BHARATGPT GROUP UNVEILS ‘HANOOMAN’

1. In February, 2024 the BharatGPT group, led by IIT Bombay and seven others, revealed plans to introduce its first ChatGPT-like service.
2. They developed the ‘Hanooman’ series of Indic language models with Seetha Mahalaxmi Healthcare (SML). The initiative is supported by Reliance Industries Ltd and the Department of Science and Technology.
3. Hanooman consists of large language models (LLMs) that can communicate in 11 Indian languages like Hindi, Tamil, and Marathi. There are plans to extend its capabilities to over 20 languages.

1. Its applications span across healthcare, governance, financial services, and education sectors.
2. Hanooman is more than just a chatbot; it's a multimodal AI tool.
3. One customized version, VizzhyGPT, is specifically tuned for healthcare using extensive medical data.

Generative Pre-trained Transformers (GPTs)

1. GPTs are large language models (LLM) that use transformer neural networks to generate human-like text.
2. Trained on vast amounts of unlabeled (data with no labels/tags) internet text data, GPTs understand and generate coherent (logical and well-organized) and relevant text.
3. They can be adapted for various tasks like language generation, sentiment analysis, machine translation, and text classification.
4. GPTs utilize self-attention mechanisms to focus on different text parts, capturing more context and enhancing performance in natural language processing (NLP) tasks.

• Natural language processing (NLP) is a computer program's ability to understand spoken and written human language. NLP is a part of artificial intelligence (AI).

Large Language Models (LLMs)

1. LLMs, utilize deep learning methods to process vast amounts of text data.
2. They analyze large text datasets to grasp the structure and meaning of the text and draw insights from it.
3. LLMs are trained to identify meanings and connections between words, enhancing their text comprehension and generation abilities.
4. The effectiveness of LLMs increases with the amount of training data they receive.
5. Training data for LLMs includes sources like Wikipedia, OpenWebText, and the Common Crawl Corpus, providing extensive text information for the models to learn and generate human-like language.

Applications of LLM

1. Medical: Aid in tasks like protein structure prediction for disease pattern recognition and outcome prediction.
2. Retail: Enhance customer experiences through interactive chatbots.
3. Software: Assist in writing software and teaching robots’ physical tasks.
4. Finance: Summarize earnings calls and generate transcripts of important meetings.
5. Marketing: Organize and analyse customer feedback.

Challenges in Developing LLMs

1. Require significant capital investment, large datasets, technical expertise, and computing infrastructure.
2. Sourcing quality datasets in Indian languages poses a major obstacle.

GPT vs LLM

1. GPTs, a type of LLM, use deep learning to generate human-like text.
2. They are "generative," "pretrained," and based on a transformer neural network architecture for text processing and generation.

Top of FormWhat is ChatGPT?

1. ChatGPT is an advanced natural language processing (NLP) model made by OpenAI.
2. It's a version of GPT-3 (Generative Pretrained Transformer 3), trained extensively on text data to mimic human-like responses.
3. The chatbot's responses aim to be technical but clear, avoiding complex terms.
4. It enables natural conversation between users and the virtual assistant.

Must Check: Best IAS Coaching In Delhi