Ysobel Sims

About

Ysobel Sims is a rising researcher at the intersection of machine learning and audio processing, with a primary focus on zero-shot learning for environmental audio. Her most-cited work, "Enhanced Embeddings in Zero-Shot Learning for Environmental Audio" (2023), tackles the challenge of recognizing unseen sound classes by improving both audio and word embeddings. Building on the VGGish model, Sims demonstrates how refined embeddings can bridge the gap between training and test classes, enabling machines to identify novel environmental sounds without explicit examples. This contribution is particularly impactful for applications in wildlife monitoring, smart cities, and acoustic surveillance, where labeled data is scarce. Though early in her career, her work has already garnered attention, with citations from researchers exploring zero-shot learning in non-visual domains. Sims’ approach—enhancing the semantic alignment between sounds and their descriptions—positions her as a promising voice in audio AI. Her research not only advances fundamental machine learning but also opens doors for more adaptable, real-world audio recognition systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Enhanced Embeddings in Zero-Shot Learning for Environmental Audio

   2 citations · 2023

Key Collaborators

Contact & Links