How can Artificial Intelligence improve African agriculture?

Published

By: Erik Bongcam-Rudloff, Professor of Bioinformatics at the Department of Animal Breeding and Genetics, SLU

As climates change and populations increase, Artificial Intelligence (AI) will be a key player in Africa in the creation of technological innovations that will improve and protect crop yield and livestock. 

Participants at “Network of Excellence in Artificial Intelligence for Development in sub-Saharan Africa” in Nairobi, Kenya, April 2019. Photo: Erik Bongcam-Rudloff

The work creating technologies that allows computers and machines to function in an intelligent manner is known as Artificial Intelligence or AI. The advantages of using AI based devices or systems are their low error rate and huge analysis capacity. If properly coded the AI systems have incredible precision, accuracy, and speed. They can also work independently in many, for humans, hard conditions and environments. One of the most interesting areas where AI is breaking into is agriculture. 

One area using AI and attracting a lot of attention is the area more known as “Precision Farming”. Precision Farming generates accurate and controlled technologies for water and nutrient management. It also gives optimal harvesting, planting times and produce solutions in many other aspects of modern agriculture.

In April 2019 a workshop was held at Strathmore University, Nairobi in with the aim to set up a “Network of Excellence in Artificial Intelligence for Development in sub-Saharan Africa”. There where 60 international participants by invitation. The meeting was supported by Swedish SIDA and organised by the International Development Research Centre and Knowledge 4 Foundation (K4A).

Plenary discussions. Photo: Erik Bongcam-Rudloff

The main goal of the workshop was to discuss the AI field with a bottom-up approach. The objectives of the workshop were to define the African Machine Learning and Artificial Intelligence (ML/AI) landscape, to create an African research roadmap and to find ways to incorporate cross continental development. Around these objectives, four thematic areas of discussion were developed: governance, skills/capacity building, applications and others. 

Discussions during a break. Photo: Erik Bongcam-Rudloff

On the last day of the workshop we visited the IBM Research – Africa in Nairobi. The staff at IBM-Africa presented several AI projects and one example related to the future of AI in agriculture was presented by Juliet Mutahi, a software Engineer working at the IBM Nairobi THINKLab. She presented “Hello Tractor” a system comparable to Uber for taxi but in this case a system that allows farmers to share tractor resources by using an app on their smartphones. This is the kind of initiatives that are created in Africa as a bottom-up approach. Juliet told the audience that she got the idea to create this system inspired by the work and needs of her parents that are coffee farmers in Kenya.

Juliet Mutahi software Engineer, IBM Nairobi THINKLab. Photo: Erik Bongcam-Rudloffï»ż

While identifying the different AI actors in the African continent, another initiative stood out among many: the “Deep Learning Indaba” initiative. This is an annual meeting of the African machine learning community. In 2018 the meeting took place in Stellenbosch, South Africa and gathered 600 participants from many African countries. The next annual meeting will take place in Nairobi, Kenya in August 2019 and the aim for this year is to gather over 700 participants. This shows the strength and vitality for this area of research in the Africa continent.

Many issues connected to agriculture will in the future be better handled using machine learning and artificial intelligence because AI can automate tasks that require human-level intelligence or beyond. This makes solutions that integrate AI better than today’s technologies. Most researchers involved in development research will in the near future learn how to use and how to incorporate AI in their work. Our young colleagues in the “Deep Learning Indaba” community are showing the way. The work in creating the “Network of Excellence in Artificial Intelligence for Development in sub-Saharan Africa” is just one of the building blocks in this process and SLU will be part of it.

Final panel discussion. Photo: Erik Bongcam-Rudloff

Watch an interview with Erik Bongcam-Rudloff talking about African bioinformatics and AI filmed at the Network of Excellence in Artificial Intelligence for Development in Nairobi, Kenya.

Report from the SIANI Annual Meeting 2019

Published

By: Dr. Alin Kadfak, Department of Urban and Rural Development, SLU

Did you know that about half of migrants are women? And more migrants migrate within developing countries than crossing North-South borders? Migration does not refer only border crossing, but moving from rural to urban as well as rural to rural within the same country. There are many misperceptions about migration and the root causes of the phenomenon, which bring us to this year SIANI Annual Meeting’s agenda!

Every year the SIANI secretariat organises a meeting in Stockholm so our members have a chance to interact with each other and to provide input for the work plan of the year ahead. This year’s theme ‘Migration, Agriculture and Rural Development’, held on 23 January, is in the spotlight as we explore different dimensions of migration and its connections with food and agriculture. Together the members can reflect on this vital topic. And this year, the meeting is well attended from academic, NGOs, government agencies and civil society.

Starting off with welcoming speech by Annika Åhnberg, Chair of the SIANI Steering Committee, who reminds us that forced and voluntary migrations are parted of the human history, and we need to understand the phenomenon in the holistic way. Our first speaker, Sigrun Rawet, SIPRI, brings to the meeting the discussing around recent UN Security Council Resolution on hunger and conflict. No doubt that conflict brings hunger, but what if ‘ending hunger can reduce conflict!’. UN World Food Programme is now doing pilot projects in four countries, hoping to reduce famine, the main cause of conflict.

Ingela Winter-Norberg, Sida, raises an important point that often refugees and immigrants are being excluded from development policy. The key question we need to ask is ‘How can we increase economic self-reliance activities for migrants?’ to ensure that they can support themselves economically in countries of destination. 

Our next speaker Jesper Bjarnesen, Nordic Africa Institute, conveys a strong message that ‘migration is by far the most positive than negative, but it has been hindered by regulations’. And the challenge to migrant problem is when the government sees migrants as threat instead of source of labour. The way forward, he suggests, is to shift the narrative from ‘migrant rights’ to ‘labour rights’.

Aster Asgedom, County board VÀstra Götaland, shares how she continues supporting rural development back home in Ethiopia, by connecting the supports from Swedish NGOs, civil societies and academic. Being migrant herself, she reflects on how nature is very important for integration process. For instance, Aster together with other organisations in Gothenburg organised outdoor activities to welcome new refugees into the country.

Round Table Dialogue is the highlight of the day. This interactive platform asks members to join and help answer ‘How can you together with other SIANI members help to minimise non-voluntary migration and address its root causes?’. Each group has one and half hour to brainstorm, discuss and agree on the main statement to help directing SIANI’s work plan for 2019. Please stay connected to see the results from Round TableDialogue at SIANI.se, with more activities to continue the migration and rural development dialogue.

Link to SIANI’s webpage with videos and documentation from the meeting.

Pushing back against disease with participatory epidemiology

Published

By: Dr. Alin Kadfak, Department of Urban and Rural Development, SLU

A herd of animals can be the only source of income for many families in areas like northern Uganda or southern Africa. Farmers with poor access to infrastructure and capital struggle to address and manage disease outbreaks, which can wipe out a household’s economy. Worldwide 750 million low-income livestock farmers depend on healthy animals for their productivity and livelihood. So, yes! Animal health matters.

But so far, we know very little about how perceptive farmers are and what their response to disease outbreaks is. A seminar ‘Exploring the Whys and the Hows of epidemiology: Integrating social sciences with veterinary research for better understanding and action towards improved animal health in the Global South’  held at the Swedish University of Agricultural Sciences, SLU on the 20th of November, 2018 explored a broader picture of how scientists  across disciplines see  the way local farmers understand the diseases and prevention mechanisms.

Why epidemiology is important?

The role of epidemiology in animal health is to identify the risk factors and the causes of disease outbreak. This knowledge provides a base for interventions which can ultimately prevent an outbreak or an epidemic. With that, epidemiology is crucial for poverty reduction and plays a vital role in increasing resilience toward risks of epidemics in low-income countries.

Even though our scientific knowledge about animal diseases has stepped forward, disease control leaves much to be desired. For instance, African swine fever continues to spread from Africa to Europe and Asia, harming many pig farms.

Karl StÄhl, Swedish National Veterinary Institute (SVA), traced an African swine fever outbreak through a trade route in Northern Uganda. He found that local pig farmers knew very well the clinical signs of the disease and how it spreads. They could even identify the potential control mechanisms. But as soon as they noticed sick pigs, they sold them to secure their income.

It became clear for Karl and his team that what they were facing in Northern Uganda is not the lack of knowledge by the farmers, but rather the lack of information on specific needs and circumstances of pig producers and supply chain actors during the outbreak. To solve the problem, Karl suggests it is necessary to bring in societal aspect into the equation. In his opinion this will help to find solutions that fit with local socio-economic and cultural context.

Is interdisciplinarity the way forward?

Social science still remains on the periphery of disease outbreak management – social sciences are rarely integrated into epidemiology. However, it is becoming apparent that quality disease prevention isn’t possible without considering socio-cultural, economic and political dimensions. This engagement can be done through an interdisciplinary approach.

Learning from the Ebola outbreak in West Africa, Paul Richards (Wageningen University) suggests to look beyond the classic idea of epidemiological and biomedical approaches and take into account human emotions.

During the early phase of the Ebola outbreak, epidemiologists built Ebola treatment centres to bring in patients, or suspected patients, hoping to isolate the patients’ from the outside world. But this created anxiety among patients and their families. The disease continued to spread.

By 2013, Paul together with locally trained field assistants used anthropological approach to collect information through real-time action-research. Human emotion appeared to be the missing link to the treatment – families needed to be able to talk to each other and have meals together. As a result, Paul’s team recommended to build more community quarantines so that families could be close the patients. Through watchful care and close communication between patients and their communities, the epidemic was eventually brought under control.

Participatory epidemiology

Participatory epidemiology has been influenced by Participatory Rural Appraisal (PRA), an approach that puts local participation at the heart of the process. However, participatory epidemiology has been adopted largely by natural scientists, who usually focus on m research quantitative methods. Meanwhile, from the experience of social scientists, statistics do not always reflect the reality of local communities. So, well-tailored solutions require real participation form the local people.

Klara Fischer and Erika Chenais’s work in Uganda is an attempt to engage a local participation in an epidemiology study. Instead of using a top-down approach, Klara and Erika let local perspectives guide their study. They used several participatory methods, such as participatory wealth-ranking and focus group discussion to understand how a community can prioritise, understand and deal with cattle diseases.

Susanna Sternberg Lewerin confirmed the value of a participatory approach in disease prevention practices. For her, it was important to create a relaxed atmosphere during participatory discussions and recruit local facilitators who could facilitate in the local language.

So, how can diseases control benefit from cross-pollination between veterinary epidemiology and social sciences? There are at least three ways:

  1. Veterinary epidemiology and social science can cooperate and increase understanding of contextualised knowledge that affects decisions relating to disease control.
  2. Using participatory methods to engage indigenous perspectives can improve disease prevention and control mechanisms.
  3. It is key to remember that there are no ‘fixed-tools’ in epidemiology. Tools need to be adapted and adjusted to the local context, addressing local heterogeneity, power dynamics and conflicts.

This article was originally published at SIANI.