Tag
This paper benchmarks auxiliary uncertainty signals from a BERT+GCN classifier to improve LLM-assisted systematic review screening, showing that targeted MAYBE-only routing maximizes efficiency with near-baseline cost.
This paper compares expert-assigned and automatically-assigned MeSH terms as features for systematic review screening classifiers, showing that evaluation design significantly affects measured performance gaps, with canonical designs showing larger gaps that attenuate under alternative designs.
This paper presents ShishuRaksha AI, a training-free multimodal framework for screening abuse-related trauma in Bangladeshi children, evaluated on a noise-aware synthetic dataset, achieving an AUC of 0.874.
A personal reflection on the history of medical ultrasound innovation, linking it to modern device builders like David Holz, and highlighting the importance of perseverance in the face of skepticism.
SelectPrism is a tool that uses AI agents to screen and interview candidates, helping companies hire faster.
A developer built a voice AI agent using open-source tools Dograh and n8n to automate mentor screening calls, reducing daily workload from 4 hours to 30 minutes.