New Google Tool Aims to Make Indian Agriculture More Efficient and Data-Centric

Google is set to introduce the Agricultural Landscape Understanding (ALU) Research API , a tool aimed at enhancing data-driven agricultural practices, with limited availability.

Jeanine Banks, Vice President and General Manager of Developer X at Google, made this announcement during the Google I/O Connect event in Bengaluru 2024.

The ALU Research API is designed to address challenges in farming , such as improving crop yields, facilitating access to capital, and enhancing market access for farm products. Google has partnered with entities like Ninjacart, Skymet, Team-Up, IIT Bombay, and the Government of India to explore the application of this tool.

The ALU API will deliver detailed landscape insights at the farm field level, a critical aspect for transforming the agricultural ecosystem. It aims to address the diverse landscape and crop requirements that vary even among neighboring fields.

Currently, agricultural insights are available at an aggregate level, but the need for individual farm-level data is essential. Utilizing high-resolution satellite imagery and machine learning , the tool plans to delineate field boundaries, helping to tackle issues like drought preparedness, irrigation management, and market access. The API provides granular data, including crop type, field size, and proximity to water, roads, and markets.

In the realm of Indian languages, Google DeepMind India unveiled updates to support developers in building language solutions for India. This includes the expansion of Project Vaani, in partnership with the Indian Institute of Science (IISc), offering over 14,000 hours of speech data across 58 languages from 80,000 speakers in 80 districts. Additionally, the team introduced IndicGenBench, a benchmark for evaluating the generation capabilities of Large Language Models in Indic languages, covering 29 languages.

Google has also open-sourced the CALM (Composition of Language Models) framework, enabling developers to integrate their specialized language models with Gemma models, creating solutions that account for India's linguistic diversity.

Ambharish Kenghe, Vice President at Google, emphasized the company's commitment to empowering Indian innovators to leverage AI's potential, addressing India's unique needs, and shaping the future of AI globally. He highlighted the vast opportunities in multimodal, mobile, and multilingual AI, expressing excitement about contributing to India's AI journey.

Google is also collaborating with the MeitY Startup Hub to support 10,000 Indian startups in their AI endeavors through Google Cloud credits, an AI-first programming curriculum, and initiatives like a nationwide GenAI Hackathon and AI Startup Bootcamp. Indian developers now have increased access to Google’s advanced AI models with the expanded 2 million token context window in Gemini 1.5 Pro and the upcoming Gemma 2 models.