TDAA: high-frequency closed-book tests paired with a human-reviewed AI production harness¶

Design · All designs
Evidence · 2 claims (2 for) · 1 study (1 design), q2 · 0 of 1 report an effect size · 2 claims rest on one study

Description¶

TDAA is a course-level design pattern coupling two cycles: students prepare with aligned materials and face a weekly closed-book gate, while an instructor-side AI harness drafts, reviews, and grades the materials under human approval. The paper states the pairing directly: "high-frequency tests preserve individual accountability, while AI agents make material production and marking scalable." Without the agent layer weekly testing is infeasible for one instructor; without the test gate AI-assisted preparation offers no reliable check on unaided performance.

Design Implications¶

Context¶

Requirements¶

A visible preparation path of learning sheets and aligned validation practice so the closed-book tests feel fair rather than arbitrary
A version-controlled agent workspace with human review before any AI draft reaches students

Constraints¶

Evidence is limited to one small, proof-heavy course without a control group; transfer to large or non-proof courses is a hypothesis for future offerings, not a result

Target Learners¶

upper-level undergraduate data-science students in a proof-heavy course

Learning Goals¶

unaided problem-solving and proof-writing in theory of computation

Claims¶

Examples¶

-

Key Sources¶

Jin-Guo Liu, Shang-Qi Lu, Xin-Ran Shi, Long-Li Zheng, Wei Wang. (2026). Test-Driven, AI-Assisted Learning: Replacing Lectures with Weekly Closed-Book Tests. https://arxiv.org/abs/2606.23315

Page metadata

Field	Value
`type`	`design`
`id`	`tdaa-test-driven-ai-assisted-pattern`
`title`	`TDAA: high-frequency closed-book tests paired with a human-reviewed AI production harness`
`description`	`TDAA is a course-level design pattern coupling two cycles: students prepare with aligned materials and face a weekly closed-book gate, while an instructor-side AI harness drafts, reviews, and grades the materials unde...`
`status`	`review`
`generated`	by `process:settle-candidates` on `2026-10-12`
`trust`	unverified — no one has confirmed this page's content
`sources`	1 — `jin-guo-liu-2026`

Frontmatter is the page's machine-readable record — see the Schema & Guide.