Materials Science

Φ-M (phi-M) is a deterministic platform for materials discovery. It transforms materials into precise mathematical objects and places them into a fixed coordinate system. Unlike conventional AI-driven approaches, Φ-M does not rely on learned embeddings, black-box models, or brute-force enumeration. Instead, it defines a stable, navigable geometry of materials based on explicit structure, enabling discovery to proceed with precision rather than trial and error.

Within this coordinate system, known materials occupy structured regions of information space, and unexplored regions correspond to physically realizable but undiscovered materials. By generating candidate compositions explicitly and encoding them into the Φ-M map, discovery becomes a problem of guided navigation rather than prediction. This allows Φ-M to operate effectively in low-data regimes, produce exactly the number of candidates required, including a single candidate, and provide a clear geometric justification for every result.

Φ-M is designed to operate upstream of simulation and machine learning, not as a replacement for them. Density functional theory, experimental validation, and AI-based tools can be applied downstream as focused validators once high-value regions of the map are identified. By separating representation and navigation from modeling and optimization, Φ-M establishes a foundational discovery layer that scales across materials classes and enables a fundamentally new paradigm for how novel materials are identified, evaluated, and ultimately realized.