While current approaches focus on "one-model-fits-all", larger models, and more data, we need architectures that can evolve, adapt, and reason across contexts while maintaining reliability and efficiency. This is where bio-symbolic composite AI offers a promising path forward, constrained by nature's proven principles rather than brute-force computation. In addition to their flexibilty, these architectures enable the integration of third-party frontier foundational models, which can offer scientific ground for experimentation in academic labs and small private actors.

Neural-symbolic AI bridges network learning with symbolic reasoning, combining the pattern recognition capabilities of neural networks with the interpretability of logical rules. At Daice Labs, we believe that bio-symbolic composite AI takes this integration further by incorporating nature's broader computational principles—from cellular information processing to evolutionary adaptation (Figure 2).

While neural-symbolic systems mirror brain-like computation, bio-symbolic architectures draw from the entire spectrum of biological intelligence, offering a richer template for resilient and adaptive systems. These architectures aren't just brain-inspired but nature-optimized, constraining system design to key evolutionary-like principles. By encoding nature's computational strategies—from cellular processing units to environmental adaptation and collective intelligence—into symbolic frameworks, we unlock new possibilities for contextually-aware AI systems that evolve and specialize like natural organisms.

Dynamic systems are characterized by constant change, interdependence, and the need to adapt to shifting conditions/environments. These systems, whether natural or human-made, are shaped by complex interactions among their components, often resulting in unpredictable behaviors and emergent phenomena. Financial systems, digital commerce platforms, and biological systems exemplify such dynamics, sharing core traits like high interconnectivity, continuous evolution, and adaptive behavior. Their inherent complexity arises from many interacting variables, often producing emergent phenomena that are difficult to predict with traditional models.

Therefore the need for innovative frameworks that understand and modulate their complex behavior.

The convergence of artificial intelligence and biological systems creates a powerful cycle of discovery and innovation. Our approach implements a variation of neurosymbolic AI, bio-symbolic composite AI, grounded in key adaptability and evolutionary principles, to navigate complex system challenges. The bidirectional exchange between AI and bioengineering (Figure 3) drives breakthroughs in both fields, leading to practical applications across dynamic domains (see our publications). We strategically focus on digital, financial, and cellular systems as ideal environments for testing and refining context-aware evolution, where our technology demonstrate its capability to adapt and learn frollowing key natural principles.