Precision Aquaculture with Inclusion of Commodity Smart Phones and Collective Learning

Introduction

Since demand for seafood is constantly growing, interest in aquaculture is on the rise. On the other hand, meticulous attention to various aspects of the cultivation process, such as stocking, feeding, and safeguarding against predators and disease is mandatory for the success of aquaculture [1, 2].

Precision aquaculture is a concept that advocates for the integration of emerging technologies to enhance the efficiency and productivity of aquaculture farms [3]. It encompasses various aspects of aquaculture, including automatic feeding, water quality measurement, disease detection, and more. The use of advanced sensors, data analytics, and automation systems enables real-time monitoring and targeted interventions to optimize production and minimize risks.

Opinion

Nevertheless, implementing a precision aquaculture farm can be expensive due to the need for advanced technologies and infrastructure [4]. Statistically speaking, a significant number of aquaculture farms falls within the category of small to medium-sized farms, often managed by individuals with moderate to modest incomes [5].

Although, it needs to be mentioned that the potential long-term benefits in terms of improved production efficiency, resource management, and profitability should be considered when evaluating the feasibility and cost- effectiveness of precision aquaculture. Technological advancements and research breakthroughs have also successfully addressed many of the challenges associated with implementing precision aquaculture farms, painting a promising picture for the future [6]. Nevertheless, despite that, implementing a precision aquaculture farm is usually economically infeasible for individual farmers and small sized farms.

In the following two technologies will be highlighted that may benefit both farms with and without precision aquaculture. First, collective learning will be presented as a technology that helps training artificial intelligence algorithms with richer data. Second, commodity smartphones will be introduced as an enabler for small farms without precision aquaculture.

Collective Learning

One of the emerging technologies that can be integrated into precision aquaculture is the use of machine learning algorithms. Typically, training these algorithms with high- quality and extensive data is a common requirement, which can pose a challenge when the algorithm is intended for specific use cases with limited available sample data [7]. An effective solution to address this challenge is the utilization of collective learning algorithms.

Collective learning algorithms are referring to a family of algorithms where the goal is to train a high-quality centralized model while training data remains distributed over many clients [8]. Federated learning as one of these algorithms is a decentralized learning framework that allows participating sites to collaborate and share data, models, and training processes [9].

Such technology can be utilized among existing farms to train machine learning models that are trained on collective data from all participant farms which results in more accurate models due to richness of data. It’s important to note that data sharing needs do not infringe upon privacy concerns, as existing technology can effectively address these issues. There could be many applications for these systems in aquaculture such as early detection of aquatic diseases that is one of the main challenges that farmers face. Moreover, other information such as information about the spread of a disease for example could help both farmers and policy makers to plan appropriately.

Consider a scenario where a disease is spreading within a specific geographic area. Early detection will become possible if it reveals that many farms in a neighborhood are detected with a disease. In such cases if the information is shared between other farmers, they can benefit from it and take necessary actions to protect their farm and maybe stop the spread of disease to other farms. While it’s true that not all data from one farm is directly applicable to another due to unique factors at play in each farm, there’s still valuable information to be gleaned.

Potential of Smartphones

It’s important to highlight that the integration of collaborative learning can be considered a step towards adopting precision aquaculture. This can be accomplished in two ways: participating in collective learning by contributing data or utilizing models trained with data from other farms. This feasibility is largely attributed to the widespread availability of affordable smartphones with adequate computing capabilities.

In addition to the mentioned applications, there are numerous other potential use cases for smartphones that can benefit small and medium-sized farms by harnessing technological advancements. Government and non- government agencies can offer valuable information to farmers through mobile apps, facilitating effective communication and support.

Conclusions

While not all farms can afford implementation and benefiting from precision aquaculture, there are some enabling technologies that allow these farms to benefit from some crucial aspects of it. Collective learning helps training of machine learning models train by rich amount of data shared by all participating farms. This data sharing is not only limited to farms with precision aquaculture but also by farms without precision aquaculture thanks to the wide accessibility of smartphones these days. A crucial application of these technologies could be early detection and localization of viral diseases that can easily transfer between different farms.