analiza aiqbench
This research introduces a novel distillation framework that leverages uncertainty-aware supervision from Large Vision-Language Models to optimize cloud-edge workflows for emergency chest X-ray triage. By treating model outputs as probabilistic samples rather than fixed targets, the approach effectively mitigates hallucinations and instability while achieving high accuracy with low latency on edge devices. The method establishes a concrete reliability-latency operating point specifically designed for clinical environments where multimodal data constraints are common.
Confidence-Gated Cloud-Edge Cascade Triage via Variational Risk Minimization for Medical Imaging
arXiv:2610.02269v1 Announce Type: cross Abstract: Emergency chest X-ray (CXR) triage has a structural modality gap: reports arrive after triage decisions, yet multimodal foundation models require image-text inputs. We present Variational Risk Minimization (VRM), a distillation framework that treats LVLM-generated report variants as Monte Carlo samples of latent clinical interpretations. Rather than distilling from a single teacher target, VRM learns from a variationally marginalized teacher distribution, enabling uncertainty-aware supervision under missing-modality constraints. Under matched encoder families, VRM outperforms direct fine-tuning baselines and improves calibration with strong recovery from hallucinated supervision. Marginalized supervision reduces report-selection instability. In our compact edge-student instantiation, a confidence-gated cascade reaches AUC 0.941 at 103ms average latency with 20.3% cloud escalation, yielding an explicit reliability-latency operating point for cloud-edge clinical workflows.
Abstract:Emergency chest X-ray (CXR) triage has a structural modality gap: reports arrive after triage decisions, yet multimodal foundation models require image-text inputs. We present Variational Risk Minimization (VRM), a distillation framework that treats LVLM-generated report variants as Monte Carlo samples of latent clinical interpretations. Rather than distilling from a single teacher target, VRM learns from a variationally marginalized teacher distribution, enabling uncertainty-aware supervision under missing-modality constraints. Under matched encoder families, VRM outperforms direct fine-tuning baselines and improves calibration with strong recovery from hallucinated supervision. Marginalized supervision reduces report-selection instability. In our compact edge-student instantiation, a confidence-gated cascade reaches AUC 0.941 at 103ms average latency with 20.3% cloud escalation, yielding an explicit reliability-latency operating point for cloud-edge clinical workflows.
14 pages, 6 figures, 14 tables. Accepted manuscript of the article published in Smart Health 41 (2026) 100689. Presented as an oral at IEEE/ACM CHASE 2026. Code: this https URL
Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
arXiv:2610.02269 [eess.IV]
(or arXiv:2610.02269v1 [eess.IV] for this version)
https://doi.org/10.1016/j.smhl.2026.100689 DOI(s) linking to related resources
Text integral sub licența deschisă a publicației · 176 cuvinte
Subiecte conexe
Article: Architecting Secure and Scalable Facial Verification Systems
InfoQ · 18 sept. 2026
Building trust in AI agents: A production-ready evaluation framework with Amazon Bedrock AgentCore
AWS · 16 oct. 2026
A Two-Stage Cascade for Near-Real-Time Forest Anomaly Detection from Sentinel-1 SAR Time Series
arXiv · 5 oct. 2026
Game Design Creativity evals
LessWrong · 12 oct. 2026