Code

You can visit my GitHub profile to explore additional projects, as well as materials from tutorials, courses, and workshops.

• SimpleMC
• nnogada
• neuralike
• Cosmo Reconstructions
• NCosmoVAE

---

SimpleMC

Cosmological parameter inference toolkit for robust parameter estimation and sampling.

Cosmological parameter estimation code originally developed by Dr. A. Slosar and Dr. J. A. Vázquez. Between 2019 and 2023, I contributed to maintenance and new features, including nested sampling, Metropolis–Hastings convergence criteria, post-processing utilities, and additional analysis options.

Links

• Library GitHub repository: ja-vazquez/SimpleMC
• Documentation: igomezv/SimpleMC/Docs
• Workshop/tutorial: igomezv/simplemc_workshop

---

nnogada

Neural networks optimized with genetic algorithms for flexible, data-driven inference.

Neural Networks Optimized by Genetic Algorithms in Data Analysis (nnogada) is a framework that combines neural networks with genetic algorithms for efficient modeling, reconstruction, and parameter inference.

Links

• Library GitHub repository: igomezv/nnogada
• Documentation: docs/nnogada
• Related: PRD paper with the method.

Projects using nnogada

• igomezv/Reconstructing-RC-with-ANN
• igomezv/LSST_DE_neural_reconstruction
• igomezv/neuralike

---

neuralike

Surrogate-assisted Bayesian inference to accelerate cosmological likelihood evaluations.

Deep learning and genetic-algorithm techniques designed to speed up Bayesian inference in cosmology, particularly within sampling pipelines where likelihood evaluations dominate the cost.

Links

• Library GitHub repository: igomezv/neuralike
• Using neuralike within SimpleMC using nested sampling from dynestylibrary: igomezv/simplemc_tests
• Related: PRD paper with the method.

---

ANN Reconstructions

Model-independent cosmological reconstructions with uncertainty-aware neural networks.

Python implementations for model-independent reconstructions of cosmological functions using artificial neural networks, incorporating uncertainty quantification and hyperparameter optimization.

Repositories

• [In progress] Library to reproduce our previous works: igomezv/alp

• Reconstructing rotation curves with Monte Carlo Dropout and gentetic algorithsm, through our nnogada code, for hyperparameter optimization: igomezv/Reconstructing-RC-with-ANN
  Related: PDU paper.

• Reconstructing SNeIa from LSST simulations with Monte Carlo Dropout and genetic algorihtms for hyperparameter optimization with our nnogada code: igomezv/LSST_DE_neural_reconstruction
  Related: PDU paper.

• Reconstructing cosmological functions with Monte Carlo Dropout and grid hyperparameter optimization: igomezv/neuralCosmoReconstruction
  Related: EPJC paper.

---

NCosmoVAE

Variational autoencoders for fast generative modeling of cosmological N-body simulations.

Variational autoencoder framework trained on N-body cosmological simulations to generate realistic dark matter halo realizations. Hyperparameter optimization with genetic algorithms using the optuna Python library.

Links

• Library repository: igomezv/NcosmoVAE
• Related: PRD paper.