Thesis roadmap
Download the standalone strategic roadmap PDF · Strategic source data · Thesis PDF
This page combines the reviewed strategic plan with a separately timestamped projection of human review tasks. The PDF preserves the reviewed strategic snapshot; the live task projection below refreshes on Pages builds. The active thesis and immutable evidence retain ownership of scientific claims.
2 Human review progress
GitHub issues own the review queue. Only section tasks closed as completed record human task completion; closure without a completed reason does not. Task completion does not establish scientific validity or acceptance of later edits.
Observed: 2026-10-22T08:45:55.216269+00:00. Snapshot: complete.
Section tasks: 0/42 completed.
| Review item | State | Planned date | Unresolved prerequisites |
|---|---|---|---|
| #786 [author review] Full manuscript: final rendered review and submission-ready decision | open | 2026-11-26 | 0 |
| #785 [author review] Full manuscript: figures, captions, tables and references | open | 2026-11-23 | 0 |
| #784 [author review] Front matter: title, acknowledgements and declarations | open | 2026-11-22 | 0 |
| #783 [author review] Appendices: metric validity and measurement details | open | 2026-11-21 | 0 |
| #782 [author review] Appendices: reproducibility and Phase-A contract | open | 2026-11-20 | 0 |
| #781 [author review] Conclusion: answers, limits and final claims | open | 2026-11-19 | 0 |
| #780 [author review] Discussion: Data collection before action-space replacement | open | 2026-11-18 | 0 |
| #779 [author review] Discussion: What the unresolved comparisons must distinguish | open | 2026-11-17 | 0 |
| #778 [author review] Discussion: What the available evidence establishes | open | 2026-11-16 | 0 |
| #777 [author review] Results: historical diagnostics, coverage and outstanding evidence | open | 2026-11-12 | 0 |
| #776 [author review] Results: one-step and continuing-value findings | open | 2026-11-11 | 0 |
| #775 [author review] Results: supervision and selected-camera behavior | open | 2026-11-10 | 0 |
| #774 [author review] Results: measurement and collected evidence | open | 2026-11-09 | 0 |
| #773 [author review] Experimental design: Policy Comparison and Statistical Protocol | open | 2026-11-06 | 0 |
| #772 [author review] Experimental design: Replay Eligibility and Finite-Horizon Learning Gate | open | 2026-11-04 | 0 |
| #771 [author review] Experimental design: Study Population and Evidence Gates | open | 2026-11-02 | 0 |
| #770 [author review] Method: Value Definition and Learning | open | 2026-10-30 | 0 |
| #769 [author review] Method: Finite Action and Causal Replay | open | 2026-10-29 | 0 |
| #768 [author review] Method: Spatial Readout and Interaction Comparisons | open | 2026-10-29 | 0 |
| #767 [author review] Method: Pose-history order and observed content | open | 2026-10-28 | 0 |
| #766 [author review] Method: Representation Comparisons | open | 2026-10-28 | 0 |
| #765 [author review] Method: Target-Conditioned Finite-Horizon Value | open | 2026-10-27 | 0 |
| #764 [author review] Method: Candidate-to-State Interaction | open | 2026-10-27 | 0 |
| #763 [author review] Method: Target, Candidate, and History Encoding | open | 2026-10-26 | 0 |
| #762 [author review] Method: Current and Intended Actor State | open | 2026-10-26 | 0 |
| #761 [author review] Data generation: Causal replay, storage and chapter synthesis | open | 2026-10-20 | 0 |
| #760 [author review] Data generation: Collection and proposal diagnostics | open | 2026-10-20 | 0 |
| #759 [author review] Data generation: Rewards, finite-horizon returns and endpoint gain | open | 2026-10-19 | 0 |
| #758 [author review] Data generation: Operational error, reliability and construct validity | open | 2026-10-19 | 0 |
| #757 [author review] Data generation: Observation updates and measurement preparation | open | 2026-10-18 | 0 |
| #756 [author review] Data generation: Candidate view generation | open | 2026-10-18 | 0 |
| #755 [author review] Data generation: Target selection and observed-target admission | open | 2026-10-17 | 0 |
| #754 [author review] Data generation: State and information access | open | 2026-10-17 | 0 |
| #753 [author review] Data generation: Dataset ecosystem | open | 2026-10-17 | 0 |
| #752 [author review] Background 2.6: Related-Work Synthesis and Research Gap | open | 2026-10-24 | 0 |
| #751 [author review] Background 2.5: Representations for Egocentric NBV | open | 2026-10-23 | 0 |
| #750 [author review] Background 2.4: Sequential Decision-Making under Partial Observability | open | 2026-10-23 | 0 |
| #749 [author review] Background 2.3: Candidate-View Support and Motion Feasibility | open | 2026-10-23 | 0 |
| #748 [author review] Background 2.2: Reconstruction Objectives and Target Scope | open | 2026-10-22 | 0 |
| #747 [author review] Background 2.1: Choosing the Next Observation | open | 2026-10-22 | 0 |
| #746 [author review] Introduction: contributions, research questions and thesis structure | open | 2026-10-22 | 0 |
| #745 [author review] Introduction: motivation and simulated acquisitions | open | 2026-10-20 | 0 |
| #744 [author review] Front matter and final candidate | open | 2026-11-30 | 0 |
| #743 [author review] Appendices | open | 2026-11-22 | 0 |
| #742 [author review] Conclusion | open | 2026-11-22 | 0 |
| #741 [author review] Discussion | open | 2026-11-22 | 0 |
| #740 [author review] Results | open | 2026-11-15 | 0 |
| #739 [author review] Experimental Design | open | 2026-11-08 | 0 |
| #738 [author review] Method | open | 2026-11-01 | 0 |
| #737 [author review] Oracle and Data Generation | open | 2026-10-26 | 0 |
| #736 [author review] Foundations and Related Work | open | 2026-10-26 | 0 |
| #735 [author review] Introduction | open | 2026-10-26 | 0 |
3 Strategic evidence schedule
REVIEW DUE — substantive review: 2026-10-02; next review due: 2026-10-16. Page built as of 2026-10-22 (UTC). Automated issue refreshes do not advance these review dates or milestone states.
The experimental schedule below was last reviewed on 2 October and is not revalidated by this author-review plan. Data-generation and training-run scheduling is deferred until LRZ throughput is established.
| Gate | Phase | Planned dates | Reviewed state | Exit gate |
|---|---|---|---|---|
| P0 | Pilot infrastructure | 2026-07-07 – 2026-08-30 | done | Committed smoke and MVP artifacts compile and remain explicitly bounded as pilot evidence. |
| P1 | Freeze experiment substrate | 2026-08-31 – 2026-10-02 | in-progress | Multi-scene immutable store, scene-level split, camera/depth and target-RRI fidelity, candidate-support admission, and throughput receipts. |
| P2 | Paired policy baselines | 2026-10-03 – 2026-10-20 | planned | Equal-budget random-valid, greedy, and bounded-lookahead comparisons report uncertainty, support, and oracle receipts. |
| P3 | Learned value evaluation | 2026-10-21 – 2026-11-10 | planned | Held-out ranking, calibration, and paired policy value are reported under actor-visible inputs. |
| P4 | Results and discussion synthesis | 2026-11-11 – 2026-11-22 | planned | Final tables, failure cases, coverage limits, and claim dispositions resolve to the immutable evidence registry. |
| P5 | Reproducibility and full-draft freeze | 2026-11-23 – 2026-11-30 | planned | Full draft, reproducible configurations, final figures, and release checks pass with no unresolved placeholders. |
| P6 | Safety buffer and submission | 2026-12-01 – 2027-01-07 | buffer | Advisor feedback, formal checks, final upload, and the authenticated submission receipt are closed without reopening the scientific core. |