How AI is Helping This IIT Alumnus Who Has Pledged to Save The Aquaculture in India

The demand for seafood in India is growing faster than ever\r\nand consumers are becoming more interested in the nutritional advantages of\r\ntheir food choices. As maintaining ocean health and wild fish stocks is\r\nbecoming a major concern, it is becoming increasingly important to take care of\r\nthe aqua community.

Indian aquaculture is a $7 billion worth industry and only a\r\nhandful of technology startups are working on this domain in India. Rajamanohar\r\nSomasundaram, an IIT Alumnus, is taking the lead in the aquaculture industry\r\nwhere he is using the power of IT, artificial intelligence and IoT skills to\r\nget a profound impact on the industry. Using these technologies he has been\r\nable to positively impact the lives of 3,000 aqua farmers in coastal India and\r\nis now planning to expand internationally to run pilots in Aceh and Sulawesi by\r\nthe end of 2020.

Analytics India Magazine got in touch with Somasundaram\r\nto get an inside view on how he is promoting sustainable fish and shrimp\r\nfarming using AI. Founded in 2017, Aquaconnect is a full-stack aquaculture technology\r\nventure that offers data-driven farm advisory solutions and market place\r\nsolutions to Shrimp and Fish farmers. They have an AI-powered farm advisor tool\r\nFarmMOJO, that is helping farmers to enhance their farm productivity.

Analytics India\r\nMagazine: Please tell us about your initiative Aqua Connects and how did it\r\ncome into existence?

Rajamanohar Somasundaram: It was a serendipitous\r\nmeeting with a shrimp farmer, Sanjai that sparked the idea of Aquaconnect.\r\nDuring the conversation with him, I got to learn about the shrimp farming\r\nindustry and its challenges such as high production cost and diseases\r\noutbreaks.

With further research, I learned that Aquaculture in India\r\nhas evolved as a viable commercial farming practice and has been showing an\r\nimpressive annual growth rate of 10-15% every year. The industry exports are\r\nvalued at around 7 billion USD in 2018. India stands top on shrimp production\r\nand it contributes 70% of Indian aquaculture export value.

Though it is a 7 billion dollar industry, it still lacks the\r\ntechnology adoption and efficiency it is expected to have. Around 2 million\r\nrural farmers and coastal communities depend on shrimp & fish aquaculture,\r\nwhere traditional farming practices prevents them in achieving production\r\nefficiency and diseases prediction.

Having founded 3 technology ventures in past, I sensed an\r\nopportunity to use artificial intelligence technology to help solve the fish\r\n& shrimp farmer’s problems. My research pointed to the fact that a lack of\r\ndata-driven farming is one of the biggest challenges yet to be addressed by the\r\nindustry to achieve sustainability.

I started Aquaconnect with Sanjai and Shanmugam with a goal\r\nto promote a pioneer Indian full-stack aquaculture technology company that\r\npromotes the data-driven farming practice. Our data-driven marketplace helps\r\naqua farmers to connect with the upstream and downstream supply chain in the\r\naqua industry.

AIM: How has\r\nartificial intelligence helped in achieving your goal of positively impacting\r\naqua farmers in a coastal area?

RS: It has\r\nimpacted us in 3 key areas:

Impact on aqua farmers: AI helps farmers in various\r\naqua farming activities such as decision making, production planning, disease\r\nprevention and growth management. And AI unleashes the real potential of farm\r\ndata and encourages farmers to practice data-driven farming. Our AI advisory\r\ngives transparency in farming activity and advisory to regulate and control the\r\noperational cost.

Feed usage optimisation: Often, unregulated feeding\r\nresults in high production cost and put pressure on limited resources. Feed\r\nconstitutes about 70% of the production cost and it is necessary to optimize\r\nthe feed usage to achieve efficiency and control cost. Growth of the animal\r\ndepends on the right feed usage. Bringing feeding efficiency can help farmers\r\nto have better returns from their farms. Artificial intelligence regulates feed\r\nusage by analysing the data from the pond level and creates patterns to avoid\r\nfeed waste and ensures the right quantities. FarmMOJO’s AI helps farmers to\r\nreduce 10% feed cost through its feed efficiency model.

Further, farm care ML algorithm analyzes the performance of\r\ncompetitive farm care products on the ground (Feed, Probiotics, Minerals) and\r\nbased on their past performance it provides a personalized recommendation to\r\nthe farmers.

Disease prediction and prevention: Implementing AI\r\ntechnologies in aquaculture make possible for aqua farmers to detect diseases\r\nwell in advance and take preventive measures to secure the animals. We have\r\nbuilt a disease prevention model call “Morby-mass” that involves reducing the\r\noccurrences of diseases. During the pilot, FarmMOJO predicted the diseases in\r\nover 120 pounds and helped farmers to take corrective actions and also minimize\r\nthe losses.

AIM: What are the AI\r\ntools and technologies that you use for your workings?

RS: We use FarmMOJO AI technologies, frameworks such as\r\nTensorFlow, languages like Python, libraries such as NumPy, sci-kit-learn and\r\nalgorithms such as linear regression, random forest, kNN, K-Means.

AIM: Where are the\r\nvarious data points collected from? What is the kind of data that you work\r\nwith?

RS: We collect farm-level data such as pond size,\r\nstocking species and stocked volume. And also pond level data such as feed\r\ninputs, water quality parameters, and animal health status at regular\r\nintervals. This data helps us create feed efficiency model, growth model and\r\ndisease prediction models. Collected data will be analysed with our feed\r\nefficiency AI models to understand animal performance with respect to various\r\nfactors. Based on the AI module analysis, FarmMOJO gives a comprehensive report\r\non pond operations and suggests relevant products to be used.

AIM: Please elaborate\r\non the problem statement that you are aiming to solve highlighting how AI is\r\nused in the process. Please elaborate on the AI use case.

RS: The Fish and shrimp farmers have been facing\r\nseveral problems such as high costs in unscientific farming practice, constant\r\nmonitoring of water quality, feed intake, identifying anomalies and biomass\r\nconversion. Poor biomass conversion and disease management are some of the\r\nmajor concerns of the industry that has not been addressed yet. Also, the lack\r\nof data and data-analytics driven decision has a negative impact on production\r\nefficiency, traceability and disease management.

Therefore, a need for the technology that enables AI-driven\r\nreal-time analysis and promotes sustainable farming to achieve production\r\nefficiency, predict diseases and traceability. To address these challenges we\r\nhave built AI models that analyse the pond level data continuously such as\r\nmorby-mass model, pond Health Monitoring model, healthcare management model,\r\ngrowth Prediction Model and feed optimization model. They have helped in\r\ndetection and reporting issues at an early stage.

AIM: Please tell us\r\nabout FarmMOJO and how it works?

RS: FarmMOJO, an AI-based mobile farm advisor, record and\r\nmonitors the real-time production data gathered from the farms through the\r\nmobile application interface. Collected data will be analyzed in real-time and\r\nour prediction model uses the deep learning algorithm to provide\r\ncontext-sensitive suggestions and alerts to improve the water quality\r\nparameters, feed consumption pattern and health management. FarmMOJO\r\nunderstands the farmer’s needs inherently based on the input given by the\r\nfarmer as well as the data captured by the IoT/Smart farm management platforms.

FarmMOJO alerts the farmers about parameters which are not\r\nin optimal levels. For instance, at any point MOJO observes poor Feed\r\nConversion Ratio (FCR), it will suggest the necessary actions and relevant\r\nproduct to be used to normalize the pond environment to boosts FCR. It\r\nsimplifies the farm operations for the farmer and improves efficiency,\r\npredictability, and transparency.

With FarmMOJO’s data intelligence and location-aware\r\ncapabilities, we connect farmers with up-streams (Processors, Certifying bodies)\r\nand downstream ( Hatcheries, Feed and healthcare) of the aquaculture supply\r\nchain.

AIM: How is the reach\r\nof aqua connects? What are the various places/clients where it is being used?

RS: Currently, Aquaconnect works with 3000 farmers in\r\nvarious states of India ie., Tamil Nadu, Andhra Pradesh & Gujarat. We\r\nhave implemented our FarmMOJO application for 1900 pounds in a short period of\r\n9 months. We are piloting our solution in Sulawesi, Indonesia.

AIM: What is your\r\nroadmap from Aqua connects in the coming future? What are some of the goals you\r\nwish to accomplish?

RS: We are aiming to further expand into three states —\r\nTamil Nadu, Andhra Pradesh and Gujarat, deploy FarmMOJO in 8000 ponds and increase\r\nthe operating revenue to 400%. We also aim to increase our strategic\r\npartnerships to positively impact aquaculture production. We also plan to\r\nestablish our presence in Indonesia by the end of the year.

With FarmMOJO’s data intelligence, Aquaconnect aims to\r\ncreate a data-driven marketplace for Indian aqua Industry by connecting\r\nstakeholders under one umbrella. We are striving to create financial and\r\ninsurance products for the aquaculture industry so that farmers can access\r\nfarmer financial products and insurance coverage for farming activities.

AIM: What have been\r\nsome of the challenges in adopting AI into your project?

\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n

RS: Some of the challenges we faced were farm data\r\ncollection and integrity of the data, the steeps technology learning curve,\r\ncost of technology adoption, availability of large data sets and data integrity\r\nis one of the critical aspects for any successful AI project. Due to low\r\ntechnology adoption among Indian farmers, we found the data collection very\r\nchallenging in the initial days. It costs time, effort and capital for us to\r\nbring awareness among farmers and train them on FarmMOJO.

Source : Analytics India Magazine