AD4Edu-Bench
AD4Edu: Benchmarking University Lecture Audio Description across Accessibility Standards, Metrics, and Learner Preferences
A benchmark for evaluating audio description of university lectures, featuring a 45-rule standard, silver references, and 400 preference pairs from BLV raters.
How to run
$ npm i -g lettokenburn-cli$ ltb bench info ad4edu-bench
External harness. No Harbor adapter has been generated for this benchmark yet. `ltb bench info` prints the known links.
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
Scores
3 sourced entriesEvery number links to where it was published. Low-confidence rows are muted and need per-cell verification.
Reproduced by the community
All runsAbout
AD4Edu-Bench comprises lecture moments with multimodal context (slides, audio), silver reference ADs, and pairwise preferences from blind/low-vision raters. It operationalizes a 45-rule accessibility standard for evaluation.
Scoring uses AD4Edu-Eval, a rule-diagnosable metric checking compliance. Findings show rule compliance does not fully align with user preferences, highlighting the need for human-in-the-loop design.