Home.
Sound demixing challenge.
Using neural networks to solve the source separation problem. These audio inverse problems are mainly for sound demixing and cinematic demixing challenge.
Investigating Semantic Roles for Emotion Role Prediction
Project on semantic role labelling using neural networks to find the impact of semantic role features for emotion analysis.
Bengali Handwritten Grapheme detection
Accelerate Bengali handwritten optical character recognition research and help enable the digitalization of educational resources. The problem is centered around given handwritten Bengali graphemes, to separately classify three constituent elements in...
Denoising diffusion for speech enhancement
We used simple diffusion for removing noise from the speech recordings. This project takes mel spectrograms and noises and denoises them for better speech intelligibility.
Jigsaw toxicity detection
Can you help detect toxic comments ― and minimize unintended model bias? Let's work to build a model that recognizes toxicity and minimizes this type of unintended bias with respect to mentions of identities.