Share your thoughts, 1 month free Claude Pro on usSee more

Fine-grained CI discrimination on LEADS 10-fold cross validation (in-domain)

93.8AUC

MLP + Qwen3-Omni-30B

Updated 24d ago

Evaluation Results

Method	Links
MLP + Qwen3-Omni-30B 2026.06		93.8
MLP + Clinical BERT 2026.06		90.2
MLP + BERT+Whisper 2026.06		89.1
MLP + BERT 2026.06		88.7
DeTAiL (self) 2026.06		86.3
DeTAiL (teacher) 2026.06		84.4
MLP + BERT (GT text)+Whisper 2026.06		84.3
MLP + Qwen2.5-Omni-7B 2026.06		83.9
MLP + BERT (GT text) 2026.06		83
Qwen2.5-Omni-7B (Zero-shot) 2026.06		71.9
LoRA + Qwen2.5-Omni-7B (Distilled GRPO) 2026.06		60.9
LoRA + Qwen2.5-Omni-7B (GRPO) 2026.06		59.9
LoRA + Qwen2.5-Omni-7B (SFT) 2026.06		53.7