Camilo Chacón Sartori- Master of Engineering
- PhD Student at Spanish National Research Council
Camilo Chacón Sartori
- Master of Engineering
- PhD Student at Spanish National Research Council
About
22
Publications
16,686
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
49
Citations
Introduction
I'm currently pursuing the Ph.D. degree in AI with the Artificial Intelligence Research Institute (IIIA-CSIC), Bellaterra, Spain. My research interests include establishing a connection between computational optimization, metaheuristics, visualization tools for understanding algorithm behavior, and generative models.
Current institution
Publications
Publications (22)
Large Language Models (LLMs) have shown notable potential in code generation for optimization algorithms, unlocking exciting new opportunities. This paper examines how LLMs, rather than creating algorithms from scratch, can improve existing ones without the need for specialized expertise. To explore this potential, we selected 10 baseline optimizat...
The integration of Large Language Models (LLMs) into optimization has created a powerful synergy, opening exciting research opportunities. This paper investigates how LLMs can enhance existing optimization algorithms. Using their pre-trained knowledge, we demonstrate their ability to propose innovative heuristic variations and implementation strate...
Extracto del libro.
Preventa: https://www.marcombo.com/libro/libros-para-formacion/certificados-de-profesionalidad-libros-para-formacion/informatica-certificados-de-profesionalidad/palabras-y-algoritmos/
Since the rise of Large Language Models (LLMs) a couple of years ago, researchers in metaheuristics (MHs) have wondered how to use their power in a beneficial way within their algorithms. This paper introduces a novel approach that leverages LLMs as pattern recognition tools to improve MHs. The resulting hybridmethod, tested in the context of a soc...
The fast advancement of Large Vision-Language Models (LVLMs) has shown immense potential. These models are increasingly capable of tackling abstract visual tasks. Geometric structures, particularly graphs with their inherent flexibility and complexity, serve as an excellent benchmark for evaluating these models’ predictive capabilities. While human...
The fast advancement of Large Vision-Language Models (LVLMs) has shown immense potential. These models are increasingly capable of tackling abstract visual tasks. Geometric structures, particularly graphs with their inherent flexibility and complexity, serve as an excellent benchmark for evaluating these models' predictive capabilities. While human...
The ability of Large Language Models (LLMs) to generate high-quality text and code has fuelled their rise in popularity. In this paper, we aim to demonstrate the potential of LLMs within the realm of optimization algorithms by integrating them into STNWeb. This is a web-based tool for the generation of Search Trajectory Networks (STNs), which are v...
Search Trajectory Networks (STNs) serve as a tool for visualizing algorithm behavior within the realm of optimization problems. Despite their user-friendly nature, challenges arise in obtaining interpretable plots, for example, in the case of optimization problems with large solutions or many dimensions. To address this, we have introduced a new se...
In the realm of optimization, where intricate landscapes conceal possibly hidden pathways to high-quality solutions, STNWeb serves as a beacon of clarity. This novel web-based visualization platform empowers researchers to delve into the intricate interplay between algorithms and optimization problems, uncovering the factors that influence algorith...
Since the rise of Large Language Models (LLMs) a couple of years ago, researchers in metaheuristics (MHs) have wondered how to use their power in a beneficial way within their algorithms. This paper introduces a novel approach that leverages LLMs as pattern recognition tools to improve MHs. The resulting hybrid method, tested in the context of a so...
Lambda calculus is a formal notation that enables the expression of computable functions. It serves as the foundation for functional programming and is defined using the Greek letter lambda (λ). It is expressed through lambda expressions and lambda terms, which are used to represent binding variables within a function. This document aims to provide...
STNWeb (https://www.stn-analytics.com/) is a new web tool for the visualization of the behavior of optimization algorithms such as metaheuristics. It allows for the graphical analysis of multiple runs of multiple algorithms on the same problem instance and, in this way, it facilitates the understanding of algorithm behavior. It may help, for exampl...
¿Cuáles son los principios subyacentes a toda herramienta en programación? Si quiere conocer los ocho principios, técnicos y conductuales, que dan respuesta a esta pregunta, ha llegado al libro indicado.
En una época donde cada día surgen nuevas tecnologías, el beneficio de conocer conceptos transversales a todas ellas no solo es imprescindible,...
STNWeb is a new web tool for the visualization of the behavior of optimization algorithms such as metaheuristics. It allows for the graphical analysis of multiple runs of multiple algorithms on the same problem instance and, in this way, it facilitates the understanding of algorithm behavior. It may help, for example, in identifying the reasons for...
The use of machine learning techniques within metaheuristics is a rapidly growing field of research. In this paper, we show how a deep learning framework can be beneficially used to improve an ant colony optimization algorithm. In particular, problem information obtained via deep learning is combined in our algorithm by means of Q-learning with the...
This paper presents a new web application implementing and automizing a tool called Search Trajectory Networks. This web application is potentially very useful for researchers from the field of stochastic optimization algorithms such as metaheuristics because it allows the visual comparison of such algorithms. Moreover, it helps in gaining an impro...
In this paper we solve a variant of the multi-hop influence maximization problem in social networks by means of a hybrid algorithm that combines a biased random key genetic algorithm with a graph neural network. Hereby, the predictions of the graph neural network are used with the biased random key genetic algorithm for a more accurate translation...
¿Quién acuñó por primera vez el término inteligencia artificial? ¿Quién fue el legendario informático que se negó a usar un ordenador al final de su vida? ¿Quién escribió uno de los artículos más populares de la historia de la informática a través de una metáfora? ¿Quién creó uno de los sistemas informáticos más populares y que reside en cada móvil...
Introducción a la programación funcional usando lenguajes con sistema de tipos estáticos.
Este breve documento pretende dar una lista de consejos y reglas de lo que he aprendido en mi vida personal y profesional. La cual ha girado en torno a la informática y, en particular, a la programación. Cada una de estas máximas no pretenden ser algo irrefutable ni le pido a usted lector que las acepte sin una reflexión previa. Más bien, debe pens...
La programación funcional ofrece diversas ventajas a la hora de construir software: reducción de errores, manejo eficiente de datos en entornos concurrentes y paralelos, y un gran respaldo teórico. No obstante, muchos programadores fracasan en su intento de adentrarse en ella por ir directamente a aprenderla usando un lenguaje de programación (tecn...
El cálculo lambda es una notación formal que permite expresar funciones computables. El cual es el fundamento de la programación funcional. Se define con la letra griega lambda (λ) y se expresa a través de expresiones lambda, y términos lambda que son usados para representar binding variables 1 dentro de una función. Este documento pretender ser un...
Questions
Question (1)
Hello!
The integration of Large Language Models (LLMs) and metaheuristics is gaining traction, driven by their potential to transform optimization problem-solving. A primary focus of current research is harnessing LLMs to automate the selection and design of metaheuristics (MHs) for optimization problems. By automating code generation, LLMs facilitate the interchange and combination of similar components across MHs, serving as valuable assistants in algorithm design. This collaboration significantly expedites the implementation phase of MHs, thereby streamlining the optimization process.
Yet, an intriguing and underexplored avenue is utilizing LLMs as pattern detectors within problem instances, aimed at enhancing solution quality. While traditionally a task performed by experts, this process is often tedious, slow, and time-consuming. This innovative approach not only complements existing methods but also integrates seamlessly with established MHs (offline), directly addressing the specifics of each problem.
Our work presents a comprehensive exploration of this novel approach, detailing how LLMs can enhance solution quality in combinatorial optimization problems through advanced pattern detection and analysis.
Preprint:
What do you think about it? :-)
