◇ OpenCyber Academy
Master's · Chapter 8 · English

Chapter 8 — Independent review and defense

Open, free learning for anyone who wants to protect people and technology. No account is required to read this text or its worked answers.

Chapter 8 — Independent review and defense

Lesson map

Lesson 08 published. Map, example, EX045.

flowchart TD
  N0["Learning outcomes"]
  N1["8.1 Independent code review"]
  N2["8.2 Preparing the defense"]
  N3["8.3 Responding to criticism"]
  N4["8.4 Corrections and archiving"]
  N5["8.5 Follow-up work"]
  N6["Synthesis"]
  N0 --> N1 --> N2 --> N3 --> N4 --> N5 --> N6
  N6 --> EX["Worked example"]
  EX --> Q["Exercise EX045"]

Logical drawing

flowchart LR
  N0["Learning outcomes"]
  N1["8.1 Independent code review"]
  N2["8.2 Preparing the defense"]
  N3["8.3 Responding to criticism"]
  N4["8.4 Corrections and archiving"]
  N5["8.5 Follow-up work"]
  N6["Synthesis"]
  N0 --> N1 --> N2 --> N3 --> N4 --> N5 --> N6 --> Q["EX045"]

[Learning outcomes]
    |
    v
[8.1 Independent code review]
    |
    v
[8.2 Preparing the defense]
    |
    v
[8.3 Responding to criticism]
    |
    v
[8.4 Corrections and archiving]
    |
    v
[8.5 Follow-up work]
    |
    v
[Synthesis]
    |
    v
[Worked example] --> [Exercise EX045]

Learning outcomes

You will (1) obtain independent review of code and results; (2) prepare for an oral defense; (3) respond to examiner criticism; (4) plan follow-up work.

8.1 Independent code review

Before submission, a doctoral student in another group reviews Kenji's code using only the repository and its documentation. She reruns all experiments from scratch on her machine and matches the stored summaries, which confirms reproducibility. She also finds that the random-edge control removed edges from the flat network uniformly, whereas a fairer control might preserve each workstation's degree. Kenji adds this alternative control as a sensitivity analysis; it narrows the identity advantage slightly but leaves it clear. He acknowledges the reviewer and records the change in his revision table.

8.2 Preparing the defense

Kenji lists likely questions: why simulation instead of field data, why breadth-first reach rather than attack likelihood, how parameters were chosen, what would change his conclusion and how practitioners should use the results. For each he prepares a short answer pointing to a section and a result. He rehearses with two classmates, who interrupt and challenge. Dr Aydın advises him to answer the question asked rather than the one he prepared, to admit uncertainty plainly and to use the overturn thresholds he wrote in chapter 4 as evidence of his discipline.

8.3 Responding to criticism

In the defense, the external examiner argues that reachability without exploit likelihood exaggerates differences between policies. Kenji agrees that reach is an upper bound, explains why he chose it for tractability and transparency, and suggests weighting edges by exploit likelihood as future work. A second examiner asks whether the template-copy rival undermines the whole recommendation. Kenji answers that it qualifies it: identity segmentation helps only if rule management prevents broad templates, which is a practical finding in itself. The examiners accept the dissertation with minor corrections.

8.4 Corrections and archiving

Kenji completes corrections within the allowed time: clarifying the bootstrap description, adding the degree-preserving control to the main text and expanding the discussion of exploit likelihood. He archives the final code, configurations and results with a persistent identifier from his university repository and includes a SHA-256 manifest of the archived files. Vantor reviews the final version and confirms that no sensitive structure is revealed. The archive allows others to rerun or extend the study with different network shapes. It also lets examiners and later readers check that the published numbers came from the archived code.

8.5 Follow-up work

Kenji lists follow-up studies: weighting edges by exploit likelihood, modelling detection and response time, testing on network shapes from other sectors and running a field pilot at Vantor that measures real rule counts and access attempts. He offers the code to another student. His reflection notes that the most useful moments were when his expectations were challenged, by the random control, by the template-copy rival and by the examiner's critique, because each made his final claim smaller but much more defensible. He now treats a challenged expectation as progress rather than as a setback.

Worked example

Kenji's code is rerun by an independent reviewer, a degree-preserving control is added, he defends the upper-bound choice and template finding and archives the corrected study with a manifest.

Synthetic case. Reviewers, examiners and outcomes are synthetic.

Errors and limits

Independent rerun confirms reproducibility, not validity of assumptions. Minor corrections can still change emphasis.

Exercises

Synthesis

Independent review and a candid defense make a master's result smaller, stronger and ready for others to build on.

Continue studying — chapter 08 of 08

← Chapter 07 (draft) · Course chapter index · Worked results ↓ · Next course via chapter index →