Oct 2026
MPLMM Research
Extending Prompt Learning for Missing-Modality Sentiment Analysis
Team Lead, IIT Bhilai · Oct 2026
Leading a team of 3 at IIT Bhilai to reproduce and extend MPLMM (ACL 2024) on CMU-MOSI, CMU-MOSEI, IEMOCAP and CH-SIMS. We're designing three extensions: 2D CNN spatial facial features, retrieval-augmented prompt context, and a Mixture-of-Experts missing-modality generator.
Try it: switch a signal off
All three modalities are present, so the model reads them directly.
A concept sketch of the idea, not live model output.