# Quantum Wave (QWAVE)

Project brief · September 2026

## Purpose

Quantum Wave is a proposed research task network for neural quantum states and variational Monte Carlo Researchers define computational questions and budgets; contributors deliver computational work; independent evaluators inspect and re-evaluate submissions

The research pilot connects EVM wallet sign-in, task records, private artifact uploads, public submissions and authorized reviewer assessments

Computations run externally and no token utility or settlement contract is connected

## Research direction

The initial scope is electronic structure at fixed nuclear geometries Neural networks represent trial wavefunctions, and variational Monte Carlo estimates their energy This work uses classical computing hardware; quantum computing hardware is not required for the first milestone

The first benchmark is neutral singlet H2 at a fixed bond length of 1·4 bohr

A real PySCF RHF and finite-basis FCI reference has been executed and recomputed locally with matching results
Both runs used the same machine and software so they do not establish independent scientific validation
A 1000-step FermiNet GPU training pilot and a same-environment repeat have completed on an RTX 4090
Model checkpoints and raw training artifacts are available — independent fixed-model evaluation remains pending

## Research workflow

- Define the system, budget and acceptance criteria
- Train a neural wavefunction and retain model checkpoints, configuration, logs and execution environment details
- Independently re-evaluate the submitted model using fresh samples
- Report energy estimates, statistical uncertainty, sampling diagnostics and reproducibility limitations
- Publish an assessment with traceable supporting artifacts

Independent evaluation provides evidence under stated assumptions; it is not a guarantee of an exact ground state or a trustless proof of scientific validity

## Network design

Scientific computation and evaluation take place off-chain Funding records, submission references and settlement are planned for Robinhood Chain, subject to implementation and testing

Researchers define tasks Compute contributors submit work Independent evaluators assess it The community may participate in decisions about research budgets; scientific conclusions remain evidence-based

## Proposed token role

The project name is Quantum Wave, its abbreviation is QWAVE, and its proposed token symbol is QWAVE (displayed as $QWAVE)

Planned uses include contribution incentives and participation in community research budgets Reward pools must be funded and acceptance rules published Task payments and ecosystem token rewards are accounted for separately Later milestones may explore proposal and performance bonds with explicit dispute and refund rules

No token contract is published on this website These utilities are not live Token ownership does not automatically confer intellectual property rights, equity or guaranteed returns

## Implementation direction

- Scientific baseline: Python, PySCF, NumPy and SciPy with RHF and finite-basis FCI
- Neural pilot: FermiNet and JAX with pinned source commits and a tested GPU dependency lock — independent evaluation pending
- Evaluation artifacts: model checkpoints, configuration, logs, sampling diagnostics and independently produced reports
- Current website: semantic HTML, responsive CSS and vanilla JavaScript; an original Canvas 2D particle illustration
- Hosting: Vercel
- Research platform: Supabase Auth, Postgres with row-level security, private Storage and an administrator access gateway
- Wallet interface: EVM message signatures with OKX support and viem address validation
- Input validation: Zod and database-side checks
- Future development: independently validated neural evaluation, compute orchestration and reviewed settlement contracts

The homepage particle field and molecular drawing are conceptual illustrations, not calculated or measured results

## Milestones

- Complete the project specification and website design
- Reproduce the hydrogen baseline and document the environment
- Demonstrate independent evaluation and publish artifacts
- Pilot task funding and settlement after technical and security review
- Expand research scope based on benchmark evidence and expert review

## Website language

All public website content, navigation, dialogs and downloadable project materials are English-only

## Research reference

FermiNet: https://github.com/google-deepmind/ferminet

Upstream tools are research references, not project partners or endorsements