Experimental research on designed representations for reasoning and decision-making in complex systems.
Aukemir is an independent experimental research project investigating whether deliberately structured representations can make important structure in complex quantitative problems easier for humans to inspect and reason about.
Our long-term direction is simple to state and difficult to earn:
Expand what humans can perceive, understand, and do.
Our first research program tests whether deliberately structured representations can improve human inference and intervention reasoning beyond competent conventional reporting.
The experimental program is designed to distinguish genuine learning and transfer from memorization, superficial cues, explicit rules, conventional visualization, and direct machine recommendation.
The current system has passed an internal synthetic reproducibility and held-out technical evaluation program.
External computational reproduction and human decision-value testing remain pending.
| Evidence layer | Current status |
|---|---|
| Research hypothesis | Active and falsifiable |
| Internal technical architecture | Built |
| Internal synthetic evaluation | Completed |
| Held-out technical evaluation | Completed internally |
| External computational reproduction | Pending |
| Human validation | Not yet begun |
| Human learning evidence | Not demonstrated |
| Far-transfer evidence | Not demonstrated |
| New functional human capability | Not demonstrated |
| Biological workflow validation | Not demonstrated |
| Commercial validation | Open |
Internal technical success is not human evidence.
Aukemir does not currently claim that learning, far transfer, counterfactual improvement, cognitive efficiency, retention, biological efficacy, or a new human capability has been demonstrated.
This GitHub organization provides selected, evidence-bounded public provenance for the Aukemir research program.
Public materials may include:
- research questions
- evidence states
- milestone records
- selected non-sensitive research documentation
- public scientific outputs
- future non-core reproducibility tools that clear disclosure review
Core mechanism details, exact algorithms, benchmark internals, answer-bearing structures, participant-facing experimental materials, and private reproducibility packages are not publicly released at the current stage.
Aukemir is built around:
- explicit hypotheses
- competing explanations
- strong controls
- versioned artifacts
- predefined evidence thresholds
- negative-result preservation
- kill criteria
Ambitious questions. Explicit falsification. Evidence before claims.
Aukemir was founded and is led by Isabella Salcedo Tuiran, medical researcher and scientific founder.
- GitHub: @salcedoisabella
- LinkedIn: Isabella Salcedo Tuiran
Public materials are released selectively and do not imply external validation, human efficacy, product validation, or commercial validation.
