Unleashing AI: Multilingual, Ethical, and Informed
Ensuring Fairness in AI, Delivering Global News Insights, and Empowering Multilingual Interactions
Ensuring Fairness in AI, Delivering Global News Insights, and Empowering Multilingual Interactions
The core concept is teaching model trivial tasks the way babies learn during their initial years. Words-shuffle, adding special characters, completing sentences, filling in the blanks, and adding the task of translation. The 8-tasks (T8) model architecture dramatically improved the translation adapter compared to just one task of translation. At its core, it remains a sequence-to-sequence (encoder-decoder) style model. We also binarize the model to get faster inference.
We are working on a project that aims to evaluate AI models and projects and provide governance ratings. We would check Fairness (language, community, etc), data security, and the unbiased nature of the model.
More details coming soon ...
Bolee Innovation Labs was envisioned by "Swapnil Ashok Jadhav" who has 12+ years of experience in Data Science, NLP, ML, CV, Speech and now GenAI. He has worked for notable companies like Walmart, Unacademy, Dailyhunt, Yubi, Haptik and Rediff.com.
He is a supporter of open-source tech and believes in Data more than algorithms. Better the Data .. better the outcome of AI systems. He has recently become avid student of Psychology and Unicellular Organisms.
LinkedIn: https://www.linkedin.com/in/swapn
Mumbai, Maharashtra, India
Email ID: boleeinnovationlabs@gmail.com LinkedIn: https://www.linkedin.com/company/bolee-innovation-labs/
We use cookies to analyze website traffic and optimize your website experience. By accepting our use of cookies, your data will be aggregated with all other user data.