Medical Visual Question Answering on Pooled Medical VQA Datasets Open 5-fold CV
0.068ACEIsotonic Regr.
Evaluation Results
| Method | Links | |
|---|---|---|
| Isotonic Regr.Backbone=LLaVA-NeXT-7B, Confidence Type=Sampling2026.04 | 0.068 | |
| HAC-PlattBackbone=Qwen3-VL-8B, Confidence Type=Verbalized2026.04 | 0.075 | |
| HAC-GateBackbone=Qwen3-VL-8B, Confidence Type=Verbalized2026.04 | 0.075 | |
| HAC-GateBackbone=LLaVA-NeXT-7B, Confidence Type=Sampling2026.04 | 0.077 | |
| Hist. BinningBackbone=Qwen3-VL-32B, Confidence Type=Verbalized2026.04 | 0.079 | |
| Platt ScalingBackbone=LLaVA-NeXT-7B, Confidence Type=Sampling2026.04 | 0.082 | |
| HAC-PlattBackbone=InternVL3-2B, Confidence Type=Sampling2026.04 | 0.083 | |
| HAC-PlattBackbone=LLaVA-NeXT-7B, Confidence Type=Verbalized2026.04 | 0.083 | |
| HAC-GateBackbone=LLaVA-NeXT-7B, Confidence Type=Verbalized2026.04 | 0.084 | |
| HAC-GateBackbone=InternVL3-2B, Confidence Type=Sampling2026.04 | 0.085 | |
| HAC-PlattBackbone=Qwen3-VL-8B, Confidence Type=Sampling2026.04 | 0.086 | |
| HAC-GateBackbone=Qwen3-VL-8B, Confidence Type=Sampling2026.04 | 0.086 | |
| HAC-PlattBackbone=InternVL3-38B, Confidence Type=Sampling2026.04 | 0.086 | |
| Isotonic Regr.Backbone=Qwen3-VL-8B, Confidence Type=Verbalized2026.04 | 0.087 | |
| Hist. BinningBackbone=Qwen3-VL-8B, Confidence Type=Verbalized2026.04 | 0.087 | |
| HAC-PlattBackbone=Qwen3-VL-32B, Confidence Type=Verbalized2026.04 | 0.087 | |
| Platt ScalingBackbone=InternVL3-2B, Confidence Type=Sampling2026.04 | 0.087 | |
| HAC-PlattBackbone=InternVL3-8B, Confidence Type=Verbalized2026.04 | 0.087 | |
| HAC-GateBackbone=Qwen3-VL-2B, Confidence Type=Verbalized2026.04 | 0.088 | |
| Isotonic Regr.Backbone=Qwen3-VL-32B, Confidence Type=Verbalized2026.04 | 0.088 | |
| HAC-GateBackbone=InternVL3-38B, Confidence Type=Sampling2026.04 | 0.088 | |
| HAC-PlattBackbone=Qwen3-VL-2B, Confidence Type=Verbalized2026.04 | 0.089 | |
| Hist. BinningBackbone=InternVL3-2B, Confidence Type=Sampling2026.04 | 0.089 | |
| HAC-GateBackbone=Qwen3-VL-32B, Confidence Type=Sampling2026.04 | 0.09 | |
| Isotonic Regr.Backbone=InternVL3-2B, Confidence Type=Sampling2026.04 | 0.09 | |
| Isotonic Regr.Backbone=InternVL3-8B, Confidence Type=Sampling2026.04 | 0.09 | |
| HAC-GateBackbone=InternVL3-8B, Confidence Type=Verbalized2026.04 | 0.091 | |
| Isotonic Regr.Backbone=InternVL3-38B, Confidence Type=Sampling2026.04 | 0.091 | |
| HAC-GateBackbone=Qwen3-VL-32B, Confidence Type=Verbalized2026.04 | 0.092 | |
| Hist. BinningBackbone=InternVL3-38B, Confidence Type=Sampling2026.04 | 0.093 | |
| Hist. BinningBackbone=LLaVA-NeXT-7B, Confidence Type=Sampling2026.04 | 0.093 | |
| Isotonic Regr.Backbone=Qwen3-VL-8B, Confidence Type=Sampling2026.04 | 0.094 | |
| Platt ScalingBackbone=InternVL3-8B, Confidence Type=Sampling2026.04 | 0.094 | |
| Isotonic Regr.Backbone=InternVL3-8B, Confidence Type=Verbalized2026.04 | 0.094 | |
| HAC-PlattBackbone=InternVL3-8B, Confidence Type=Sampling2026.04 | 0.094 | |
| Platt ScalingBackbone=Qwen3-VL-8B, Confidence Type=Verbalized2026.04 | 0.095 | |
| Platt ScalingBackbone=Qwen3-VL-32B, Confidence Type=Verbalized2026.04 | 0.095 | |
| Platt ScalingBackbone=InternVL3-2B, Confidence Type=Verbalized2026.04 | 0.096 | |
| Hist. BinningBackbone=InternVL3-8B, Confidence Type=Verbalized2026.04 | 0.096 | |
| Isotonic Regr.Backbone=InternVL3-38B, Confidence Type=Verbalized2026.04 | 0.096 | |
| HAC-PlattBackbone=LLaVA-NeXT-7B, Confidence Type=Sampling2026.04 | 0.096 | |
| HAC-GateBackbone=InternVL3-8B, Confidence Type=Sampling2026.04 | 0.097 | |
| Platt ScalingBackbone=Qwen3-VL-8B, Confidence Type=Sampling2026.04 | 0.098 | |
| HAC-GateBackbone=LLaVA-NeXT-34B, Confidence Type=Sampling2026.04 | 0.098 | |
| Platt ScalingBackbone=InternVL3-38B, Confidence Type=Sampling2026.04 | 0.099 | |
| HAC-PlattBackbone=Qwen3-VL-32B, Confidence Type=Sampling2026.04 | 0.1 | |
| HAC-PlattBackbone=Qwen3-VL-2B, Confidence Type=Sampling2026.04 | 0.102 | |
| Platt ScalingBackbone=InternVL3-8B, Confidence Type=Verbalized2026.04 | 0.102 | |
| Isotonic Regr.Backbone=Qwen3-VL-32B, Confidence Type=Sampling2026.04 | 0.103 | |
| Hist. BinningBackbone=Qwen3-VL-8B, Confidence Type=Sampling2026.04 | 0.104 | |
| Isotonic Regr.Backbone=InternVL3-2B, Confidence Type=Verbalized2026.04 | 0.104 | |
| Hist. BinningBackbone=InternVL3-38B, Confidence Type=Verbalized2026.04 | 0.104 | |
| Platt ScalingBackbone=LLaVA-NeXT-7B, Confidence Type=Verbalized2026.04 | 0.104 | |
| Hist. BinningBackbone=InternVL3-8B, Confidence Type=Sampling2026.04 | 0.105 | |
| Platt ScalingBackbone=InternVL3-38B, Confidence Type=Verbalized2026.04 | 0.105 | |
| Isotonic Regr.Backbone=Qwen3-VL-2B, Confidence Type=Sampling2026.04 | 0.106 | |
| HAC-GateBackbone=Qwen3-VL-2B, Confidence Type=Sampling2026.04 | 0.106 | |
| Hist. BinningBackbone=InternVL3-2B, Confidence Type=Verbalized2026.04 | 0.106 | |
| Isotonic Regr.Backbone=LLaVA-NeXT-7B, Confidence Type=Verbalized2026.04 | 0.106 | |
| Hist. BinningBackbone=Qwen3-VL-32B, Confidence Type=Sampling2026.04 | 0.107 | |
| Hist. BinningBackbone=LLaVA-NeXT-7B, Confidence Type=Verbalized2026.04 | 0.107 | |
| Platt ScalingBackbone=Qwen3-VL-2B, Confidence Type=Verbalized2026.04 | 0.108 | |
| Isotonic Regr.Backbone=Qwen3-VL-2B, Confidence Type=Verbalized2026.04 | 0.109 | |
| Platt ScalingBackbone=Qwen3-VL-2B, Confidence Type=Sampling2026.04 | 0.11 | |
| Platt ScalingBackbone=Qwen3-VL-32B, Confidence Type=Sampling2026.04 | 0.111 | |
| Hist. BinningBackbone=LLaVA-NeXT-34B, Confidence Type=Sampling2026.04 | 0.112 | |
| HAC-PlattBackbone=InternVL3-2B, Confidence Type=Verbalized2026.04 | 0.115 | |
| HAC-PlattBackbone=InternVL3-38B, Confidence Type=Verbalized2026.04 | 0.116 | |
| HAC-GateBackbone=LLaVA-NeXT-34B, Confidence Type=Verbalized2026.04 | 0.116 | |
| Hist. BinningBackbone=Qwen3-VL-2B, Confidence Type=Verbalized2026.04 | 0.118 | |
| HAC-GateBackbone=InternVL3-2B, Confidence Type=Verbalized2026.04 | 0.118 | |
| Isotonic Regr.Backbone=LLaVA-NeXT-34B, Confidence Type=Verbalized2026.04 | 0.118 | |
| HAC-GateBackbone=InternVL3-38B, Confidence Type=Verbalized2026.04 | 0.119 | |
| Hist. BinningBackbone=LLaVA-NeXT-34B, Confidence Type=Verbalized2026.04 | 0.119 | |
| HAC-PlattBackbone=LLaVA-NeXT-34B, Confidence Type=Sampling2026.04 | 0.119 | |
| Hist. BinningBackbone=Qwen3-VL-2B, Confidence Type=Sampling2026.04 | 0.12 | |
| Isotonic Regr.Backbone=LLaVA-NeXT-34B, Confidence Type=Sampling2026.04 | 0.12 | |
| HAC-PlattBackbone=LLaVA-NeXT-34B, Confidence Type=Verbalized2026.04 | 0.12 | |
| Platt ScalingBackbone=LLaVA-NeXT-34B, Confidence Type=Verbalized2026.04 | 0.124 | |
| Platt ScalingBackbone=LLaVA-NeXT-34B, Confidence Type=Sampling2026.04 | 0.133 | |
| UncalibratedBackbone=InternVL3-8B, Confidence Type=Sampling2026.04 | 0.228 | |
| UncalibratedBackbone=InternVL3-2B, Confidence Type=Sampling2026.04 | 0.249 | |
| UncalibratedBackbone=LLaVA-NeXT-7B, Confidence Type=Sampling2026.04 | 0.256 | |
| UncalibratedBackbone=InternVL3-38B, Confidence Type=Sampling2026.04 | 0.294 | |
| UncalibratedBackbone=Qwen3-VL-32B, Confidence Type=Verbalized2026.04 | 0.321 | |
| UncalibratedBackbone=Qwen3-VL-2B, Confidence Type=Sampling2026.04 | 0.345 | |
| UncalibratedBackbone=Qwen3-VL-32B, Confidence Type=Sampling2026.04 | 0.349 | |
| UncalibratedBackbone=InternVL3-8B, Confidence Type=Verbalized2026.04 | 0.352 | |
| UncalibratedBackbone=LLaVA-NeXT-34B, Confidence Type=Sampling2026.04 | 0.36 | |
| UncalibratedBackbone=Qwen3-VL-8B, Confidence Type=Verbalized2026.04 | 0.363 | |
| UncalibratedBackbone=Qwen3-VL-8B, Confidence Type=Sampling2026.04 | 0.4 | |
| UncalibratedBackbone=InternVL3-38B, Confidence Type=Verbalized2026.04 | 0.406 | |
| UncalibratedBackbone=Qwen3-VL-2B, Confidence Type=Verbalized2026.04 | 0.411 | |
| UncalibratedBackbone=InternVL3-2B, Confidence Type=Verbalized2026.04 | 0.421 | |
| UncalibratedBackbone=LLaVA-NeXT-7B, Confidence Type=Verbalized2026.04 | 0.539 | |
| UncalibratedBackbone=LLaVA-NeXT-34B, Confidence Type=Verbalized2026.04 | 0.551 |