Papers by Kristina T. Johnson
N-CORE: N-View Consistency Regularization for Disentangled Representation Learning in Nonverbal Vocalizations (2025.emnlp-main)
Copied to clipboard
| Challenge: | Nonverbal vocalizations are an essential component of human communication, conveying rich information without linguistic content. |
| Approach: | They propose a backbone-agnostic framework to disentangle emotion and speaker information from nonverbal vocalizations by leveraging N views of audio samples to learn invariance to specific transformations. |
| Outcome: | The proposed framework achieves competitive performance compared to state-of-the-art methods on the VIVAE, ReCANVo, and ReCANVO-Balanced datasets. |
ROSCO-Omni: Multimodal LLM-Based Communication Understanding for Non- and Minimally-Speaking Autistic Individuals (2026.findings-acl)
Copied to clipboard
| Challenge: | 30% of autistic individuals remain non- or minimally-speaking throughout their lives . however, caregivers rely on simultaneous integration of visual cues, auditory signals, and contextual understanding to infer intent. |
| Approach: | They propose a framework that fine tunes a teacher-student MLLM for domain-specialized inference. |
| Outcome: | The proposed framework achieves comparable performance to closed-source models . |