In-house
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Classification | In-house | Accuracy92.27 | 26 | |
| Mean Deviation prediction | In-house (test) | R2 Score0.229 | 22 | |
| Speech Recognition | In-House dataset | CER0.0305 | 19 | |
| Multimodal Research | In-house 2 | Accuracy37.7 | 18 | |
| Multimodal Research | In-house 1 | Accuracy52.5 | 18 | |
| Medical VQA | In-house (Held-out) | Accuracy80.5 | 16 | |
| Segmentation | Held-out in-house (test) | IOU72.09 | 13 | |
| Polyp Detection | In-house 1.0 (test) | Precision94.2 | 12 | |
| Clinical text revision | In house Radiology Report | Mistral Score0.79 | 11 | |
| PET reconstruction | In-House (1% Count) | SSIM0.974 | 10 | |
| PET reconstruction | In-House 10% Count real (test) | SSIM0.9852 | 10 | |
| Abnormal classification | In-house (test) | AUC88.4 | 10 | |
| Asymmetric classification | In-house (test) | AUC0.907 | 10 | |
| Text-to-Image Generation | in-house (test) | Overall HP Score5.39 | 9 | |
| Colorectal polyp classification | In-house WLI (test) | Accuracy0.802 | 8 | |
| Thyroid Ultrasound Segmentation | In-house | DSC82.59 | 7 | |
| Lung nodule malignancy prediction | In-house (test) | AUC (<10mm)0.899 | 7 | |
| Contrast-enhanced breast MRI synthesis | In-house | SSIM90.9 | 6 | |
| Semantic Segmentation | In-house | Accuracy93.7 | 6 | |
| Explanation quality evaluation | In-house Dataset | Helpfulness80.8 | 6 | |
| Dynamic Reconstruction | in-house dataset | SSIMa66.81 | 6 | |
| Staging & Diagnosis | In-house n=640 (test) | BAcc84.3 | 6 | |
| Automatic Speech Recognition | In-house EN music (test) | WER24.61 | 5 | |
| Automatic Speech Recognition | In-house EN speech (test) | WER9.42 | 5 | |
| Automatic Speech Recognition | In-house ZH domain H (test) | WER3.59 | 5 |
Showing 25 of 33 rows