Question → Target
Target Board
Every internal research target with the evidence it has, the range that evidence covers and the check that would change its reading. A target is written before evaluation and is not a result.
Structure-Based AI Small-Molecule Discovery
Compare structural evidence, experimentally measured binding values and predicted properties for therapeutic targets and previously reported compounds.
- Internal target
- Binding free-energy error of 1 kcal/mol or less
- Evidence so far
- Common synthetic checks
- What the evidence covers
- The computational benchmark covers scores and chemical metrics. Patent novelty, synthesis, toxicity and therapeutic effects each require their own evidence.
- What would change it
- Select records with measured binding values, specify assay conditions and lock the split by scaffold, target family and date.
Early Boost for Core Health R&D Technology
Align variables and clinical time series by schema, then assess predictive utility and privacy protection together.
- Internal target
- At least 95% of downstream predictive performance retained
- Evidence so far
- Synthetic harness (synth8) · Common synthetic checks
- What the evidence covers
- TCGA, GEO and MIMIC-IV are separate sources. Validation follows each source's endpoint, access terms and verified linkage.
- What would change it
- Run the same harness on the fixed TCGA-BRCA institution split with the baseline generator, then train the main generator and require it to beat the baseline on every endpoint at the same privacy budget.
Medication Administration and Dispensing Safety
Define reference cases for pill identification and prescription review, covering lighting, occlusion, look-alike products and unit conversion errors.
- Internal target
- Zero missed contraindications or look-alike medication errors
- Evidence so far
- Synthetic harness (dispense18) · Common synthetic checks
- What the evidence covers
- Report a zero miss rate with the frozen test denominator. Product codes and source rules require verification across medical settings.
- What would change it
- Train the vision modules on the fixed ePillID split, build per-category calibration sets from its validation split and fill the same table on its test split; categories below the minimum calibration size are reported as uncertified with no automatic pass.
Rare Disease Diagnostic Research
Rank candidate genes using variants, frequencies and symptom codes. Report conflicting evidence, uncertainty and reference versions.
- Internal target
- At least 90% of causal genes in the top five
- Evidence so far
- Published-case evaluation · Public-record audit (acmg19) · Synthetic replay (rare19) · Common synthetic checks
- What the evidence covers
- Evaluation can overstate performance when inputs already include the variant annotation or gene–disease association used as the reference answer.
- What would change it
- Read the literature for every expert-pathogenic missense variant instead of a sample, so that functional, segregation and case findings enter the ledger. Then re-measure sensitivity on genes outside the current 40, with the condition that specificity does not fall.
Non-Pharmacological Dementia Research
Simulate response delay and fatigue during memory, attention and executive-function tasks and inspect responses to difficulty transitions and stop criteria.
- Internal target
- Difficulty convergence in at least 95% of simulated episodes
- Evidence so far
- Synthetic harness (cog16) · Common synthetic checks
- What the evidence covers
- Convergence in simulation evaluates the behaviour of the training policy. Clinical cognitive improvement requires different evidence.
- What would change it
- Import stratum-level summary statistics inside the data-use zone, re-freeze the fixture and re-run; then test a response generator whose functional form differs from the scheduler's model.
Multimodal Dementia Diagnostic Research
Evaluate MRI and PET features along with pauses, vocabulary and sentence features from speech. Compare fixed models on independent cohorts.
- Internal target
- External cohort AUROC of at least 0.90
- Evidence so far
- Synthetic harness (cog16) · Common synthetic checks
- What the evidence covers
- Single-modality evaluation is available pending imaging and speech matched to the same participant and diagnostic visit.
- What would change it
- After approved data import, run the same harness with ADNI and OASIS-3 as the two external directions and evaluate speech separately; the synthetic figures are then replaced.
Fall Prediction and Prevention
Evaluate a warning policy on timestamped sensor streams. Separate detection of falls from prediction before impact.
- Internal target
- At least 92% early-warning sensitivity in the high-risk group
- Evidence so far
- Common synthetic checks
- What the evidence covers
- A laboratory fall label and an advance alert for a patient are different endpoints. Measure lead time only for records that cover at least three seconds before impact.
- What would change it
- Register participant and event splits, alert lead time and the time denominator for false alarms.
Surgical Environment Error Prevention
Evaluate tracking within the visible field. Use fixed sequences with occlusion to test arrival and removal counts.
- Internal target
- Zero instrument count omissions on the fixed test
- Evidence so far
- Common synthetic checks
- What the evidence covers
- In-frame presence and segmentation cover a different observation space from total instrument inventories. Report the synthetic missed-count denominator.
- What would change it
- Lock visibility, occlusion and tracking ID criteria and define separate video and synthetic count metrics.
Suicide Prevention Research for People with Mental Illness
Align event markers and note excerpts by time in offline research. Test for information leakage after the reference time and missed signals across subgroups using controlled examples.
- Internal target
- At least 90% recall for acute crisis signals
- Evidence so far
- Synthetic harness (prevent12) · Common synthetic checks
- What the evidence covers
- Recall in actual records requires a predefined population, observation window, timestamped independent reference and an eligible positive denominator.
- What would change it
- Use structured event markers for synthetic evaluation. Define the requirements for longitudinal labels in external evaluation.
Suicide Prevention Research for Bereaved People and Attempt Survivors
Evaluate events and transition rules in synthetic follow-up timelines. Verify that constrained responses stay within the designated state machine.
- Internal target
- Zero clinically disallowed response events on the fixed test
- Evidence so far
- Synthetic harness (prevent12) · Common synthetic checks
- What the evidence covers
- Transition behaviour and prevention after discharge require different follow-up evidence. The source and access terms for the follow-up cohort remain unverified.
- What would change it
- Lock the synthetic transition table and denominator. Define the independent outcomes required for further follow-up.
Social Isolation and Suicide Prevention Research
Use aggregate features for prolonged changes in social connection and count false alerts during ordinary interruptions.
- Internal target
- At least 85% precision for detecting transitions into sustained isolation
- Evidence so far
- Synthetic harness (prevent12) · Common synthetic checks
- What the evidence covers
- Living alone, short-term stress, subjective loneliness and sustained isolation are separate endpoints. Calculate precision only against verified transition labels.
- What would change it
- Register the scale, time window and eligibility criteria for isolation. Test missing sensor readings and ordinary interruptions in simulation.
Student Suicide Prevention Research
Evaluate age-appropriate language context and the handoff separately. Use fixed inputs to check that unrestricted responses cease following a crisis event.
- Internal target
- 100% immediate handoff after a detected crisis state
- Evidence so far
- Synthetic harness (prevent12) · Common synthetic checks
- What the evidence covers
- Detection sensitivity and handoff rate use different denominators. An independent test corpus needs age, language and timestamped crisis annotations.
- What would change it
- Lock decision, success state and maximum delay for synthetic events. Register eligibility criteria for adolescent reference material.