Patient-Level Multimodal Question Answering from Multi-Site Auscultation Recordings
Jul 1, 2026·
,,,,,,,,,,·
0 min read
Fan Wu
Tzu-Ning Wang
Nicolas Zumarraga
Nan Wang
Marc Kreft
Kai O'Sullivan
Elgar Fleisch
Oliver Aalami
Paul Schmiedmayer
Robin Jakob
Philipp Langer

Abstract
A patient-level multimodal question answering approach built from multi-site auscultation recordings, combining audio signals with language models for clinical reasoning.
Publication
Interspeech