An extensive collection of Speech Emotion Recognition (SER) datasets across multiple languages, including English, Mandarin, Hindi, Spanish, Tamil, Arabic, and more. Perfect for training emotion detection models in diverse linguistic and cultural contexts.

Static website for aaivulabs.com — AI Research & Innovation

Transformer-based model for accurate and scalable panel time series forecasting. Builds on Temporal Fusion Transformer with innovations like segment-wise attention, multi-scale decomposition, and cross-entity attention.

EmoTa is an open-access Tamil Speech Emotion Recognition dataset with 936 utterances from 22 native speakers, covering five emotions (anger, happiness, sadness, fear, and neutrality). It supports emotion classification tasks and advances Tamil language processing.

A deep learning-based Speech Emotion Recognition (SER) model trained primarily on Indian languages. Designed for applications in call centers, sentiment analysis, and accessibility tools.