Research Laboratory
About TEYA SOL LAB
An independent research laboratory investigating how information is encoded in the physical world. We study the spectral signatures of bounded systems and develop computational methods to reconstruct internal structure from external measurements.
Origin
How It Started
I was fascinated by Tesla's stories about resonance - the idea that matching a system's frequency could amplify energy seemed like a hidden key to reality.
But when I tried to explain what resonance actually is - mathematically - I couldn't. I once attempted to describe it at a meetup. Someone corrected me. Someone else laughed.
Instead of defending myself, I decided to understand it properly. I bought a soldering iron, opened a linear algebra textbook, and started rebuilding from zero.
Then I noticed something: the same mathematics - eigenvalues, operators, entropy, optimization - keeps appearing in springs, circuits, black holes, and neural networks.
This lab is my attempt to trace that idea wherever it leads - and to build the tools and models that come from truly understanding the math that connects different domains.
Mission
To understand and formalize the relationship between the internal structure of physical systems and their observable spectral behavior - and to build the computational tools that make this knowledge actionable.
TEYA SOL LAB operates at the intersection of physics, information theory, and machine learning. We believe that the signals emitted, reflected, and transmitted by physical systems carry deep structural information that can be decoded systematically. Our work spans from theoretical foundations to hands-on laboratory experiments with real hardware.
Axioms
Core Principles
Principle 01
Every physical system has a spectral signature.
Any bounded physical system, when probed with energy, produces a response whose frequency content encodes the internal structure of that system. This spectral fingerprint is unique, measurable, and information-rich.
Principle 02
Information can be extracted from boundary measurements.
The observable response at a system's boundary contains sufficient information to reconstruct aspects of its internal arrangement. Non-invasive exterior measurements carry structural data.
Principle 03
Machine learning can reconstruct internal structure from signal data.
Modern representation learning and inverse-problem techniques can map from spectral observations back to structural descriptions, solving the reconstruction problem computationally.
Program
Research Lines
Three interconnected research directions exploring how the same mathematical structures appear across physics, information theory, and machine learning.
Cross-Domain Mathematics
The same operators, eigenvalues, and entropy measures appear in springs, circuits, black holes, and neural networks. We map these isomorphisms and extract transferable methods.
Spectral Structure Analysis
Using Fourier analysis, spectral entropy, and drift detection to extract structural information from signals - whether mechanical vibrations, acoustic responses, or neural network weight spectra.
Computational Reconstruction
Building ML models that solve the inverse problem: reconstructing internal structure from external measurements. From physics-informed neural networks to cross-domain similarity scoring.
Domains
Research Focus Areas
Signal Acquisition
Hardware and software systems for capturing high-fidelity boundary measurements from physical systems.
Spectral Analysis
Fourier, wavelet, and empirical mode decomposition methods for extracting spectral features from raw signals.
Information Theory
Entropy analysis, information encoding, and holographic-principle-inspired approaches to boundary information.
Machine Learning
Neural networks and statistical models trained to solve the inverse problem of structure-from-signal.
Physical Experiments
Laboratory experiments with mechanical, acoustic, and electromagnetic systems to validate theoretical predictions.
Reconstruction Algorithms
Algorithms that map spectral data to structural descriptions, from simple parameter estimation to full 3D reconstruction.
Infrastructure
Technology Stack
Frontend
Backend
ML Services
Data
Infrastructure
Get in Touch
Contact
Have a question about our research or want to collaborate? Send us a message.