Item analysis
Question statistics
| Question | n | Difficulty p | Discrimination r_pb | Verdict |
|---|---|---|---|---|
| How does deformable attention in Deformable DETR reduce computational complexity compared to standard self-attention? | 6 | 0.33 | 0.79 | OK |
| What enables DETR to avoid using non-maximum suppression (NMS) for duplicate removal? | 6 | 0.33 | 0.79 | OK |
| Why is bipartite matching necessary when training an object detector that outputs a fixed number of prediction slots? | 6 | 0.50 | 0.80 | OK |
| Why is Generalized IoU (GIoU) preferred over standard IoU for bounding box regression during training? | 6 | 0.67 | 0.69 | OK |
| Why is the Vision Transformer (ViT) forced to use relatively large patches for high-resolution images? | 6 | 0.83 | 0.49 | OK |
Rejection rule (from 5 attempts): a question is flagged when its difficulty p falls outside 0.30–0.85 (share answering correctly) or its point-biserial discrimination drops below 0.15 (correlation with the rest of the attempt score). Retiring and regenerating flagged questions runs offline: uv run python -m app.quality … --regenerate.