The Most Incredible Agriculture Technology Innovations You Have Never Heard Of
Earlier this month, Pinduoduo and China Agricultural University co-organized the Smart Agriculture Competition in collaboration with the Food and Agriculture Organization of the United Nations. Through the competition, 4 technology teams employed data analysis, intelligent sensors, and greenhouse automation to grow strawberries. Their efforts yielded 196% more strawberries than their traditional grower counterparts.
As a part of the event, Pinduoduo also hosted several agriculture technology experts from around the world: Ranveer Chandra, chief scientist of Azure Global at Microsoft; Lesly Goh, former CTO of the World Bank; and Hila Cohen, head of business development at the World Food Program’s Innovation Accelerator. Each of these leaders shared insights on the challenges and opportunities that agriculture development faces.
Ranveer Chandra on democratizing access to data driven farming
Ranveer shared more about the cutting edge work of FarmBeats, a project within Microsoft Research that enables more food to be grown to feed the growing global population without harming the environment. Specifically, FarmBeats increases the feasibility and impact of incorporating data driven practices in agriculture, which ultimately improves yields, reduces costs, and demands less of the environment.
Through data driven practices, farmers can map their soil moisture, pH, temperature, and nutrient compositions at the very nuanced level. With this information, farmers can apply water and fertilizer only where and when needed to conserve resources. However, data driven farming is not widely adopted largely because of the cost, so FarmBeats has worked to lower the cost in the following key ways:
- Connectivity: FarmBeats realized that rural areas had poor internet connectivity but had many empty TV channels, so they embedded WiFi signals in these empty channels in a way that preserves propagation efficacy while not interfering with TV reception in adjacent channels.
- Accuracy: FarmBeats knew that it was quite expensive to put a sensor into the ground every 10 meters to comprehensively measure soil composition. So, they developed AI to combine available sensor data with aerial imagery from drones and satellite imagery in order to predict real time soil characteristics in areas without physical sensors. Through combining multiple different data sources (not only the sensor data but also the drone and satellite imagery), the prediction results improved by 3X in accuracy. Here, FarmBeats faced and tackled an additional challenge: satellite imagery is shrouded by clouds 77% of the time, so FarmBeats combined optical images with radar data from SAR satellites to leverage multispectral imagery to reconstruct any obstructed image data. With this enhanced technology, FarmBeats could not only see through the clouds but far below as well.
- Last mile data delivery: FarmBeats found that many farmers did not have the technology needed to receive and work with farm data. So, they shipped farmers a PC that does compute in the farmer’s home or office using vision based stitching and machine learning on the edge. Subsequently, summaries are sent to the cloud, and the cloud merges this information with other data streams to return the data to farmers.
- Local specificity: to make the data hyperlocal, FarmBeats leveraged transfer learning. In one example of the powerful results, a farmer was able to see how his cows were pooping and how his grass was growing, which were key indicators for his farm health.
Ranveer and the Microsoft team hope that FarmBeats can be a platform for any farmers with a smartphone to gain a deeper real time understanding of their farms and formulate a more targeted and resource efficient strategy to farming. For example, with just a smartphone, farmers no longer need access to expensive sensors to measure soil moisture; instead, the time it takes the phone signal to travel through the soil can indicate the level of moisture (because moisture impacts conductivity).
Through FarmBeats, more machine learning engineers can improve and add algorithms on the platform, and partners can add features to build more solutions for farmers. For broader farm planning, people can even create and customize 3D farm simulations.
In the long term, FarmBeats aims to provide training and education programs to bring AI and data knowledge to farmers everywhere.
Hila Cohen on using frontier technologies to drive agriculture innovation
Hila shared more about the World Food Program’s Innovation Accelerator which provides hands-on, on the ground support in launching and running pilot programs for agriculture technology startups in locations including Tanzania, Jordan, and South Sudan.
She also underscored the program’s interest in frontier innovations leveraging futuristic technologies. She shared several of these:
- SKAI: the platform maps houses in advance of disasters so when there is a climate shock, the system can compare the before and after images and better predict the level of damage in the next disaster
- Building Blocks: ablockchain solution has been developed to help people in Bangladesh safely get their World Food Programme entitlements during COVID
- EShop: in the wake of COVID, this platform has helped people in Somalia order from the safety of their homes
- H2Grow: hydroponics are an incredibly efficient way to farm; they need less soil, less water, and less space for a given production level. However, hydroponic construction is very expensive, which has kept the use of the technology largely outside of the developing world. Through H2Grow, hydroponics were brought to remote areas in impoverished areas, including in the middle of the desert. H2Grow found local materials to reduce the cost of construction (from $40K to $3K in one instance!).
Lesly Goh on how agriculture and fintech can improve women’s outcomes
Lesley underscores several key challenges that that agriculture industry faces, as outlined in her forthcoming Brookings Institute chapter on digital agricultural transformation:
- Agriculture makes up 26% of employment globally, but 700 million of these people live below the poverty line
- 1 billion of the 1.7 billion people excluded from the financial system are women
- The gender gap in access to financial markets is as great as 30% in some countries
- Women are disproportionately hit by COVID induced agriculture setbacks because they bear the bulk of the burden of home and farm labor
- If women participated in the economy equally, global GDP could increase by $28 trillion by 2025
Lesley also outlines several key innovations to grow the productivity of agriculture and lift people out of poverty:
- Farmer advisory: provide farmers with comprehensive information on prices, crop planning, weather, pests, and disease control
- Peer to peer lending: enable more farmer to access credit with lower interest rates
- Traceability: first mile data collection to enable traceability across supply chains
- Digital marketplace: connect farmers to buyers and suppliers more seamlessly
- Mechanization platform: support farm activities with machinery like robotics and drone imaging
- Digital financial products: in collaboration with the Better than cash Alliance, Women’s World Banking, and the World Bank, better underwrite small and medium sized businesses (which are predominantly run by women) to provide better structured government payments and loans
- Trusting computing environment: ensure holistic security and privacy as more technologically driven solutions are developed and deployed
Cheng Biao, the leader of the winning team in Pinduoduo’s Smart Agriculture Competition, notes that in order to create more sustainable agriculture innovations, collaboration from both the farmers and technologists are needed. As Adre Zhu, Pinduoduo Senior Vice President notes, “technology is the force multiplier that helps both the people who grow food and the people who eat it. We want to enable farmers by giving them a tech toolbox that takes the guesswork out of planting.”
The future is indeed bright for agriculture, and Pinduoduo is excited to be a part of supporting the next generation of agriculture innovators.