About This Project

We are prototyping a VO2 Symbolic Resonance Array, a hybrid analog-digital dev kit for sustainable computing. VO2 pillars interact through an excite, relax, read cycle, and the readout is quantized into symbol vectors. We will deliver a VO2 resonance tile and a compact controller with a simple Python API, release a reproducible model, and measure stability, vector error rate, symbol rate, and energy per cycle. Success would show a low-power path for AI and general computing at the edge.

Ask the Scientists

What is the context of this research?

Artificial intelligence is advancing quickly, but the energy and cost of computation keep rising, especially for on-device systems where heat and battery life are limiting. This project explores a different path: a brain-inspired Symbolic Resonance Array (SRA). The SRA uses phase-transition materials such as VO₂ arranged in symmetric arrays. Through an excite, relax, read cycle the device produces symbol vectors that can feed conventional software or small models, aiming to deliver useful computation at much lower energy.

We will build a first VO₂ test module and open tools to evaluate stability, vector error rate, symbol rate, and energy per cycle. The goal is evidence, not hype: can a compact, low-power resonance front end reduce downstream compute for sensing, control, compression, and AI at the edge? If successful, the results will outline a practical path toward more sustainable AI and general computing.

What is the significance of this project?

This project could reshape the global conversation on AI by offering an alternative to faster and larger digital machines. The framework I propose is: Safe, designed to reduce risks of unintended harm; Preventive, anticipating problems before they escalate; Resonant, able to align with human meaning and empathy; Compatible, designed to integrate with human values and systems; Economical, requiring minimal cost and energy; Green, sustainable and environmentally friendly; Analog, using natural material dynamics instead of energy-hungry digital computation; and Scalable, adaptable from small prototypes to global applications. The SRA will help guide AI toward sustainability and safety so humanity benefits from technology rather than fears it.

What are the goals of the project?

Digital computing keeps growing while edge energy and heat budgets are fixed. This project will provide reproducible evidence for a different path: a VO2-based Symbolic Resonance Array (SRA) that converts device dynamics into symbol vectors.

We will answer:

• What is the energy per cycle, and how does it trade with symbol rate?

• What vector error rate and information throughput are achievable?

• How stable are symbols across temperature and 10^4–10^5 cycles?

• How do coupling and array size affect performance?

Deliverables: an open model, a dataset, and a small dev kit with a Python API. These benchmarks enable fair comparison with memristors, oscillatory networks, and other mixed-signal front ends. If metrics meet practical thresholds (for example, sub-millijoule per 10^3 vectors at low error), SRA can reduce downstream compute for sensing, control, compression, and AI. Clear results, positive or negative, will guide sustainable computing.