Large Language Models

Context:

As per Principal Scientific Advisor, India will set up a “high powered committee” to explore the development of Large Language Models (LLMs), tools that harness Artificial Intelligence to create applications that can understand and process human language.

Relevance:

GS III: Science and Technology

Dimensions of the Article:

1. Large Language Models (LLMs)
2. Deep Tech
3.  Draft National Deep Tech Startup Policy (NDTSP), 2023

Large Language Models (LLMs):

• LLMs are a distinct category of generative AI models designed to understand and generate human-like text.
• They leverage deep learning techniques, particularly neural networks, in their construction.
• LLMs are capable of producing coherent and contextually relevant text when provided with a prompt or input.

Prominent Example:

• One of the most renowned instances of LLMs is OpenAI’s GPT (Generative Pre-trained Transformer).

Generative AI:

• Generative AI is a segment of artificial intelligence that concentrates on developing systems capable of generating content resembling human-produced material.
• These systems learn from patterns in existing data and employ that knowledge to produce fresh, original content across various forms, including text, images, music, and more.

US-India Collaboration:

• India and the United States have forged a strong partnership, particularly in the domain of deep tech collaboration.
• India’s draft policy on deep tech highlights the significant presence of over 10,000 startups across various deep tech sectors in its Startup India database, aligning well with the collaborative potential of the U.S.-India relationship.

Deep Tech:

• Deep tech, short for deep technology, represents a category of startup businesses that center their innovations on substantial engineering breakthroughs, scientific discoveries, and technological advancements.
• These startups typically operate within fields like agriculture, life sciences, chemistry, aerospace, and green energy, although their scope is not restricted to these areas.
• Key deep tech domains encompass Artificial Intelligence, advanced materials, blockchain, biotechnology, robotics, drones, photonics, and quantum computing. These fields are transitioning rapidly from early-stage research to practical market applications.

Impact of Deep Tech:

• Deep tech innovations are characterized by their radical nature, capable of disrupting existing markets or establishing entirely new ones.
• These innovations often have profound effects on individuals, economies, and societies.
• The development timeline for deep technology is significantly longer than that of shallow technology, as exemplified by mobile apps and websites.
• Achieving market-ready maturity for deep tech innovations typically involves extensive research and development, prototyping, hypothesis validation, and technology refinement. For instance, artificial intelligence has been in development for decades and is still evolving.

Challenges Faced by Deep Tech Startups:

• Securing funding is a primary challenge for deep tech startups, with fewer than 20% of them successfully obtaining financial support.
• Government funds are often underutilized, and there is a shortage of domestic capital available for such startups.
• Additional challenges encompass talent acquisition, accessing relevant markets, obtaining research guidance, enhancing investors’ understanding of deep tech, customer acquisition, and managing talent-related costs.

 Draft National Deep Tech Startup Policy (NDTSP), 2023

The Draft National Deep Tech Startup Policy (NDTSP), 2023, is a proposed policy framework with the following key objectives and measures:

Objectives:

• Support Research and Development: The policy aims to strengthen research and development efforts in deep tech startups. These startups focus on addressing fundamental and technical challenges rather than simply commercializing existing technologies.
• Financial Support: It seeks to identify approaches to provide financial support to deep tech startups during critical stages of their development, particularly before they introduce their products or ideas to the market.

Facilitation of Startups:

• Intellectual Property Simplification: The NDTSP suggests simplifying the intellectual property regime for deep tech startups, making it easier for them to protect and leverage their innovations.
• Regulatory Easing: The policy aims to reduce regulatory requirements, making it more straightforward for deep tech startups to operate and innovate.
• Export Promotion Board: It proposes the creation of an Export Promotion Board to lower entry barriers for Indian deep tech startups in foreign markets. It also recommends including provisions in foreign trade agreements to facilitate market access for these startups.

-Source: The Hindu