RESEARCH · PERSONAL
Speech emotion recognition research
Graduate research exploring Conformer networks and pretrained speech representations for speech emotion recognition.
Independent / personalExperimental / concept2023–2024 · Graduate work
⌘
01 / THE IDEA
What it is
Master’s-level investigation of how models can learn emotional cues in speech from time-dependent acoustic features and representations.
02 / THE BUILD
How it works
The research focused on Conformer-based modeling and pretrained speech representations, with experimentation and evaluation in an academic setting.
- Research into spoken emotion classification and acoustic representations
- Conformer architecture and pretrained audio representation experiments
- Graduate research, not a public clinical or production system
Tools and technologies
03 / STATUS
Where it stands
Graduate research. No peer-reviewed publication or production deployment is claimed.
This is independent work by the founder, not company-owned Kado software.
Explore another project
More Kado products, developer tools and independent experiments are collected in one place.