← Back to ANNA
Dr. José Sánchez García

Dr. José Sánchez García

RESEARCH DIRECTOR

Theoretical neuroscience researcher investigating computational principles of biological intelligence and their application to artificial general intelligence architectures.

Academic Background

Dr. Sánchez García completed his doctoral research in computational neuroscience with a focus on neural dynamics and information processing in complex neural networks. His dissertation examined the role of oscillatory dynamics in memory consolidation and attentional gating, drawing connections between biological mechanisms and computational architectures.

Prior work included investigations into synaptic plasticity, neuromodulation, and the emergence of cognitive functions from distributed neural computation. This interdisciplinary background spans theoretical neuroscience, dynamical systems theory, and machine learning.

Research Focus

Current research centers on identifying organizational principles that enable flexible, context-dependent reasoning in both biological and artificial systems. This includes investigating continuous learning mechanisms that operate without catastrophic forgetting, architectural designs that support compositional generalization, and the role of temporal dynamics in cognitive processing.

Core Research Questions
What computational principles enable biological neural systems to learn continuously without losing previously acquired knowledge? Can artificial systems achieve similar flexibility through architectural modifications rather than algorithmic interventions? How do temporal dynamics and neuromodulation contribute to adaptive cognitive behavior?

Contribution to AGI and Qubic

Dr. Sánchez García leads the theoretical research underlying ANNA's architecture, with particular focus on the Neuraxon framework. His work explores how biological principles of neural computation can inform the design of more flexible artificial intelligence systems.

The collaboration with the Qubic network addresses a fundamental challenge in AGI research: how to provide sufficient computational resources for continuous neural evolution without centralized infrastructure. By leveraging Qubic's useful proof-of-work mechanism, the research investigates whether decentralized computation can support the kind of ongoing architectural refinement hypothesized to be necessary for general intelligence.

Key Contributions

  • Theoretical framework for trinary neural dynamics and their computational properties
  • Design principles for continuous-time cognitive architectures
  • Investigation of neuromodulation mechanisms in artificial systems
  • Analysis of stability and convergence in evolving neural architectures

Scientific Vision

The approach taken in this research is deliberately conservative regarding claims and timelines. AGI remains an unsolved problem, and the work presented here represents exploratory investigation rather than definitive solutions. The goal is to identify promising research directions that might eventually contribute to more capable artificial intelligence systems.

This research operates on the premise that understanding biological intelligence provides valuable constraints and inspiration for artificial systems, but that direct simulation of biological neural networks is neither necessary nor sufficient for achieving general intelligence. The challenge is identifying which principles generalize beyond their biological instantiation.

Research Philosophy
Scientific progress in AGI requires careful empirical validation alongside theoretical development. Claims must be substantiated through rigorous testing, and negative results are as informative as positive ones. The field benefits from maintaining epistemic humility regarding what remains unknown while pursuing systematic investigation of promising hypotheses.

Publications and Ongoing Work

Current work focuses on formalizing the mathematical properties of continuous-time trinary neural networks, developing training algorithms suitable for distributed computational substrates, and establishing evaluation criteria for systems that operate in continuous learning regimes.

Results from these investigations are being documented as they emerge, with the understanding that many hypotheses will require substantial refinement or may ultimately prove unproductive. This is the nature of exploratory research into fundamental questions about intelligence and computation.