Synthetic Data Generation
When does generated data add information a model can use, and when does it add bias or noise?
Open questions
- Which variations matter for a task?
- How large is the synthetic-to-real gap?
- When does synthetic data stop helping?
Synthex Research & Innovation
Exploring synthetic data generation, virtual-world simulation, annotation quality and reproducible evaluation for real-world AI applications.
Research directions and methods. No Synthex experimental results are published yet; every claim is traceable in our evidence register.
Research themes
Each theme pairs a scientific question with an engineering one. Answers are not assumed; they are what the research is for.
When does generated data add information a model can use, and when does it add bias or noise?
Open questions
How should scenarios be specified so they are representative, repeatable and comparable?
Open questions
How can label errors be found and measured before they distort training and evaluation?
Open questions
How can evaluation reveal where and why models fail, not just how often?
Open questions
Featured research projects
Each project shows its objective, approach and actual evidence status. None is labelled complete, because no completed results are documented yet.
Medical imaging
Evidence: Conceptual · R-006Agriculture
Evidence: Conceptual · R-007Autonomous systems
Evidence: Conceptual · R-009Research lifecycle
Every project follows the same lifecycle, so results can be compared, repeated and challenged.
STEP 01
A specific, falsifiable question with a defined success criterion.
STEP 02
Licensed data and scenario definitions, with provenance recorded.
STEP 03
Baselines, controls, splits and metrics fixed before results are seen.
STEP 04
Versioned generation or preprocessing with recorded parameters.
STEP 05
Automated checks and expert review of labels.
STEP 06
Fixed configurations and documented random seeds.
STEP 07
Per-condition results, error analysis and negative findings.
STEP 08
Re-run from the manifest; confirm results before any claim.
Decided before results
Technical demonstrations
Educational, frontend-only demonstrations built from static, original fixtures. They show how research questions are examined — not research results.
Publications and outputs
This section lists only real, approved outputs. None have been published yet, so it describes the artefacts each project is designed to produce.
No research outputs have been published yet.
Methods, results and limitations for each completed study, including negative findings.
Peer-reviewed or preprint papers where work is mature enough to be scrutinised.
Licensed synthetic or derived datasets with documentation and dataset cards.
Code and configurations that re-create a published result from its manifest.
Versioned scenario suites with fixed protocols and comparable results.
Experiment manifests, metric definitions and data governance notes.
Evidence and provenance
Every research statement on this site is recorded in an evidence register with its source, verification class and publication decision.
| ID | Claim | Status | Public decision |
|---|---|---|---|
| R-001 | Synthex is developing synthetic data, simulation, annotation and evaluation infrastructure for physical-world AI. Evidence: Conceptual Published as a statement of intent and direction, never as an available product. | Evidence: Conceptual | Published as a statement of intent and direction, never as an available product. |
| R-002 | LIDC-IDRI is a public thoracic CT collection with radiologist lesion annotations, hosted by The Cancer Imaging Archive. Evidence: Verified Cited as public research context with attribution. No images from the collection are reproduced. Public source: cancerimagingarchive.net | Evidence: Verified | Cited as public research context with attribution. No images from the collection are reproduced. |
| R-003 | Synthex lung CT lesion segmentation results (Dice, IoU, precision, recall or sensitivity values). Evidence: Unavailable Not published. Pages show the comparison protocol with 'no verified result' cells. | Evidence: Unavailable | Not published. Pages show the comparison protocol with 'no verified result' cells. |
| R-004 | Prior academic or dissertation work on lung CT segmentation and synthetic augmentation. Evidence: Unavailable Not described as Synthex research and not referenced on public pages. | Evidence: Unavailable | Not described as Synthex research and not referenced on public pages. |
| R-005 | Synthetic augmentation improves lesion segmentation performance. Evidence: Unavailable Presented only as an open research question, including the possibility that it does not help. | Evidence: Unavailable | Presented only as an open research question, including the possibility that it does not help. |
| R-006 | Medical imaging research workflow: preprocessing, masks, augmentation, training, metric evaluation, error analysis. Evidence: Conceptual Published as proposed research methodology, not clinical deployment. | Evidence: Conceptual | Published as proposed research methodology, not clinical deployment. |
| R-007 | Synthetic crop and weed dataset research for agricultural perception. Evidence: Conceptual Published as a research direction with schematic illustrations. | Evidence: Conceptual | Published as a research direction with schematic illustrations. |
| R-008 | Agricultural segmentation benchmark values (IoU, Dice, precision, recall by condition). Evidence: Unavailable Not published. | Evidence: Unavailable | Not published. |
| R-009 | Simulation-driven autonomous perception research (scenarios, synthetic sensor data, robustness evaluation). Evidence: Conceptual Published as a research direction. | Evidence: Conceptual | Published as a research direction. |
| R-010 | Integration with CARLA, Unreal Engine, Unity, AirSim or another simulator. Evidence: Unavailable Simulators described only as technologies under exploration; no integration or partnership claimed. | Evidence: Unavailable | Simulators described only as technologies under exploration; no integration or partnership claimed. |
| R-011 | CARLA is an open-source simulator for autonomous driving research, built on a fork of Unreal Engine; code MIT, assets CC-BY. Evidence: Verified Described factually as a candidate research tool; no affiliation implied. Public source: carla.org | Evidence: Verified | Described factually as a candidate research tool; no affiliation implied. |
| R-012 | Descriptions of Unity, AirSim and Gazebo used on Sprint 2–3 pages. Evidence: Unavailable Retained as generic descriptions; flagged for owner verification. Research pages name only CARLA and Unreal Engine. | Evidence: Unavailable | Retained as generic descriptions; flagged for owner verification. Research pages name only CARLA and Unreal Engine. |
| R-013 | Synthex publications, technical reports or preprints. Evidence: Unavailable No entries displayed; output types described as intended artefacts. No scholarly structured data. | Evidence: Unavailable | No entries displayed; output types described as intended artefacts. No scholarly structured data. |
| R-014 | Open datasets, reproducible notebooks or public benchmark results from Synthex. Evidence: Unavailable Not displayed. | Evidence: Unavailable | Not displayed. |
| R-015 | Interactive research demonstrations use static, original fixtures (schematic drawings and descriptive text). Evidence: Verified Published with 'Illustrative · static fixture' labels. No backend, generation or inference. | Evidence: Verified | Published with 'Illustrative · static fixture' labels. No backend, generation or inference. |
| R-016 | Definitions of Dice, IoU, precision, recall and sensitivity. Evidence: Verified Published as definitions only, with the caveat that technical metrics do not establish clinical effectiveness. | Evidence: Verified | Published as definitions only, with the caveat that technical metrics do not establish clinical effectiveness. |
| R-017 | Research partners, collaborators, customers or institutional affiliations. Evidence: Unavailable None claimed. Collaboration is offered, not asserted. | Evidence: Unavailable | None claimed. Collaboration is offered, not asserted. |
| R-018 | Regulatory approval, clinical validation, medical device certification or safety certification. Evidence: Unavailable Explicitly disclaimed on medical and autonomy pages. | Evidence: Unavailable | Explicitly disclaimed on medical and autonomy pages. |
| R-019 | Data Engine copy stating the direction was 'informed by work with public research datasets such as LIDC-IDRI'. Evidence: Unavailable Reworded in Sprint 4: the wording implied prior work for which no evidence exists in the repository. | Evidence: Unavailable | Reworded in Sprint 4: the wording implied prior work for which no evidence exists in the repository. |
| R-020 | No private research assets, patient data, credentials or proprietary datasets are published by the website. Evidence: Verified Guard retained in the e2e suite. | Evidence: Verified | Guard retained in the e2e suite. |
Collaborate
We are looking for collaborators who care about rigorous, reproducible evidence as much as we do.