. The strong appeal of the HRL framework has led to a rich and diverse body of… ▽ More Developing agents capable of exploring, planning and learning in complex open-ended environments is a grand challenge in artificial intelligence (AI). Hierarchical reinforcement learning (HRL) offers a promising solution to this challenge by discovering and exploiting the temporal structure within a stream of experience. The strong appeal of the HRL framework has led to a rich and diverse body of literature attempting to discover a useful structure. However, it is still not clear how one might define what constitutes good structure in the first place, or the kind of problems in which identifying it may be helpful. This work aims to identify the benefits of HRL from the perspective of the fundamental challenges in decision-making, as well as highlight its impact on the performance trade-offs of AI agents. Through these benefits, we then cover the families of methods that discover temporal structure in HRL, ranging from learning directly from online experience to offline datasets, to leveraging large language models (LLMs). Finally, we highlight the challenges of temporal structure discovery and the domains that are particularly well-suited for such endeavours. △ Less

Submitted 16 June, 2025; originally announced June 2025.

The LHC as an Axion-Photon Collider

Authors: Sergio Barbosa, Matheus Coelho, Sylvain Fichet, Gustavo Gil da Silveira, Magno Machado

Abstract: Assuming the existence of an axion-like particle (ALP), beams of relativistic particles emit fluxes of quasi-real ALPs, analogous to the photon fluxes described by Weizsäcker-Williams-type approximations. Consequently, ALP-ALP and ALP-photon collisions can occur at the LHC. We initiate the study of the LHC as an ALP collider, and show that ALP collisions provide competitive probes of certain ALP c… ▽ More Assuming the existence of an axion-like particle (ALP), beams of relativistic particles emit fluxes of quasi-real ALPs, analogous to the photon fluxes described by Weizsäcker-Williams-type approximations. Consequently, ALP-ALP and ALP-photon collisions can occur at the LHC. We initiate the study of the LHC as an ALP collider, and show that ALP collisions provide competitive probes of certain ALP couplings. We show that ALP fluxes from heavy ions are suppressed relative to those from protons, unlike their photon counterpart. As a result, the most likely processes are ALP-photon collisions occurring in the proton-ion ($p$A) ultraperipheral collisions. Using our implementation of ALP fluxes in simulation tools, we show that ALP-photon collisions in $p$Pb efficiently probe ALP couplings to third generation fermions. LHC data with realistic $p$Pb luminosity can constrain the product of ALP couplings to nucleons and top quarks at the level of $O(0.01$ TeV$^{-1})$. Notably, ALP-photon collisions naturally provide the leading probe of ALP flavor-violating couplings to the top quark. We suggest that the 2016 $p$Pb dataset collected at CMS, ATLAS, and LHCb should be explored for evidence of such collisions. △ Less

Submitted 31 October, 2025; v1 submitted 11 June, 2025; originally announced June 2025.

Comments: 6 pages, 2 figures. v2: text slightly expanded, matches PRL version

Journal ref: Phys. Rev. Lett. 135, 181801. Published 30 October, 2025

Universality of scaling entropy in charged hadron multiplicity distributions at the LHC

Authors: L. S. Moriggi, F. S. Navarra, M. V. T. Machado

Abstract: In this work, we investigate the scaling behavior of the entropy associated with the charged hadron multiplicity distribution P(N) in proton-proton collisions at the LHC. We show that the growth of this entropic indicator as a function of the Bjorken x variable exhibits a universal behavior, consistent with observations from deep inelastic scattering (DIS). This universality suggests that the entr… ▽ More In this work, we investigate the scaling behavior of the entropy associated with the charged hadron multiplicity distribution P(N) in proton-proton collisions at the LHC. We show that the growth of this entropic indicator as a function of the Bjorken x variable exhibits a universal behavior, consistent with observations from deep inelastic scattering (DIS). This universality suggests that the entropy scaling is a property of the initial state and reflects the diffusive nature of gluon dynamics at small x. Furthermore, we demonstrate that high-multiplicity events are not accurately described by traditional KNO scaling and require a more precise description based on a diffusion scaling framework. This new scaling emerges naturally from the universal growth of partonic entropy and offers a deeper insight into the dynamics of particle production in high-energy hadronic collisions. △ Less

Submitted 13 November, 2025; v1 submitted 11 June, 2025; originally announced June 2025.

Comments: 11 pages, 7 figures. Final version to be published in Physical Review D

The Cell Must Go On: Agar.io for Continual Reinforcement Learning

Authors: Mohamed A. Mohamed, Kateryna Nekhomiazh, Vedant Vyas, Marcos M. Jose, Andrew Patterson, Marlos C. Machado

Abstract: Continual reinforcement learning (RL) concerns agents that are expected to learn continually, rather than converge to a policy that is then fixed for evaluation. This setting is well-suited to environments that the agent perceives as changing over time, rendering any static policy ineffective. In continual RL, researchers often simulate such changes either by modifying episodic environments to inc… ▽ More Continual reinforcement learning (RL) concerns agents that are expected to learn continually, rather than converge to a policy that is then fixed for evaluation. This setting is well-suited to environments that the agent perceives as changing over time, rendering any static policy ineffective. In continual RL, researchers often simulate such changes either by modifying episodic environments to incorporate task shifts during interaction or by designing simulators that explicitly model continual dynamics. However, transforming episodic problems into continual ones primarily captures scenarios involving abrupt changes in the data stream and still relies on episodic structure. Meanwhile, the few simulators explicitly designed for empirical continual RL research are often limited in scope or complexity. In this paper, we introduce AgarCL, a research platform for continual RL that enables agents to progress toward increasingly sophisticated behaviour. AgarCL is based on the game Agar.io, a non-episodic, high-dimensional problem with stochastic, ever-evolving dynamics, continuous actions, and partial observability. We provide benchmark results for DQN, PPO, and SAC on the primary continual RL challenge, as well as across a suite of smaller tasks within AgarCL. These smaller tasks isolate aspects of the full environment and allow us to characterize the distinct challenges posed by different components of the game. We further evaluate three continual learning methods-Shrink and Perturb, ReDo, and Continual Backpropagation-and observe little improvement over standard RL algorithms, suggesting that the challenges posed by AgarCL extend beyond the stability-plasticity dilemma. △ Less

Submitted 8 August, 2026; v1 submitted 23 May, 2025; originally announced May 2025.

Reward-Aware Proto-Representations in Reinforcement Learning

Authors: Hon Tik Tse, Siddarth Chandrasekar, Marlos C. Machado

Abstract: In recent years, the successor representation (SR) has attracted increasing attention in reinforcement learning (RL), and it has been used to address some of its key challenges, such as exploration, credit assignment, and generalization. The SR can be seen as representing the underlying credit assignment structure of the environment by implicitly encoding its induced transition dynamics. However,… ▽ More In recent years, the successor representation (SR) has attracted increasing attention in reinforcement learning (RL), and it has been used to address some of its key challenges, such as exploration, credit assignment, and generalization. The SR can be seen as representing the underlying credit assignment structure of the environment by implicitly encoding its induced transition dynamics. However, the SR is reward-agnostic. In this paper, we discuss a similar representation that also takes into account the reward dynamics of the problem. We study the default representation (DR), a recently proposed representation with limited theoretical (and empirical) analysis. Here, we lay some of the theoretical foundation underlying the DR in the tabular case by (1) deriving dynamic programming and (2) temporal-difference methods to learn the DR, (3) characterizing the basis for the vector space of the DR, and (4) formally extending the DR to the function approximation case through default features. Empirically, we analyze the benefits of the DR in many of the settings in which the SR has been applied, including (1) reward shaping, (2) option discovery, (3) exploration, and (4) transfer learning. Our results show that, compared to the SR, the DR gives rise to qualitatively different, reward-aware behaviour and quantitatively better performance in several settings. △ Less

Submitted 30 January, 2026; v1 submitted 22 May, 2025; originally announced May 2025.

Comments: NeurIPS 2025

Absence of loops for the Wasserstein-$\mathcal{H}^1$ problem: the concentration/blow-up argument

Authors: Jo{ã}o Miguel Machado

Abstract: In the present work we prove that minimizers of the Wasserstein-$\mathscr{H}^1$ problem, introduced recently by Chambolle et. al., are trees in two cases: when the target measure is a sum of finitely many Dirac masses or when it has a bounded density. In the present work we prove that minimizers of the Wasserstein-$\mathscr{H}^1$ problem, introduced recently by Chambolle et. al., are trees in two cases: when the target measure is a sum of finitely many Dirac masses or when it has a bounded density. △ Less

Submitted 20 April, 2026; v1 submitted 14 May, 2025; originally announced May 2025.

Brazilian Report on Dark Matter 2024

Authors: I. F. M. Albuquerque, J. Alcaniz, A. Alves, J. Amaral, C. Bonifazi, H. A. Borges, S. Carneiro, L. Casarini, D. Cogollo, A. G. Dias, G. C. Dorsch, A. Esmaili, G. Gil da Silveira, C. Gobel, V. P. Gonçalves, A. S. Jesus, D. Hadjimichef, P. C. de Holanda, R. F. L. Holanda, E. Kemp, A. Lessa, A. Machado, M. V T. Machado, M. Makler, V. Marra , et al. (29 additional authors not shown)

Abstract: One of the key scientific objectives for the next decade is to uncover the nature of dark matter (DM). We should continue prioritizing targets such as weakly-interacting massive particles (WIMPs), Axions, and other low-mass dark matter candidates to improve our chances of achieving it. A varied and ongoing portfolio of experiments spanning different scales and detection methods is essential to max… ▽ More One of the key scientific objectives for the next decade is to uncover the nature of dark matter (DM). We should continue prioritizing targets such as weakly-interacting massive particles (WIMPs), Axions, and other low-mass dark matter candidates to improve our chances of achieving it. A varied and ongoing portfolio of experiments spanning different scales and detection methods is essential to maximize our chances of discovering its composition. This report paper provides an updated overview of the Brazilian community's activities in dark matter and dark sector physics over the past years with a view for the future. It underscores the ongoing need for financial support for Brazilian groups actively engaged in experimental research to sustain the Brazilian involvement in the global search for dark matter particles △ Less

Submitted 8 May, 2025; v1 submitted 22 April, 2025; originally announced April 2025.

Comments: Editor-Convener: Farinaldo Queiroz. Report for the Latin American Association for High Energy, Cosmology and Astroparticle Physics

Kimina-Prover Preview: Towards Large Formal Reasoning Models with Reinforcement Learning

Authors: Haiming Wang, Mert Unsal, Xiaohan Lin, Mantas Baksys, Junqi Liu, Marco Dos Santos, Flood Sung, Marina Vinyes, Zhenzhe Ying, Zekai Zhu, Jianqiao Lu, Hugues de Saxcé, Bolton Bailey, Chendong Song, Chenjun Xiao, Dehao Zhang, Ebony Zhang, Frederick Pu, Han Zhu, Jiawei Liu, Jonas Bayer, Julien Michel, Longhui Yu, Léo Dreyfus-Schmidt, Lewis Tunstall , et al. (15 additional authors not shown)

Abstract: We introduce Kimina-Prover Preview, a large language model that pioneers a novel reasoning-driven exploration paradigm for formal theorem proving, as showcased in this preview release. Trained with a large-scale reinforcement learning pipeline from Qwen2.5-72B, Kimina-Prover demonstrates strong performance in Lean 4 proof generation by employing a structured reasoning pattern we term \textit{forma… ▽ More We introduce Kimina-Prover Preview, a large language model that pioneers a novel reasoning-driven exploration paradigm for formal theorem proving, as showcased in this preview release. Trained with a large-scale reinforcement learning pipeline from Qwen2.5-72B, Kimina-Prover demonstrates strong performance in Lean 4 proof generation by employing a structured reasoning pattern we term \textit{formal reasoning pattern}. This approach allows the model to emulate human problem-solving strategies in Lean, iteratively generating and refining proof steps. Kimina-Prover sets a new state-of-the-art on the miniF2F benchmark, reaching 80.7% with pass@8192. Beyond improved benchmark performance, our work yields several key insights: (1) Kimina-Prover exhibits high sample efficiency, delivering strong results even with minimal sampling (pass@1) and scaling effectively with computational budget, stemming from its unique reasoning pattern and RL training; (2) we demonstrate clear performance scaling with model size, a trend previously unobserved for neural theorem provers in formal mathematics; (3) the learned reasoning style, distinct from traditional search algorithms, shows potential to bridge the gap between formal verification and informal mathematical intuition. We open source distilled versions with 1.5B and 7B parameters of Kimina-Prover △ Less

Submitted 15 April, 2025; originally announced April 2025.

Comments: 22 pages

Command A: An Enterprise-Ready Large Language Model

Authors: Team Cohere, :, Aakanksha, Arash Ahmadian, Marwan Ahmed, Jay Alammar, Milad Alizadeh, Yazeed Alnumay, Sophia Althammer, Arkady Arkhangorodsky, Viraat Aryabumi, Dennis Aumiller, Raphaël Avalos, Zahara Aviv, Sammie Bae, Saurabh Baji, Alexandre Barbet, Max Bartolo, Björn Bebensee, Neeral Beladia, Walter Beller-Morales, Alexandre Bérard, Andrew Berneshawi, Anna Bialas, Phil Blunsom , et al. (205 additional authors not shown)

Abstract: In this report we describe the development of Command A, a powerful large language model purpose-built to excel at real-world enterprise use cases. Command A is an agent-optimised and multilingual-capable model, with support for 23 languages of global business, and a novel hybrid architecture balancing efficiency with top of the range performance. It offers best-in-class Retrieval Augmented Genera… ▽ More In this report we describe the development of Command A, a powerful large language model purpose-built to excel at real-world enterprise use cases. Command A is an agent-optimised and multilingual-capable model, with support for 23 languages of global business, and a novel hybrid architecture balancing efficiency with top of the range performance. It offers best-in-class Retrieval Augmented Generation (RAG) capabilities with grounding and tool use to automate sophisticated business processes. These abilities are achieved through a decentralised training approach, including self-refinement algorithms and model merging techniques. We also include results for Command R7B which shares capability and architectural similarities to Command A. Weights for both models have been released for research purposes. This technical report details our original training pipeline and presents an extensive evaluation of our models across a suite of enterprise-relevant tasks and public benchmarks, demonstrating excellent performance and efficiency. △ Less

Submitted 14 April, 2025; v1 submitted 1 April, 2025; originally announced April 2025.

Comments: 55 pages

Nonextensive aspects of gluon distribution and the implications to QCD phenomenology

Authors: Lucas Soster Moriggi, Magno Valerio Trindade Machado

Abstract: This study presents new insights into gluon transverse momentum distributions through nonextensive statistical mechanics, addressing their implications for QCD phenomenology. The saturation physics and scaling laws present in high energy collision data are investigated as a consequence of gluon distribution modification at high density regime. The analysis explores how these modifications influenc… ▽ More This study presents new insights into gluon transverse momentum distributions through nonextensive statistical mechanics, addressing their implications for QCD phenomenology. The saturation physics and scaling laws present in high energy collision data are investigated as a consequence of gluon distribution modification at high density regime. The analysis explores how these modifications influence observables across different collision systems, such as proton-proton, proton-nucleus, and relativistic heavy-ion collisions. Both high and low $p_T$ regions are successfully described in hadron production. △ Less

Submitted 9 January, 2025; originally announced January 2025.

Comments: 17 pages, 6 Figures. Contribution to MDPI Physics Special Issue "Complexity in High Energy and Statistical Physics"

Precise determination of pomeron intercept via scaling entropy analysis

Authors: Lucas Soster Moriggi, Magno Valério Trindade Machado

Abstract: In this work, we confront the geometrical scaling properties of inclusive DIS cross section ($e+p\rightarrow e +X$) with the scaling entropy obtained from event multiplicity. We show that these two quantities are equivalent in the kinematic range probed by H1 Collaboration data. We propose that scaling entropy associated with partonic interactions is a more efficient way to detect scaling in exper… ▽ More In this work, we confront the geometrical scaling properties of inclusive DIS cross section ($e+p\rightarrow e +X$) with the scaling entropy obtained from event multiplicity. We show that these two quantities are equivalent in the kinematic range probed by H1 Collaboration data. We propose that scaling entropy associated with partonic interactions is a more efficient way to detect scaling in experimental data. We used a combined analysis of the inclusive cross section and entropy obtained from multiplicities $P(N)$ of final-state hadrons to accurately determine the value of the Pomeron intercept. The approach could provide new constraints for future hadron collider experiments and deepen our understanding of parton saturation. △ Less

Submitted 20 December, 2024; originally announced December 2024.

Comments: 8 pages, 6 Figures. Version to be published in Physical Review D

MaestroMotif: Skill Design from Artificial Intelligence Feedback

Authors: Martin Klissarov, Mikael Henaff, Roberta Raileanu, Shagun Sodhani, Pascal Vincent, Amy Zhang, Pierre-Luc Bacon, Doina Precup, Marlos C. Machado, Pierluca D'Oro

Abstract: Describing skills in natural language has the potential to provide an accessible way to inject human knowledge about decision-making into an AI system. We present MaestroMotif, a method for AI-assisted skill design, which yields high-performing and adaptable agents. MaestroMotif leverages the capabilities of Large Language Models (LLMs) to effectively create and reuse skills. It first uses an LLM'… ▽ More Describing skills in natural language has the potential to provide an accessible way to inject human knowledge about decision-making into an AI system. We present MaestroMotif, a method for AI-assisted skill design, which yields high-performing and adaptable agents. MaestroMotif leverages the capabilities of Large Language Models (LLMs) to effectively create and reuse skills. It first uses an LLM's feedback to automatically design rewards corresponding to each skill, starting from their natural language description. Then, it employs an LLM's code generation abilities, together with reinforcement learning, for training the skills and combining them to implement complex behaviors specified in language. We evaluate MaestroMotif using a suite of complex tasks in the NetHack Learning Environment (NLE), demonstrating that it surpasses existing approaches in both performance and usability. △ Less

Submitted 11 December, 2024; originally announced December 2024.

Plastic Learning with Deep Fourier Features

Authors: Alex Lewandowski, Dale Schuurmans, Marlos C. Machado

Abstract: Deep neural networks can struggle to learn continually in the face of non-stationarity. This phenomenon is known as loss of plasticity. In this paper, we identify underlying principles that lead to plastic algorithms. In particular, we provide theoretical results showing that linear function approximation, as well as a special case of deep linear networks, do not suffer from loss of plasticity. We… ▽ More Deep neural networks can struggle to learn continually in the face of non-stationarity. This phenomenon is known as loss of plasticity. In this paper, we identify underlying principles that lead to plastic algorithms. In particular, we provide theoretical results showing that linear function approximation, as well as a special case of deep linear networks, do not suffer from loss of plasticity. We then propose deep Fourier features, which are the concatenation of a sine and cosine in every layer, and we show that this combination provides a dynamic balance between the trainability obtained through linearity and the effectiveness obtained through the nonlinearity of neural networks. Deep networks composed entirely of deep Fourier features are highly trainable and sustain their trainability over the course of learning. Our empirical results show that continual learning performance can be drastically improved by replacing ReLU activations with deep Fourier features. These results hold for different continual learning scenarios (e.g., label noise, class incremental learning, pixel permutations) on all major supervised learning datasets used for continual learning research, such as CIFAR10, CIFAR100, and tiny-ImageNet. △ Less

Submitted 27 October, 2024; originally announced October 2024.

The glue that binds us all -- Latin America and the Electron-Ion Collider

Authors: A. C. Aguilar, A. Bashir, J. J. Cobos-Martínez, A. Courtoy, B. El-Bennich, D. de Florian, T. Frederico, V. P. Gonçalves, M. Hentschinski, R. J. Hernández-Pinto, G. Krein, M. V. T. Machado, J. P. B. C. de Melo, W. de Paula, R. Sassot, F. E. Serna, L. Albino, I. Borsa, L. Cieri, I. M. Higuera-Angulo, J. Mazzitelli, Á. Miramontes, K. Raya, F. Salazar, G. Sborlini , et al. (1 additional authors not shown)

Abstract: The Electron-Ion Collider, a next generation electron-hadron and electron-nuclei scattering facility, will be built at Brookhaven National Laboratory. The wealth of new data will shape research in hadron physics, from nonperturbative QCD techniques to perturbative QCD improvements and global QCD analyses, for the decades to come. With the present proposal, Latin America based physicists, whose exp… ▽ More The Electron-Ion Collider, a next generation electron-hadron and electron-nuclei scattering facility, will be built at Brookhaven National Laboratory. The wealth of new data will shape research in hadron physics, from nonperturbative QCD techniques to perturbative QCD improvements and global QCD analyses, for the decades to come. With the present proposal, Latin America based physicists, whose expertise lies on the theory and phenomenology side, make the case for the past and future efforts of a growing community, working hand-in-hand towards developing theoretical tools and predictions to analyze, interpret and optimize the results that will be obtained at the EIC, unveiling the role of the glue that binds us all. This effort is along the lines of various initiatives taken in the U.S., and supported by colleagues worldwide, such as the ones by the EIC User Group which were highlighted during the Snowmass Process and the Particle Physics Project Prioritization Panel (P5). △ Less

Submitted 29 April, 2025; v1 submitted 26 September, 2024; originally announced September 2024.

Comments: White Paper contribution to the Latin American Strategy Forum for Research Infrastructure (III LASF4RI). Updated version includes an extended discussion of the "Perspectives and Challenges" section and an additional section about "Current projects and future Latin American contributions to the EIC". Accepted for publication in Brazilian Journal of Physics

Evaluating Low-Resource Lane Following Algorithms for Compute-Constrained Automated Vehicles

Authors: Beñat Froemming-Aldanondo, Tatiana Rastoskueva, Michael Evans, Marcial Machado, Anna Vadella, Rickey Johnson, Luis Escamilla, Milan Jostes, Devson Butani, Ryan Kaddis, Chan-Jin Chung, Joshua Siegel

Abstract: Reliable lane-following is essential for automated and assisted driving, yet existing solutions often rely on models that require extensive computational resources, limiting their deployment in compute-constrained vehicles. We evaluate five low-resource lane-following algorithms designed for real-time operation on vehicles with limited computing resources. Performance was assessed through simulati… ▽ More Reliable lane-following is essential for automated and assisted driving, yet existing solutions often rely on models that require extensive computational resources, limiting their deployment in compute-constrained vehicles. We evaluate five low-resource lane-following algorithms designed for real-time operation on vehicles with limited computing resources. Performance was assessed through simulation and deployment on real drive-by-wire electric vehicles, with evaluation metrics including reliability, comfort, speed, and adaptability. The top-performing methods used unsupervised learning to detect and separate lane lines with processing time under 10 ms per frame, outperforming compute-intensive and poor generalizing deep learning approaches. These approaches demonstrated robustness across lighting conditions, road textures, and lane geometries. The findings highlight the potential for efficient lane detection approaches to enhance the accessibility and reliability of autonomous vehicle technologies. Reducing computing requirements enables lane keeping to be widely deployed in vehicles as part of lower-level automation, including active safety systems. △ Less

Submitted 2 March, 2025; v1 submitted 4 September, 2024; originally announced September 2024.

Comments: Supported by the National Science Foundation under Grants No. 2150292 and 2150096

A Roadside Unit for Infrastructure Assisted Intersection Control of Autonomous Vehicles

Authors: Michael Evans, Marcial Machado, Rickey Johnson, Anna Vadella, Luis Escamilla, Beñat Froemming-Aldanondo, Tatiana Rastoskueva, Milan Jostes, Devson Butani, Ryan Kaddis, Chan-Jin Chung, Joshua Siegel

Abstract: Recent advances in autonomous vehicle technologies and cellular network speeds motivate developments in vehicle-to-everything (V2X) communications. Enhanced road safety features and improved fuel efficiency are some of the motivations behind V2X for future transportation systems. Adaptive intersection control systems have considerable potential to achieve these goals by minimizing idle times and p… ▽ More Recent advances in autonomous vehicle technologies and cellular network speeds motivate developments in vehicle-to-everything (V2X) communications. Enhanced road safety features and improved fuel efficiency are some of the motivations behind V2X for future transportation systems. Adaptive intersection control systems have considerable potential to achieve these goals by minimizing idle times and predicting short-term future traffic conditions. Integrating V2X into traffic management systems introduces the infrastructure necessary to make roads safer for all users and initiates the shift towards more intelligent and connected cities. To demonstrate our control algorithm, we implement both a simulated and real-world representation of a 4-way intersection and crosswalk scenario with 2 self-driving electric vehicles, a roadside unit (RSU), and a traffic light. Our architecture reduces acceleration and braking through intersections by up to 75.35%, which has been shown to minimize fuel consumption in gas vehicles. We propose a cost-effective solution to intelligent and connected intersection control to serve as a proof-of-concept model suitable as the basis for continued research and development. Code for this project is available at https://github.com/MMachado05/REU-2024. △ Less

Submitted 4 March, 2025; v1 submitted 1 September, 2024; originally announced September 2024.

Comments: Supported by the National Science Foundation under Grants No. 2150292 and 2150096

Demystifying the Recency Heuristic in Temporal-Difference Learning

Authors: Brett Daley, Marlos C. Machado, Martha White

Abstract: The recency heuristic in reinforcement learning is the assumption that stimuli that occurred closer in time to an acquired reward should be more heavily reinforced. The recency heuristic is one of the key assumptions made by TD($λ$), which reinforces recent experiences according to an exponentially decaying weighting. In fact, all other widely used return estimators for TD learning, such as $n$-st… ▽ More The recency heuristic in reinforcement learning is the assumption that stimuli that occurred closer in time to an acquired reward should be more heavily reinforced. The recency heuristic is one of the key assumptions made by TD($λ$), which reinforces recent experiences according to an exponentially decaying weighting. In fact, all other widely used return estimators for TD learning, such as $n$-step returns, satisfy a weaker (i.e., non-monotonic) recency heuristic. Why is the recency heuristic effective for temporal credit assignment? What happens when credit is assigned in a way that violates this heuristic? In this paper, we analyze the specific mathematical implications of adopting the recency heuristic in TD learning. We prove that any return estimator satisfying this heuristic: 1) is guaranteed to converge to the correct value function, 2) has a relatively fast contraction rate, and 3) has a long window of effective credit assignment, yet bounded worst-case variance. We also give a counterexample where on-policy, tabular TD methods violating the recency heuristic diverge. Our results offer some of the first theoretical evidence that credit assignment based on the recency heuristic facilitates learning. △ Less

Submitted 26 August, 2024; v1 submitted 18 June, 2024; originally announced June 2024.

Comments: RLC 2024. 18 pages, 8 figures, 1 table

Journal ref: Reinforcement Learning Journal, vol. 1, no. 1, 2024

Learning Continually by Spectral Regularization

Authors: Alex Lewandowski, Michał Bortkiewicz, Saurabh Kumar, András György, Dale Schuurmans, Mateusz Ostaszewski, Marlos C. Machado

Abstract: Loss of plasticity is a phenomenon where neural networks can become more difficult to train over the course of learning. Continual learning algorithms seek to mitigate this effect by sustaining good performance while maintaining network trainability. We develop a new technique for improving continual learning inspired by the observation that the singular values of the neural network parameters at… ▽ More Loss of plasticity is a phenomenon where neural networks can become more difficult to train over the course of learning. Continual learning algorithms seek to mitigate this effect by sustaining good performance while maintaining network trainability. We develop a new technique for improving continual learning inspired by the observation that the singular values of the neural network parameters at initialization are an important factor for trainability during early phases of learning. From this perspective, we derive a new spectral regularizer for continual learning that better sustains these beneficial initialization properties throughout training. In particular, the regularizer keeps the maximum singular value of each layer close to one. Spectral regularization directly ensures that gradient diversity is maintained throughout training, which promotes continual trainability, while minimally interfering with performance in a single task. We present an experimental analysis that shows how the proposed spectral regularizer can sustain trainability and performance across a range of model architectures in continual supervised and reinforcement learning settings. Spectral regularization is less sensitive to hyperparameters while demonstrating better training in individual tasks, sustaining trainability as new tasks arrive, and achieving better generalization performance. △ Less

Submitted 27 October, 2024; v1 submitted 10 June, 2024; originally announced June 2024.

Multiplicity dependence of the $p_T$-spectra for charged particles and its relationship with partonic entropy

Authors: L. S. Moriggi, G. S. Ramos, M. V. T. Machado

Abstract: We investigate the multiplicity dependence of the transverse momentum $p_T$ spectra of hadrons produced in high-energy collisions. We propose that the partonic distribution be parameterized by its non-extensive entropy and the parton saturation scale $Q_s(x)$. These two variables can be identified from the produced charged hadron distributions and provide important information on the gluon dynamic… ▽ More We investigate the multiplicity dependence of the transverse momentum $p_T$ spectra of hadrons produced in high-energy collisions. We propose that the partonic distribution be parameterized by its non-extensive entropy and the parton saturation scale $Q_s(x)$. These two variables can be identified from the produced charged hadron distributions and provide important information on the gluon dynamics at the moment of interaction. From this perspective we interpret data from different ALICE multiplicity classes at $\sqrt{s}= 13$ TeV and $\sqrt{s}= 5.02$ TeV. A multiplicity dependent scaling function is presented and the dependence of the interaction area on multiplicity is also investigated. △ Less

Submitted 8 July, 2024; v1 submitted 2 May, 2024; originally announced May 2024.

Comments: 11 pages, 7 figures. Version to be published in Physical Review D

Planning the path with Reinforcement Learning: Optimal Robot Motion Planning in RoboCup Small Size League Environments

Authors: Mateus G. Machado, João G. Melo, Cleber Zanchettin, Pedro H. M. Braga, Pedro V. Cunha, Edna N. S. Barros, Hansenclever F. Bassani

Abstract: This work investigates the potential of Reinforcement Learning (RL) to tackle robot motion planning challenges in the dynamic RoboCup Small Size League (SSL). Using a heuristic control approach, we evaluate RL's effectiveness in obstacle-free and single-obstacle path-planning environments. Ablation studies reveal significant performance improvements. Our method achieved a 60% time gain in obstacle… ▽ More This work investigates the potential of Reinforcement Learning (RL) to tackle robot motion planning challenges in the dynamic RoboCup Small Size League (SSL). Using a heuristic control approach, we evaluate RL's effectiveness in obstacle-free and single-obstacle path-planning environments. Ablation studies reveal significant performance improvements. Our method achieved a 60% time gain in obstacle-free environments compared to baseline algorithms. Additionally, our findings demonstrated dynamic obstacle avoidance capabilities, adeptly navigating around moving blocks. These findings highlight the potential of RL to enhance robot motion planning in the challenging and unpredictable SSL environment. △ Less

Submitted 23 April, 2024; originally announced April 2024.

Comments: 12 pages, 3 figures, 3 tables

KIF: A Wikidata-Based Framework for Integrating Heterogeneous Knowledge Sources

Authors: Guilherme Lima, João M. B. Rodrigues, Marcelo Machado, Elton Soares, Sandro R. Fiorini, Raphael Thiago, Leonardo G. Azevedo, Viviane T. da Silva, Renato Cerqueira

Abstract: We present a Wikidata-based framework, called KIF, for virtually integrating heterogeneous knowledge sources. KIF is written in Python and is released as open-source. It leverages Wikidata's data model and vocabulary plus user-defined mappings to construct a unified view of the underlying sources while keeping track of the context and provenance of their statements. The underlying sources can be t… ▽ More We present a Wikidata-based framework, called KIF, for virtually integrating heterogeneous knowledge sources. KIF is written in Python and is released as open-source. It leverages Wikidata's data model and vocabulary plus user-defined mappings to construct a unified view of the underlying sources while keeping track of the context and provenance of their statements. The underlying sources can be triplestores, relational databases, CSV files, etc., which may or may not use the vocabulary and RDF encoding of Wikidata. The end result is a virtual knowledge base which behaves like an "extended Wikidata" and which can be queried using a simple but expressive pattern language, defined in terms of Wikidata's data model. In this paper, we present the design and implementation of KIF, discuss how we have used it to solve a real integration problem in the domain of chemistry (involving Wikidata, PubChem, and IBM CIRCA), and present experimental results on the performance and overhead of KIF △ Less

Submitted 24 July, 2024; v1 submitted 15 March, 2024; originally announced March 2024.

Testing the double-logarithm asymptotic gluon density in ultraperipheral heavy ion collisions at the Large Hadron Collider

Authors: D. A. Fagundes, M. V. T. Machado

Abstract: In this paper, we analyze the application of an analytical gluon distribution based on double-asymptotic scaling to the photoproduction of vector mesons in coherent $pp$, $pA$, and $AA$ collisions at LHC energies, using the color dipole formalism. Predictions for the rapidity distribution are presented for $ρ^0$, $J/ ψ$, $ψ(2S)$, and $Υ(1S)$ mesons photoproduction. An analysis of the uncertainties… ▽ More In this paper, we analyze the application of an analytical gluon distribution based on double-asymptotic scaling to the photoproduction of vector mesons in coherent $pp$, $pA$, and $AA$ collisions at LHC energies, using the color dipole formalism. Predictions for the rapidity distribution are presented for $ρ^0$, $J/ ψ$, $ψ(2S)$, and $Υ(1S)$ mesons photoproduction. An analysis of the uncertainties associated with different implementations of the dipole--proton amplitude is performed. The vector meson photoproduction accompanied by electromagnetic dissociation is also analyzed. △ Less

Submitted 26 June, 2025; v1 submitted 19 February, 2024; originally announced February 2024.

Comments: 16 pages, 12 figures, 2 tables. Matches the final version

Journal ref: Physics 2025, 7, 24

Aya Dataset: An Open-Access Collection for Multilingual Instruction Tuning

Authors: Shivalika Singh, Freddie Vargus, Daniel Dsouza, Börje F. Karlsson, Abinaya Mahendiran, Wei-Yin Ko, Herumb Shandilya, Jay Patel, Deividas Mataciunas, Laura OMahony, Mike Zhang, Ramith Hettiarachchi, Joseph Wilson, Marina Machado, Luisa Souza Moura, Dominik Krzemiński, Hakimeh Fadaei, Irem Ergün, Ifeoma Okoh, Aisha Alaagib, Oshan Mudannayake, Zaid Alyafeai, Vu Minh Chien, Sebastian Ruder, Surya Guthikonda , et al. (8 additional authors not shown)

Abstract: Datasets are foundational to many breakthroughs in modern artificial intelligence. Many recent achievements in the space of natural language processing (NLP) can be attributed to the finetuning of pre-trained models on a diverse set of tasks that enables a large language model (LLM) to respond to instructions. Instruction fine-tuning (IFT) requires specifically constructed and annotated datasets.… ▽ More Datasets are foundational to many breakthroughs in modern artificial intelligence. Many recent achievements in the space of natural language processing (NLP) can be attributed to the finetuning of pre-trained models on a diverse set of tasks that enables a large language model (LLM) to respond to instructions. Instruction fine-tuning (IFT) requires specifically constructed and annotated datasets. However, existing datasets are almost all in the English language. In this work, our primary goal is to bridge the language gap by building a human-curated instruction-following dataset spanning 65 languages. We worked with fluent speakers of languages from around the world to collect natural instances of instructions and completions. Furthermore, we create the most extensive multilingual collection to date, comprising 513 million instances through templating and translating existing datasets across 114 languages. In total, we contribute four key resources: we develop and open-source the Aya Annotation Platform, the Aya Dataset, the Aya Collection, and the Aya Evaluation Suite. The Aya initiative also serves as a valuable case study in participatory research, involving collaborators from 119 countries. We see this as a valuable framework for future research collaborations that aim to bridge gaps in resources. △ Less

Submitted 9 February, 2024; originally announced February 2024.