Search

This paper introduces a hierarchical framework for decision-making and optimization, leveraging Large Language Models (LLMs) for adaptive code generation. Instead of direct decision-making, LLMs generate and refine executable control policies through a meta-policy that guides task generation and a base policy for operational actions. Applied to a simplified microgrid scenario, the approach achieves up to 15 percent cost savings by iteratively improving battery control strategies. The proposed methodology lays a foundation for integrating LLM-based tools into planning and control tasks, offering adaptable and scalable solutions for complex systems while addressing challenges of uncertainty and reproducibility.

Helicenes are a class of molecules potentially suitable in several technological applications with intrinsic structural chirality which makes them interesting scaffolds for chiroptical properties. As desirable is tuning chiroptical property by synthesis, we combine experimental optical characterization and ab-initio calculations to study how different substituents influence the optical properties of [6]helicene. We explore anchoring groups presenting a variety of sizes and chemical nature, finding that both electron withdrawing and donating groups redshift and dwindle the optical activity of the molecule. We suggest the observed dumping in transitions energy and intensity is connected to the strength of the perturbation induced by the substituent on the pi-conjugation of the aromatic rings. Such observations demonstrate how helicenes' chiroptical properties can be fine-tuned via stereochemical control of the substituents and validates a simple yet effective computational setup to model the optics of those systems.

Early detection of chronic and Non-Communicable Diseases (NCDs) is crucial for effective treatment during the initial stages. This study explores the application of wearable devices and Artificial Intelligence (AI) in order to predict weight loss changes in overweight and obese individuals. Using wearable data from a 1-month trial involving around 100 subjects from the AI4FoodDB database, including biomarkers, vital signs, and behavioral data, we identify key differences between those achieving weight loss (>= 2% of their initial weight) and those who do not. Feature selection techniques and classification algorithms reveal promising results, with the Gradient Boosting classifier achieving 84.44% Area Under the Curve (AUC). The integration of multiple data sources (e.g., vital signs, physical and sleep activity, etc.) enhances performance, suggesting the potential of wearable devices and AI in personalized healthcare., Comment: 15 pages, 5 figures, 6 tables, 1 appendix

This paper addresses the convergence analysis of a variant of the LevenbergMarquardt method (LMM) designed for nonlinear least-squares problems with non-zero residue. This variant, called LMM with Singular Scaling (LMMSS), allows the LMM scaling matrix to be singular, encompassing a broader class of regularizers, which has proven useful in certain applications. In order to establish local convergence results, a careful choice of the LMM parameter is made based on the gradient linearization error, dictated by the nonlinearity and size of the residual. Under completeness and local error bound assumptions we prove that the distance from an iterate to the set of stationary points goes to zero superlinearly and that the iterative sequence converges. Furthermore, we also study a globalized version of the method obtained using linesearch and prove that any limit point of the generated sequence is stationary. Some examples are provided to illustrate our theoretical results.

High-Energy Physics experiments are facing a multi-fold data increase with every new iteration. This is certainly the case for the upcoming High-Luminosity LHC upgrade. Such increased data processing requirements forces revisions to almost every step of the data processing pipeline. One such step in need of an overhaul is the task of particle track reconstruction, a.k.a., tracking. A Machine Learning-assisted solution is expected to provide significant improvements, since the most time-consuming step in tracking is the assignment of hits to particles or track candidates. This is the topic of this paper. We take inspiration from large language models. As such, we consider two approaches: the prediction of the next word in a sentence (next hit point in a track), as well as the one-shot prediction of all hits within an event. In an extensive design effort, we have experimented with three models based on the Transformer architecture and one model based on the U-Net architecture, performing track association predictions for collision event hit points. In our evaluation, we consider a spectrum of simple to complex representations of the problem, eliminating designs with lower metrics early on. We report extensive results, covering both prediction accuracy (score) and computational performance. We have made use of the REDVID simulation framework, as well as reductions applied to the TrackML data set, to compose five data sets from simple to complex, for our experiments. The results highlight distinct advantages among different designs in terms of prediction accuracy and computational performance, demonstrating the efficiency of our methodology. Most importantly, the results show the viability of a one-shot encoder-classifier based Transformer solution as a practical approach for the task of tracking.

Additive manufacturing (AM; 3D printing) has clear benefits in the production of lightweight mirrors for astronomy: it can create optimised lightweight structures and combine multiple components into one. New capabilities in AM ceramics, silicon carbide infiltrated with silicon and fused silica, offer the possibility to combine the design benefits of AM with a material suitable for visible, ultraviolet and X-ray applications. This paper will introduce the printing methods and post-processing steps to convert AM ceramic samples into reflective mirrors. Surface roughness measurements after abrasive polishing of the AM ceramics will be presented., Comment: 15 pages, 15 figures, submitted to SPIE Astronomical Telescopes & Instrumentation, Advances in Optical and Mechanical Technologies for Telescopes and Instrumentation VI (Conference 13100, Paper 123)

Track reconstruction is a vital aspect of High-Energy Physics (HEP) and plays a critical role in major experiments. In this study, we delve into unexplored avenues for particle track reconstruction and hit clustering. Firstly, we enhance the algorithmic design effort by utilising a simplified simulator (REDVID) to generate training data that is specifically composed for simplicity. We demonstrate the effectiveness of this data in guiding the development of optimal network architectures. Additionally, we investigate the application of image segmentation networks for this task, exploring their potential for accurate track reconstruction. Moreover, we approach the task from a different perspective by treating it as a hit sequence to track sequence translation problem. Specifically, we explore the utilisation of Transformer architectures for tracking purposes. Our preliminary findings are covered in detail. By considering this novel approach, we aim to uncover new insights and potential advancements in track reconstruction. This research sheds light on previously unexplored methods and provides valuable insights for the field of particle track reconstruction and hit clustering in HEP.

This work is motivated by the interest in determining the effect of the micro-anatomy of the spinal subarachnoid space on the cerebrospinal fluid flow and on the associated transport of solutes. To that aim, we focus on a canonical model problem in which a circular post of radius $a$, confined between two parallel plates separated by a distance $2h$, is subjected to an oscillatory flow. In particular, we are concerned with the steady, viscous, time-averaged flow that persists in the vicinity on the cylinder, when the stroke length of the oscillating flow, $U_{\infty}/\omega$ is smaller or comparable to the radius of the post, $a$, where $U_{\infty}$ and $\omega$ are the mean velocity amplitude and the corresponding angular frequency, respectively. First, we analyze the asymptotic limit of small values of the stroke length varying the aspect ratio of the post $\lambda$ and the Womersley number $M$. First-order steady-streaming corrections at leading order have been computed, together with associated Stokes-drift component, the sum of both yielding the mean Lagrangian velocity field. The confinement effect is seen to induce the three-dimensionality of flow. Moreover, in the mid plane the time-averaged steady flow exhibits a recirculating vortex attached to the wall of the post that decreases as $M$ increases for low values of $M$. However, for values of $M$ larger than a critical one, $M_{cr}(\lambda)$, a second, outer vortex is also formed. The dependence $M_{cr}(\lambda)$ has been quantified. The analysis has been corroborated experimentally for a fixed aspect ratio. Consideration of an anharmonic oscillating flow shows the fort-and-aft symmetry of the steady flow is broken. Finally, the analysis is experimentally extended to consider an array of equally spaced posts., Comment: 24 pages, 13 figures

A simplified model of an initially excited oscillator as a quantum system interacting with a large number of oscillators acting as a reservoir has been developed in this work. All these oscillators are in their ground state uncoupled each other and at the limit of the weak coupling between the system and the reservoir. This system could be an oscillator excited in a microcavity that interacts with the vacuum's electromagnetic field at zero temperature. This work's primary goal is to obtain the system's density matrix's exact solution in these conditions. The general approach calculates all oscillators' evolution as a single isolated entity using the evolution operator. Starting from a total initial state that can be factored between the system and the reservoir, the evolution is unitary, and the partial trace is taken in all the degrees of freedom of the environment to obtain the density matrix of the system at any instant of time; this procedure requires diagonalizing Hamiltonian. The results are evaluated for a reservoir of N=1000 oscillators, particular values of the coupling force, and ohmic order of the spectral density, contrasted with the corresponding Markovian solution described in section [2.3.1]., Comment: Acknowledgment to my advisor Dr. Hernan Castillo Egoavil. Master's thesis, in Spanish language, http://hdl.handle.net/20.500.12404/19564

In this work was researched the problems of classicalization and measurement unifying them in a single problem: the problem of collapse, for this purpose was developed a working program -- the classicalization's program -- in an interpretation of objective collapse that was proposed for this purpose, and which overcomes the limitations of the official interpretation. This program explain that the classicization arises by alternating the unitary evolution and the collapses, due to the redistribution of energy, so the classical mechanics emerges as a good approximation; however, was defined deep chaos as the regime where macrosystems are so sensitive to the initial conditions that are affected by the collapses and indeterminism of quantum mechanics, allowing the emergence of thermodynamics with its irreversibility; at this level the observation occurs because the measurement device (classic) extends their collapses at microsystem through quantum entanglement. In this work, was did an extension of the classicalization's program at the black holes, and was explained -- at a level still in development -- about the role of the collapses in its thermodynamic; in addition, the program of quantum decoherence was critized, and was presented a theory of physical information that was developing., Comment: Bachelor's thesis, in Spanish language. https://hdl.handle.net/20.500.12952/3539

We give extensive empirical evidence against the common belief that variational learning is ineffective for large neural networks. We show that an optimizer called Improved Variational Online Newton (IVON) consistently matches or outperforms Adam for training large networks such as GPT-2 and ResNets from scratch. IVON's computational costs are nearly identical to Adam but its predictive uncertainty is better. We show several new use cases of IVON where we improve finetuning and model merging in Large Language Models, accurately predict generalization error, and faithfully estimate sensitivity to data. We find overwhelming evidence that variational learning is effective., Comment: Published at International Conference on Machine Learning (ICML), 2024. The first two authors contributed equally. Code is available here: https://github.com/team-approx-bayes/ivon

Intermolecular interactions are crucial in determining the morphology of solution-processed semiconducting polymer thin films. However, these random interactions often lead to disordered or short-range ordered structures. Achieving long-range order in these films has been a challenge due to limited control over microscopic interactions in current techniques. Here, we present a molecular-level methodology that leverages spatial matching of intermolecular dynamics among solutes, solvents, and substrates to induce directional molecular assembly in weakly bonded polymers. Within the optimized dynamic scale of 2.5 \r{A} between polymer side chains and self-assembled monolayers (SAMs) on nanogrooved substrates, our approach transforms random aggregates into unidirectional fibers with a remarkable increase in the anisotropic stacking ratio from 1 to 11. The Flory-Huggins-based molecular stacking model accurately predicts the transitioning order on various SAMs, validated by morphologic and spectroscopic observations. The enhanced structural ordering spans over 3 orders of magnitude in length, raising from the smallest 7.3 nm random crystallites to >14 um unidirectional fibers on sub-millimeter areas. Overall, this study provides insights into the control of complex intermolecular interactions and offers enhanced molecular-level controllability in solution-based processes., Comment: 40 pages, 6 figures, 2 tables

The trajectories of a single bubble rising in the vicinity of a vertical solid wall are experimentally investigated. Distinct initial wall-bubble distances are considered for three different bubble rising regimes, i.e. rectilinear, planar zigzag, and spiral. The problem is defined by three control parameters, namely the Galilei number, $Ga$, the Bond number, $Bo$, and the initial dimensionless distance between the bubble centroid and the wall, $L$. We focus on high-Bond numbers, varying $L$ from 1 to 4, and compare the results with the corresponding unbounded case, $L \rightarrow \infty$. In all cases, the bubble deviates from the expected unbounded trajectory and migrates away from the wall as it rises due to the overpressure generated in the gap between the bubble and the wall. This repulsion is more evident as the initial wall-bubble distance decreases. Moreover, in the planar zigzagging regime, the wall is found to impose a preferential zigzagging plane perpendicular to it when $L$ is small enough. Only slight wall effects are observed in the velocity or the oscillation amplitude and frequency. The wall migration effect is more evident for the planar zigzagging case and less relevant for the rectilinear one. Finally, the influence of the vertical position of the wall is also investigated. When the wall is not present upon release, the bubbles have the expected behavior for the unbounded case and experience the migration only instants before reaching the wall edge. This repulsion is, in general, more substantial than in the initially-present-wall case.

A Q-matrix is a binary matrix that defines the relationship between items and latent variables and is widely used in diagnostic classification models (DCMs), and can also be adopted in multidimensional item response theory (MIRT) models. The construction process of the Q-matrix is typically carried out by experts in the subject area of the items and statistical procedures can be used to verify its suitability. In DCMs, different approaches have been proposed for validating the Q-matrix through iterative algorithms. This article presents a method of empirical Q-matrix validation for MIRT models. A simulation study and an application to real data of morphology skills of elementary students are conducted to examine the viability of the method. Relevant issues regarding the implementation of the method and the results obtained are discussed. [This paper will be published in "Applied Measurement in Education."]

The role of type Ia supernovae (SN Ia), mainly the Delay Time Distributions (DTDs) determined by the binary systems, and the yields of elements created by different explosion mechanisms, are studied by using the MulChem chemical evolution model applied to our Galaxy. We explored 15 DTDs and 12 tables of elemental yields produced by different SN Ia explosion mechanisms, doing a total of 180 models. Chemical abundances for $\alpha$-elements (O, Mg, Si, Ca) and Fe derived from these models are compared with recent solar region observational data of $\alpha$-elements over Fe relative abundances, [X/Fe], as a function of [Fe/H] and age. A multidimensional maximum likelihood analysis shows that 52 models are able to fit all these datasets simultaneously, considering the 1$\sigma$ confidence level. The combination of STROLG1 DTD and LN20181 SN Ia yields provides the best fit. The exponential model with very prompt events is a possible DTD, but a combination of several channels is more probable. The SN Ia yields that include MCh or Near MCh correspond to 39 (75%) of the 52 best models. Regarding the DTD, 31 (60%) of the 52 most probable models correspond to the SD scenario, while the remaining 21 (40%) are based on the DD scenario. Our results also show that the relatively large dispersion of the observational data may be explained by the stellar migration from other radial regions, and/or perhaps a combination of DTDs and explosion channels., Comment: 26 pages, 28 figures. Accepted for publication in MNRAS

We investigate the linear Floquet stability of two fluid layers undergoing oscillations in the direction parallel to the flexible wall that separates them. This canonical configuration is inspired by the cerebrospinal fluid flow in the spinal canal of subjects with hydro-/syringomyelia.The analysis focuses on the marginal conditions for the onset of instability, and how these depend on the spatial wavelength of the perturbation, and on the values of the control parameters, which are the two channel widths, the Reynolds number, and the wall stiffness. Unstable perturbations are found to oscillate synchronous with the base flow. The wavelength of the most unstable perturbation, of the order of the stroke length of the basic oscillatory motion, depends strongly on the wall stiffness, but is only weakly influenced by the channel widths and the Reynolds number. In general, around criticality, it was found that increasing the Reynolds number has a destabilizing effect, and that decreasing the canal widths stabilizes the instability. The wall stiffness on the other hand has a non-monotonic effect, exhibiting an intermediate value for which the instability is maximally amplified. The present analysis is a first step towards a better understanding of the physical mechanisms that govern many (bio)fluid mechanical problems that involve oscillatory flows near compliant walls.

Inspired by certain regularization techniques for linear inverse problems, in this work we investigate the convergence properties of the Levenberg-Marquardt method using singular scaling matrices. Under a completeness condition, we show that the method is well-defined and establish its local quadratic convergence under an error bound assumption. We also prove that the search directions are gradient-related allowing us to show that limit points of the sequence generated by a line-search version of the method are stationary for the sum-of-squares function. The usefulness of the method is illustrated with some examples of parameter identification in heat conduction problems for which specific singular scaling matrices can be used to improve the quality of approximate solutions.

Due to its proximity, the stellar populations of the Galactic bulge (GB) can be resolved and can be studied in detail. This allows tracing the bulge metallicity distribution function (MDF) for different spatial regions within the bulge, which may give us clues about the bulge formation and evolution scenarios. In this work, we developed a chemical evolution model (CEM), taking into account the mass distribution in the bulge and disc, to derive the radial dependence of this time-scale in the Galaxy. Since the infall rate depends on that time scale in the CEM, the results of the model were used to test a scenario where the bulge is formed inside-out. The obtained results for the $[\alpha/\mbox{Fe}]$ vs. [Fe/H] relationship, the MDF and the [Fe/H] radial gradient in the bulge have been compared to available data in the literature. The model is able to reproduce most of the observational data: the spread in the relation $[\alpha/\mbox{Fe}]$ vs. [Fe/H], the MDF shape in different regions of the bulge, the [Fe/H] radial gradient inside it and the age-metallicity relation, as well as the [$\alpha$/Fe] evolution with age. The results of the model point to a scenario where the bulk of the bulge stars pre-existed the boxy/peanut X-shape bar formation. As a result, the classical origin of the GB is not ruled out and this scenario may be invoked to explain the chemical properties of the Galactic bulge., Comment: 16 pages, 11 figures. Accepted for publication in MNRAS