Scanning Parameter Configuration on Instruct-Obs2Param 1.0 (test)
97.9Sample Freq Exact AccuracyLogistic Regression
Evaluation Results
| Method | Links | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Logistic RegressionCategory=Instruction-Only Methods2026.05 | 97.9 | 98.6 | 98 | 93.1 | 91.8 | 75.3 | 92.7 | 74.1 | 83.3 | 96.5 | 81.5 | 90.2 | 95.5 | 88.3 | 88 | 95.8 | 86.7 | |
| KNNCategory=Instruction-Only Methods2026.05 | 92.6 | 95.7 | 92.6 | 83.4 | 82.1 | 74.8 | 91.9 | 73.3 | 80.9 | 95.5 | 77.7 | 79.3 | 93.5 | 75.2 | 82.2 | 94.1 | 80.2 | |
| ScanHDCategory=Our Model2026.05 | 90.6 | 94.3 | 90.5 | 96.5 | 97.1 | 90.6 | 99.7 | 90.3 | 86.4 | 99.9 | 81.5 | 99.3 | 100 | 99.4 | 92.7 | 98.5 | 91.8 | |
| Early FusionCategory=Fusion Methods2026.05 | 85.1 | 90.9 | 84.2 | 91.2 | 91.6 | 84.5 | 100 | 85.2 | 86.2 | 99.8 | 85.6 | 98.8 | 99.4 | 98.6 | 89.2 | 97.5 | 89 | |
| Late FusionCategory=Fusion Methods2026.05 | 76 | 89.2 | 70.4 | 88 | 86.5 | 76.2 | 99.9 | 75.6 | 75.1 | 94.2 | 68 | 98.7 | 99.1 | 98.2 | 82.8 | 95.6 | 79.7 | |
| Qwen3-VL-8B-InstructCategory=Multimodal Large Language Models, zero-shot=true2026.05 | 64.3 | 89.5 | 48.9 | 63.3 | 46 | 34.2 | 69.3 | 27.2 | 53.1 | 87.1 | 42.9 | 34.9 | 78.5 | 30.1 | 49.9 | 81.1 | 39 | |
| Qwen3-VL-8B-ThinkingCategory=Multimodal Large Language Models, zero-shot=true2026.05 | 62.6 | 86.9 | 57.5 | 61.4 | 45.8 | 33.3 | 84.1 | 33.2 | 47.6 | 82 | 40.6 | 43.3 | 82.7 | 33.8 | 49.6 | 83.9 | 42.2 | |
| Qwen3-VL-4B-ThinkingCategory=Multimodal Large Language Models, zero-shot=true2026.05 | 62 | 84.3 | 59.5 | 60 | 47.1 | 36.7 | 80.2 | 31.3 | 52.9 | 84.3 | 43.2 | 27.8 | 78.4 | 27.7 | 47.9 | 81.8 | 41.8 | |
| Qwen3-VL-4B-InstructCategory=Multimodal Large Language Models, zero-shot=true2026.05 | 60 | 92.8 | 55.4 | 59.3 | 48.7 | 24 | 89.2 | 18.9 | 65.6 | 94.3 | 53.5 | 40.2 | 79.7 | 36.6 | 49.8 | 89 | 42.6 | |
| ViTCategory=Observation-Only Methods2026.05 | 44.2 | 65 | 31.3 | 71.4 | 72.2 | 72.2 | 100 | 72.1 | 74.1 | 100 | 77 | 100 | 100 | 100 | 72.4 | 91.3 | 70.5 | |
| ResnetCategory=Observation-Only Methods2026.05 | 42.2 | 62.6 | 28.5 | 71.2 | 70.5 | 74.8 | 100 | 73.7 | 73.7 | 100 | 74.8 | 100 | 100 | 100 | 72.4 | 90.7 | 69.5 | |
| Rule-based HeuristicCategory=Rule-based Baselines2026.05 | 39.9 | 59.1 | 23.3 | 56.5 | 24 | 47.7 | 78.4 | 21.9 | 53.1 | 89.2 | 25.3 | 54.7 | 100 | 23.8 | 50.4 | 81.7 | 23.7 |