By using data from Brazilian school meals, It elaborates a portfolio of food menus with a tradeoff among nutrition indicators, the variety of components, and the prioritization of organics and regional products. 

Generation of improved menus from Brazilian school meal data by multiobjective optimization and autoencoders

large_Recomendação de cardápios escolares - versao 2_Img0.png

medium_Recomendação de cardápios escolares - versao 2_Img0.png

small_Recomendação de cardápios escolares - versao 2_Img0.png

thumbnail_Recomendação de cardápios escolares - versao 2_Img0.png

Recomendação de cardápios escolares - versao 2_Img0.png

Food-Menus

Siflor4Decision is a tool for the analysis and visualization of models generated by the DAMICORE methodology from the SiFlor Cerrado system.

Multicriteria decision-making tool  from SiFlor-Cerrado System

large_DALL·E 2024-09-16 10.56.27 - A forest scene with sunlight streaming through the trees, similar to the previous image, but instead of a Bayesian structure model, there is a large p.webp

medium_DALL·E 2024-09-16 10.56.27 - A forest scene with sunlight streaming through the trees, similar to the previous image, but instead of a Bayesian structure model, there is a large p.webp

small_DALL·E 2024-09-16 10.56.27 - A forest scene with sunlight streaming through the trees, similar to the previous image, but instead of a Bayesian structure model, there is a large p.webp

thumbnail_DALL·E 2024-09-16 10.56.27 - A forest scene with sunlight streaming through the trees, similar to the previous image, but instead of a Bayesian structure model, there is a large p.webp

DALL·E 2024-09-16 10.56.27 - A forest scene with sunlight streaming through the trees, similar to the previous image, but instead of a Bayesian structure model, there is a large p.webp

Siflor4Decision

Innovative tool that uses FAIR quality metrics, Openness, and data maturity to automatically assess the quality and transparency of databases through the attribution of scores and detailed reports.

Automatically evaluate databases quality and transparency.

medium_WEB-EVALUATE42.jpg

small_WEB-EVALUATE42.jpg

thumbnail_WEB-EVALUATE42.jpg

WEB-EVALUATE42.jpg

DBScore

Cellular Neural Networks (CelNN) for Evolutionary Partial Differential Equations enable dealing with complex (chaotic) systems from Nature, producing a kind of unsupervised Physics-Informed Neural Network.

Recent advances in Cellular Neural Networks (CelNN) for Evolutionary Partial Differential Equations provide solutions to complex and chaotic systems found in Nature

medium_Screenshot 2024-09-12 at 13.02.34.png

small_Screenshot 2024-09-12 at 13.02.34.png

thumbnail_Screenshot 2024-09-12 at 13.02.34.png

Screenshot 2024-09-12 at 13.02.34.png

CelNN for Nature modeling

Development of an automatic calibration strategy for flood models using multiobjective evolutionary optimization algorithms to enrich watershed data. Models can be knowledge-based, data-driven, and hybrid. 

Spatially-informed multiobjective optimization for integrating multi-source heterogeneous data and flood models

large_emergency-4314503_1920.png

medium_emergency-4314503_1920.png

small_emergency-4314503_1920.png

thumbnail_emergency-4314503_1920.png

emergency-4314503_1920.png

Hetero-flood

This work focuses on Heterogeneous Data System Models (MsHD) for integrating complex, multidisciplinary data, often using geospatial data and choropleth maps to simplify analysis. Automating the creation of these maps through machine learning can improve decision-making, especially in critical situations. The MultiMapas system supports this process with six key modules, including geolocation, data aggregation, thematic layers, decision-making, visualization, and data repository.

Multiobjective generator of choropleth maps from heterogeneous multisource data

medium_Screenshot 2024-09-12 at 20.54.42.png

small_Screenshot 2024-09-12 at 20.54.42.png

thumbnail_Screenshot 2024-09-12 at 20.54.42.png

Screenshot 2024-09-12 at 20.54.42.png

Multimaps

Constructing elementary cell of Cellular Neural Networks (CelNN) in an IBM physical quantum computer. An elementary cell works as a Full Adder for digital computers. CelNN are among a dozen candidates for quantum NN.

CellNN has high locality and cells that can use low amounts of qubits, which may enable them in a Noise-Intermediate State Quantum (NISQ) device. 

medium_Screenshot 2024-09-12 at 09.51.50.png

small_Screenshot 2024-09-12 at 09.51.50.png

thumbnail_Screenshot 2024-09-12 at 09.51.50.png

Screenshot 2024-09-12 at 09.51.50.png

CelNN in a quantum device

The current core (of Almes) focuses on providing a detailed description of agricultural price index datasets.
The Almes Core Reference Guide defines metadata and provides structured data with vocabulary schemes like Agrovoc and Agrotermos. 

Agriculture and Livestock Metadata Elements Set., fundamental for a Confident and Broad AI.

large_Screenshot 2024-09-12 at 21.36.10.png

medium_Screenshot 2024-09-12 at 21.36.10.png

small_Screenshot 2024-09-12 at 21.36.10.png

thumbnail_Screenshot 2024-09-12 at 21.36.10.png

Screenshot 2024-09-12 at 21.36.10.png

Almes-core

An ontology of types (or a classification ontology), meaning it does not contain individuals and is intended as a classification system for agricultural commodities price index databases in Brazil. 

Projects

Demo Outreach of MClimate and AgriBio