Key Points
Businesses, similarly, can build multilingual virtual assistants simply by adding local content (documents, databases, etc.) and training the model on it, Ankush Sabharwal, co-founder and chief executive officer (CEO) of CoRover, told me..
While ChatGPT, the chatbot developed by OpenAI, and most other LLMs in the world are trained predominantly from English databases, companies working on Indian LLMs have the unenvious task of training their systems on languages that arent fully digitized..
The lab has also partnered with Sarvam AI, a GenAI startup founded by Vivek Raghavan and Pratyush Kumarboth were co-founders of AI4Bharatto develop LLMs specifically for India called the OpenHathi Series. Sarvam AI, on its part, say it will work with Indian enterprises to co-build domain-specific AI models on their data..
It is Indias first full-stack AI" solution; it is a GenAI foundational model, built from scratch; it is trained on more than two trillion tokens and is comparable to GPT-4, created by OpenAI; it can understand 20 Indian languages and generate content in 10 Indian languages including Marathi, Hindi, Telugu, Kannada, and Odia..
Many of the 22 official Indian languages do not have digital data, which makes it challenging to build and train an AI model with local datasets..
You might be interested in
Why India risks falling behind in the AI race
30, Jun, 23Indias startup landscape, meanwhile, is caught in a time warp, with embarrassed investors marking down their stakes in Byjus, an online education company collapsing under the weight of its own reckless growth.
How are Indian languages faring in the age of AI and language models?
30, May, 23As large language models like ChatGPT find more applications around the world, their adoption also passively spreads a prejudice against languages other than English, including Indian languages. Some researchers are working to remedy this.
Here's Proof You Can Train an AI Model Without Slurping Copyrighted Content
20, Mar, 24OpenAI claimed it's "impossible" to build good AI models without using copyrighted data. An “ethically created” large language model and a giant AI dataset of public domain text suggest otherwise.
Free online AI training programme in Indian languages launched
16, Jul, 23Pradhan, the Minister for Education and Skill Development and Entrepreneurship, said technology should not be a prisoner of language, and called for tech courses in Indian languages. He added that this is a good beginning towards dismantling language barrier in technology education and future-proofing our Yuva Shakti, particularly those in rural areas, an official statement said. The minister also said that India is a technology-savvy country and the success story in adoption of digital payments in India is a case in point.
AI models make stuff up. How can hallucinations be controlled?
03, Mar, 24It is hard to do so without also limiting models’ power
Indian AI: What is it, and can we make one?
19, Dec, 23What is colloquially referred to as ‘Indian’ AI is currently aspirational, referring to datasets that foundational AI models are trained on.
Mint Explainer: The mercurial rise of India-focused LLMs
18, Dec, 23Ola co-founder Bhavish Aggarwal’s new company, Krutrim, is the latest to join the growing band of companies that are focusing on building India-specific large language models–the need of the hour but easier said than done, given high computing costs and paucity of good Indian datasets