-
Program Synthesis using Inductive Logic Programming for the Abstraction and Reasoning Corpus
Authors:
Filipe Marinho Rocha,
Inês Dutra,
Vítor Santos Costa
Abstract:
The Abstraction and Reasoning Corpus (ARC) is a general artificial intelligence benchmark that is currently unsolvable by any Machine Learning method, including Large Language Models (LLMs). It demands strong generalization and reasoning capabilities which are known to be weaknesses of Neural Network based systems. In this work, we propose a Program Synthesis system that uses Inductive Logic Progr…
▽ More
The Abstraction and Reasoning Corpus (ARC) is a general artificial intelligence benchmark that is currently unsolvable by any Machine Learning method, including Large Language Models (LLMs). It demands strong generalization and reasoning capabilities which are known to be weaknesses of Neural Network based systems. In this work, we propose a Program Synthesis system that uses Inductive Logic Programming (ILP), a branch of Symbolic AI, to solve ARC. We have manually defined a simple Domain Specific Language (DSL) that corresponds to a small set of object-centric abstractions relevant to ARC. This is the Background Knowledge used by ILP to create Logic Programs that provide reasoning capabilities to our system. The full system is capable of generalize to unseen tasks, since ILP can create Logic Program(s) from few examples, in the case of ARC: pairs of Input-Output grids examples for each task. These Logic Programs are able to generate Objects present in the Output grid and the combination of these can form a complete program that transforms an Input grid into an Output grid. We randomly chose some tasks from ARC that dont require more than the small number of the Object primitives we implemented and show that given only these, our system can solve tasks that require each, such different reasoning.
△ Less
Submitted 10 May, 2024;
originally announced May 2024.
-
Segmented zero-inflated Poisson mixed effects model with random changepoint
Authors:
Paulo Dourado,
Antonio C. Pedroso-de-Lima,
Francisco M. M. Rocha
Abstract:
The COVID-19 pandemic has had a substantial impact on hospital services, as many institutions have observed a surge in healthcare-associated infections (HAIs) despite heightened adherence to isolation protocols and hand hygiene. According to the World Health Organization (WHO), HAIs are among the leading causes of mortality and morbidity of hospitalized patients. This study aims to examine the eff…
▽ More
The COVID-19 pandemic has had a substantial impact on hospital services, as many institutions have observed a surge in healthcare-associated infections (HAIs) despite heightened adherence to isolation protocols and hand hygiene. According to the World Health Organization (WHO), HAIs are among the leading causes of mortality and morbidity of hospitalized patients. This study aims to examine the effect of the COVID-19 pandemic on the incidence of central venous catheter-related bloodstream infections (CR-BSIs) of hospitals in the city of São Paulo. Initially we considered segmented zero-inflated Poisson (ZIP) mixed-effects models with known changepoint, which can be estimated applying the standard framework of ZIP mixed-effects models. However, we found that the changepoint could occur at varying times across different hospitals. We present an effective iterative procedure to estimate segmented ZIP mixed-effects models with random changepoints in a likelihood-based framework. The suggested procedure is a practical approach employing conventional computational tools for estimating standard mixed-effects zero-inflated Poisson (ZIP) models. Prior to its implementation to the CR-BSI data, simulation studies were conducted to examine the accuracy of the estimation under various scenarios.
△ Less
Submitted 10 October, 2023; v1 submitted 2 October, 2023;
originally announced October 2023.
-
Controlling fluid adhesion force with electric fields
Authors:
Pedro H. A. Anjos,
Francisco M. Rocha,
Eduardo O. Dias
Abstract:
Develo** adhesives whose bond strength can be externally manipulated is a topic of considerable interest for practical and scientific purposes. In this work, we propose a method of controlling the adhesion force of a regular fluid, such as water and/or glycerol, confined between two parallel plates by applying an external electric field. Our results show the possibility of enhancing or diminishi…
▽ More
Develo** adhesives whose bond strength can be externally manipulated is a topic of considerable interest for practical and scientific purposes. In this work, we propose a method of controlling the adhesion force of a regular fluid, such as water and/or glycerol, confined between two parallel plates by applying an external electric field. Our results show the possibility of enhancing or diminishing the bond strength of the liquid sample by appropriately tuning the intensity and direction of the electric current generated by the applied electric field. Furthermore, we verify that, for a given direction of the electric current, the adhesion force can be reduced enough for the fluid to lose its adhesive properties and begin exerting a force to move apart the confining plates. In these circumstances, we obtain an analytical expression for the minimum electric current required to detach the plates without requiring the action of an external force.
△ Less
Submitted 10 September, 2022;
originally announced September 2022.
-
Diversity of self-propulsion speeds reduces motility-induced clustering in confined active matter
Authors:
Pablo de Castro,
Francisco M. Rocha,
Saulo Diles,
Rodrigo Soto,
Peter Sollich
Abstract:
Self-propelled swimmers such as bacteria agglomerate into clusters as a result of their persistent motion. In 1D, those clusters do not coalesce macroscopically and the stationary cluster size distribution (CSD) takes an exponential form. We develop a minimal lattice model for active particles in narrow channels to study how clustering is affected by the interplay between self-propulsion speed div…
▽ More
Self-propelled swimmers such as bacteria agglomerate into clusters as a result of their persistent motion. In 1D, those clusters do not coalesce macroscopically and the stationary cluster size distribution (CSD) takes an exponential form. We develop a minimal lattice model for active particles in narrow channels to study how clustering is affected by the interplay between self-propulsion speed diversity and confinement. A mixture of run-and-tumble particles with a distribution of self-propulsion speeds is simulated in 1D. Particles can swap positions at rates proportional to their relative self-propulsion speed. Without swap**, we find that the average cluster size $L_\text{c}$ decreases with diversity and follows a non-arithmetic power mean of the single-component $L_\text{c}$'s, unlike the case of tumbling-rate diversity previously studied. Effectively, the mixture is thus equivalent to a system of identical particles whose self-propulsion speed is the harmonic mean self-propulsion speed of the mixture. With swap**, particles escape more quickly from clusters. As a consequence, $L_\text{c}$ decreases with swap** rates and depends less strongly on diversity. We derive a dynamical equilibrium theory for the CSDs of binary and fully polydisperse systems. Similarly to the clustering behaviour of one-component models, our qualitative results for mixtures are expected to be universal across active matter. Using literature experimental values for the self-propulsion speed diversity of unicellular swimmers known as choanoflagellates, which naturally differentiate into slower and faster cells, we predict that the error in estimating their $L_\text{c}$ via one-component models which use the conventional arithmetic mean self-propulsion speed is around $30\%$.
△ Less
Submitted 16 September, 2021; v1 submitted 8 July, 2021;
originally announced July 2021.
