Research map / Vision-language and multimodal / Human and embodied modeling
Visual Knowledge Tracing
Neehar Kondapaneni, Pietro Perona, Oisin Mac Aodha
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add ground-truth label y of the query image as input (ablation) | phi_direct · Eyes (test split) | micro average precision: 0.35 -> 0.75 | Improved |
| add ground-truth label y of the query image as input (ablation) | phi_cls_pred · Eyes (test split) | micro average precision: 0.44 -> 0.74 | Improved |
| add query image representation z to the input (ablation) | phi_direct · Eyes (test split) | micro average precision: 0.75 -> 0.75 | No clear effect |
| add query image representation z to the input (ablation) | phi_cls_pred · Eyes (test split) | micro average precision: 0.74 -> 0.75 | No clear effect |
| use recurrent direct model instead of static linear classifier | phi_static · Eyes (test split) | micro average precision: 0.67 -> 0.75 | Improved |
| make static classifier time-dependent (one classifier per step) | phi_static · Eyes (test split) | micro average precision: 0.67 -> 0.33 | Hurt |
| add image information (instance-level) versus class-level DKT | phi_dkt · Eyes (test split) | micro average precision: 0.74 -> 0.75 | Improved |
| replace LSTM with Transformer | phi_direct · Eyes (test split) | micro average precision: 0.75 -> 0.37 | Hurt |
| use prototype cognitive model instead of recurrent model | phi_direct · Eyes (test split) | micro average precision: 0.75 -> 0.65 | Hurt |
| use exemplar cognitive model instead of recurrent model | phi_direct · Eyes (test split) | micro average precision: 0.75 -> 0.68 | Hurt |
| use cls_pred instead of direct recurrent model | phi_direct · Eyes (test split) | micro average precision: 0.75 -> 0.74 | No clear effect |
| add per-class accuracy meta-information to the input | phi_static, phi_direct, phi_cls_pred · all datasets | average precision | Improved |
Closest papers by meaning
- CleverBirds: A Multiple-Choice Benchmark for Fine-grained Human Knowledge Tracing (NeurIPS 2025 · similarity 0.87)
- L-WISE: Boosting Human Visual Category Learning Through Model-Based Image Selection and Enhancement (ICLR 2025 · similarity 0.86)
- How Well Do Unsupervised Learning Algorithms Model Human Real-time and Life-long Learning? (NeurIPS 2022 · similarity 0.85)
- Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations (IJCV 2017 · similarity 0.85)
- Cognitive Fluctuations Enhanced Attention Network for Knowledge Tracing (AAAI 2025 · similarity 0.84)
- Modeling Collaborator: Enabling Subjective Vision Classification With Minimal Human Effort via LLM Tool-Use (CVPR 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
