Home
Alex Baker · Research Documentation
Technical and scientific reference for my astrophysics, astro-ML, and computational physics research projects.
Technical library · scientific reference · reproducibility layer
One place to understand,reproduce, and extend my research
This home page is the editorial entry point to three open research programs: stellar spectroscopy, galaxy morphology, and Bayesian cosmology. It is designed to help visitors quickly understand what exists here, how the projects connect, and where to begin.
Projects
Three research programs
Each project has its own scientific identity, documentation logic, and repository surface.
AstroSpectro
A physics-first research program for stellar spectral classification, feature engineering, interpretability, and latent-structure analysis from LAMOST DR5 spectra.
- LAMOST DR5 × Gaia DR3 ecosystem
- Interpretability, validation, and reproducibility
- Dimensionality reduction as scientific analysis
AstroVision
A multimodal deep-learning project for galaxy morphology, uncertainty-aware classification, and the physical interpretation of visual representations.
- Galaxy imaging and morphology
- DINOv2, morphometrics, and multimodal fusion
- A bridge toward cross-modal astro-AI
ξ Dark Energy
A reproducible Bayesian inference pipeline comparing dark-energy model families with supernovae and BAO measurements in an explicitly exploratory framework.
- ΛCDM, CPL, and oscillating dark-energy models
- Transparent inference assumptions and caveats
- Reproducible MCMC workflow
Navigation
Where to start
Choose an entry point based on what you need from the documentation today.
I’m new here
Start with the ecosystem map, understand how the three projects relate, then choose a project overview.
Open Research EcosystemI want results first
Jump directly into the strongest scientific directions, current findings, and high-level project takeaways.
Start with AstroSpectroI want to reproduce something
Use the guided reading paths and section roots to move quickly from architecture to implementation and practical workflows.
See Reading PathsI need concepts and definitions
Use the atlas and section overviews to recover terminology, structure, and cross-project reference material.
Browse the AtlasArchitecture
How this site is organised
The ecosystem is intentionally split into three complementary layers.
Personal site
Narrative hub
Who I am, why these projects exist, and how they fit into my broader academic and research trajectory.
Go there ↗This documentation site
Technical reference
How the projects work, what they contain, how they are organised, and where to start for methods, data, results, and reproducibility.
Go thereGitHub repositories
Project entry points
Code, releases, notebooks, readmes, and repository-level context for each research program.
Go there ↗Why this documentation exists
Scientific directions worth opening first
This site is not only a code manual. It is a scientific reference built around method, validation, interpretation, and reproducibility.
AstroSpectro
The central direction is not only performance, but understanding what physically meaningful structure modern models learn from stellar spectra.
Explore the projectAstroVision
AstroVision extends the ecosystem into galaxy imaging, uncertainty-aware classification, and multimodal visual representations with astrophysical meaning.
Open AstroVision docsξ Dark Energy
The cosmology branch emphasizes explicit assumptions, reproducible inference, and careful documentation of what remains exploratory or unresolved.
Open ξ Dark Energy docsBrowse by type
Use the site like a library, not just a sidebar
These shortcuts help visitors move across projects by intent, not only by folder structure.
Research ecosystem
See how the three projects relate intellectually, scientifically, and technically.
Open sectionReading paths
Choose a path by persona: supervisor, reproducer, student, or future Alex.
Open sectionData & feature reference
Jump into structured dataset and feature documentation, starting from AstroSpectro.
Open sectionAtlas
Use the cross-project reference layer for terms, maps, and high-level navigation.
Open section