Skip to content

neuraparse/qantis

Repository files navigation

QANTIS

Quantum Autonomous Navigation, Tracking & Intelligence System

A quantum-native platform for next-generation autonomous systems.

Edition Status License Python Built by


Community Edition Notice

You are looking at the public, community-facing distribution of QANTIS. This repository is intentionally a basic edition — it ships a clean, didactic surface of the platform so that researchers, students, and partners can understand the architecture and integrate at the API level.

The full QANTIS platform, including production-grade modules, advanced optimisation backends, hardened mitigation pipelines, internal tooling, and the complete experimental harness, is maintained in a private workspace reserved for Neura Parse collaborators and contracted partners.

Real benchmark runs, hardware campaign results, comparative studies, and superior performance figures obtained on production backends are not published in this repository. Selected outcomes are disclosed only through peer-reviewed publications, partner briefings, and formal engagements with Neura Parse Ltd.


About

QANTIS is a quantum-native platform designed by Neura Parse Ltd for systems that must make decisions under uncertainty — drone swarms, GPS-denied navigation, multi-target surveillance, computational life sciences, and intelligent decision making in adversarial environments.

The platform is organised as three layers: a framework at the foundation, a decision engine at the core, and a portfolio of purpose-built applications running on top.

Positioning. Deterministic optimisers solve deterministic models. QANTIS optimises uncertain belief-to-action loops — the harder online problem in which noisy observations must become calibrated beliefs, calibrated risk estimates, feasible actions, and verifiable decisions under a fixed compute budget.


The QANTIS Decision Engine

The Decision Engine is the product-level core of QANTIS. It exposes a small, opinionated surface that turns raw observations into trustworthy decisions through four complementary capabilities. The Community Edition exposes a basic surface of this engine; the calibrated, production-grade variant lives in the Collaborator Edition.

Infer

Calibrated belief

Turn noisy, partial, or rare observations into a calibrated posterior belief — the single source of truth that every downstream step depends on.

Risk

Event & tail-risk

Estimate the probability of operationally critical events and tail-risk quantities directly from belief, with confidence intervals suitable for safety-aware decision making.

Optimise

Feasible decisions

Produce decisions that respect hard constraints by construction, rather than relying on penalty terms that can silently violate feasibility under stress.

Verify

Trust & diagnostics

Provide feasibility checks, repair, posterior fidelity, and closed-loop diagnostics so that every decision is accompanied by a defensible quality report.

This four-part decomposition is the throughline of QANTIS: every application in the platform is, ultimately, a specialisation of Infer → Risk → Optimise → Verify.


The Platform

QANTIS is delivered as a layered platform. The framework provides the infrastructure; the decision engine provides the core capabilities; the applications specialise those capabilities for specific operational domains.

Framework

Quantum Common

The foundation layer. Provides the unified backend abstraction, configuration system, error-mitigation pipeline, benchmarking, and reproducibility infrastructure that every QANTIS module is built on.

POMDP Application

Decision Making Under Uncertainty

Belief-state planning for autonomous agents operating with partial information. Targets navigation, mission planning, and policy execution in environments where perception is incomplete or unreliable.

MHT Application

Multi-Target Tracking

Multi-hypothesis tracking for drone swarm surveillance and dense multi-target scenarios in which classical association becomes intractable at scale.

The full QANTIS platform extends well beyond autonomy. Additional applications — including Quantum-Bio Intelligence for computational life sciences, a dedicated CRISPR application for quantum-assisted genome design and target analysis, alongside modules covering sensor fusion, adversarial robustness, and mission-level orchestration — are developed inside the private collaborator workspace and are not part of this Community Edition.


Editions

QANTIS is distributed in two editions. This repository corresponds to the first row.

Community Edition (this repository) Collaborator Edition (private)
Audience Public, academic, evaluation Neura Parse partners & collaborators
Framework core Included, basic surface Full production framework
Decision Engine Basic surface (Infer / Risk / Optimise / Verify) Calibrated, production-grade engine
Applications Reference subset (POMDP, MHT) Complete suite — incl. Quantum-Bio Intelligence, CRISPR, and more
Backends Standard public connectors Hardened, optimised, multi-vendor
Error mitigation Baseline pipeline Full mitigation & calibration stack
Benchmarks & datasets Illustrative only Full experimental harness
Real hardware results Not published here Reserved for partners
Support Community, best-effort Dedicated engineering support

Access to the Collaborator Edition is granted on a case-by-case basis through formal engagement with Neura Parse Ltd.


Design Principles

  • Framework-first. A single, well-defined core powers every QANTIS application. No application owns its own infrastructure.
  • Backend-agnostic. Applications are written once and dispatched across simulators, gate-based hardware, and quantum annealers through a uniform interface.
  • Reproducibility by construction. Every run is configuration-driven, versioned, and traceable.
  • Incremental adoption. Heavy quantum SDKs are optional; the framework installs cleanly with a minimal scientific Python stack.
  • Hybrid by default. Quantum routines are designed to interoperate with classical baselines, not to replace them.

Getting Started

git clone https://github.com/neuraparse/qantis.git
cd qantis
uv sync

Optional quantum backends and visualisation extras are available as installable groups. See the in-package documentation for the relevant extras for your environment.

The Community Edition is intended for architectural exploration, integration testing, and educational use. It is not a substitute for the Collaborator Edition in production or research-grade evaluation contexts.


What Is Not Included

To set expectations clearly, the following are deliberately outside the scope of this Community Edition:

  • Production-grade application variants and internal QANTIS modules
  • Calibrated, hardware-tuned mitigation and optimisation pipelines
  • Real measurement data, raw hardware traces, and campaign artefacts
  • Comparative benchmarks and performance figures vs. classical state-of-the-art
  • Confidential datasets, scenarios, and mission profiles developed with partners
  • Any results that constitute commercial or research advantage for Neura Parse Ltd

These remain proprietary to the Collaborator Edition. Public communication of selected results occurs through formal channels only.


Research Artefacts & Reviewer Access

This Community Edition serves as the public companion repository for QANTIS-related publications by Neura Parse Ltd. Where a peer-reviewed paper references this repository, the intent is to point readers to the open architecture, framework surface, and reference implementations that accompany the work — not to redistribute the underlying experimental campaigns.

Category Availability
Framework core and application reference code Included in this repository
Reference scenarios and illustrative examples Included in this repository
Raw hardware measurements, run identifiers, calibration windows, transpile manifests Archived by Neura Parse Ltd under the conference artefact policy and provided to programme committees and accredited reviewers on request
Aggregated benchmark figures and end-to-end campaign results Reserved for the Collaborator Edition and formal partner channels
Confidential datasets, scenario libraries, and mission profiles Not distributed

Reviewers and programme committees evaluating QANTIS-related submissions may request artefact access through the corresponding author or directly through Neura Parse Ltd. Access is granted under standard conference artefact-evaluation conditions, including time-limited review periods and confidentiality undertakings where applicable.

This separation is deliberate: the public surface is stable, citable, and easy to inspect; the experimental surface is governed under research-integrity and partner-confidentiality requirements and is therefore handled out-of-band.


Roadmap (Public Track)

The Community Edition follows a separate, slower release track than the internal platform. Public milestones focus on:

  • Stabilising the framework API surface
  • Expanding didactic examples and integration guides
  • Publishing reference notebooks aligned with peer-reviewed outputs
  • Hardening installation and onboarding for new contributors

Items on the internal roadmap — including production applications, advanced backends, and benchmark releases — are governed independently and are not announced here.


Contributing

Community contributions are welcome for documentation, examples, integration fixes, and onboarding improvements. Contributions affecting the framework core or any QANTIS application are reviewed against the internal platform and may be deferred or adapted to maintain compatibility with the Collaborator Edition.

For substantial collaboration, joint research, or evaluation under the Collaborator Edition, please contact Neura Parse Ltd directly.


Citation

If you reference QANTIS in academic or technical work, please cite the official Neura Parse publications associated with the platform. Citing this repository alone is not sufficient to attribute results obtained on the full platform.


License

Released under the MIT License. See LICENSE for details.

The MIT license applies to the source code in this Community Edition repository. It does not extend to internal QANTIS modules, datasets, calibration data, or experimental results maintained outside this repository.


QANTIS — engineered by Neura Parse Ltd.

Quantum technology for autonomous systems that have to work in the real world.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors