On AI and Agriculture Sustainability. Q&A with Amanda McClerren

Growing up on a farm influenced my passion for science and ultimate career path. I’m a scientist at heart and have always liked data and facts and being objective. Great science is what motivates me – that and working with very smart and passionate people to achieve Bayer’s mission of “Health for all, hunger for none.”

1. You’ve recently been named CIO and Head of Digital Transformation & IT at Bayer Crop Science. What are your key responsibilities in this role, and can you talk about your path to getting here?

As CIO and Head of Digital Transformation & IT, I’m responsible for the digital transformation of all major functions within the Crop Science organization to enable business and digital strategy objectives. 

I’m a Biochemist by training and that is how I started out in my career. I’ve held roles in our Biotech & Breeding organizations and served as Head of Trait & Pipeline Delivery in R&D, where I partnered with other technology teams to deliver our global breeding pipeline and developed global protected culture facilities (glasshouse in Arizona) to transform how we design and create products for farmers – this project in particular was done in close alignment with our IT teams to deliver a truly digital, automated experience. 

I’ve spent the last four years leading IT for our R&D organization, driving our digital road map across multiple technical teams to deliver business outcomes that improve product design and advancement decisions and an industry leading field-testing platform across our Crop Science division to enhance product characterization and on-farm placement. 

Growing up on a farm influenced my passion for science and ultimate career path. I’m a scientist at heart and have always liked data and facts and being objective. Great science is what motivates me – that and working with very smart and passionate people to achieve Bayer’s mission of “Health for all, hunger for none.”

2. Can you share a couple key projects where you use AI to help farmers advance agriculture sustainably? 

AI is a core technology used to bring value to growers, optimizing productivity and enhancing on farm sustainability. Here are a few examples of how we’re using AI to advance agriculture sustainably: 

• AI-driven seed R&D: Bayer uses data science to accelerate the pace and success of plant breeding innovation by leveraging years of germplasm performance data together with genetic data on the latest material in our pipeline to accurately predict the performance of new potential products and eliminate years of costly field testing. Breeders use machine learning to prescribe new possible genetic combinations and anticipate a new plant variety’s performance in thousands of micro-level agroclimatic and soil conditions, helping breeders design and create products more targeted for growers’ needs.
• Discovery of new crop protection solutions: We use AI to effectively screen our leading library of 2.6 million compounds, narrowing down the molecules that meet our leading safety and sustainability standards and show potential efficacy. This accelerates the process of identifying the next generation of modulators and new modes of action. Today, Bayer analyzes billions of molecules in a virtual environment to help us design and create only those with the highest potential for success.
• Insights and Actionable Recommendations: Climate FieldView is our flagship digital farming platform, and it’s being used in more than 20 countries on over 220 million acres. Our farmer customers use FieldView to collect data from different farm equipment types and brands, across multiple in-field activities like planting, crop protection application, and harvest. It brings them instant cause-and-effect insights that leads to more confident decisions. Through this platform we’re also able to aggregate farm data and combine it with with public data sets like weather, sensory data like satellite imagery, as well as Bayer’s proprietary research data and use artificial intelligence to deliver recommendations to farmers that inform those planting, spraying, and harvesting decisions to help increase their productivity and profitability, and better protect the environment.

3. Let’s focus on your aerial imagery platform (Imagery Pipeline). What do you use deep learning for? 

Our aerial imagery pipeline uses everything in the data science toolkit – a strategic blend of traditional computer vision methods and machine and deep learning models. For example, we’ve developed a deep learning model that can segment the target crop such as soybeans or corn from soil and weeds. This model was fed with millions of images that were not hand labeled at all. Instead, we trained it using images labeled with a traditional computer vision technique, and then used a unique architecture and active learning approach to train a model way better than the traditional one.  

We also use deep learning models to measure things like flowering, maturity, emergence, and disease. There are dozens of models both in production and under development across our row crop, vegetable, and crop protection pipelines. An exciting area of development for us is how we can leverage existing models to bootstrap and quickly train new ones – for example a model trained to count corn plants can help accelerate development of a model to count other crop types.

4. Developing better products for farmers requires a significant amount of data about in-field performance. AI is fed with petabytes of data being collected every year. How do you produce such big data? 

For generations, scientists walked the fields during trials, visually observing crops to make decisions. Today, we leverage digital tools and in-field expertise to test and develop our seeds/traits and crop protection products for farmers. During trials, we collect valuable crop performance data using our Precision Field Testing Platform, a global network of sensors and devices that capture growth conditions, while a high-throughput digital approach helps our teams analyze results from more than 8,000 internal field-testing locations annually. We plant millions of seeds at precise, GPS-guided coordinates and use high-resolution imagery from drones (or UAVs) to capture 100+million plot images containing critical plant data. We also use UAVs to make spray applications during crop protection testing and capture phenotyping data in both field and tree crops. 

Our UAV Platform at Bayer is creating value at scale across continents, crops, and organizations. Using multiple sensor modalities, ultra-high resolution and scalable software, UAVs have generated more than 200 million phenotyping datapoints, over 2 petabytes of imagery and have taken flight more than 20,000 times. About 70% of all in-field images collected in breeding come from a UAV and in 2020, we leveraged our automated imagery platform to capture data from our crop protection pipeline for the first time–transforming how we learn about plants and products in the field. Check out this video to learn more. 

5. How do you ensure the quality of this data? 

We spent a lot of time building scalability and quality into our imagery processing and imagery analytics engines. To deliver useful data to R&D, we need high-quality, high-resolution imagery as well as the ability to co-register every imagery pixel and our trial data at centimeter precision. To accomplish this, we put a lot of work into how we selected and standardized our field technology in-field processes. We build customized, highly integrated software for mission planning, flight control, and data packaging to help make this easy for our operators. We’ve built models into our automation that improve both efficiency and quality, such as automated detection of ground control points, identification of imagery issues like blur and distortion, and diagnosing misalignment of our trial plots with the imagery. Many of our phenotyping models have their own built-in QC methods that flag datasets for follow up and keep potential bad data from flowing downstream. And last but certainly not least, we have a dedicated team of imagery specialists that act on any alerts from our automated QC systems, respond to user feedback, and perform a white glove QC on many of our most important internal imagery missions.  

6. Is the data quality affected by meteorological conditions? And if yes, what do you about it? 

Yes, it certainly is. Light is one of the biggest issues, as we of course cannot control the lighting conditions in the field the way we can in the greenhouse or lab. This is particularly challenging to generating useful data from multi and hyperspectral sensors. We install calibration panels in the field, and feed the reflectance data from these to methods that adjust the actual imagery data on the trial, so we attain reasonable consistency across fields and time. Variability in lighting also challenges the exposure in images – you can easily get images that are way too bright or dark. We use a combination of ambient light sensors and a set of simple rules for our operators to modify their settings depending on field conditions. The sun angle can cause shadows that wreak havoc with your models if you aren’t careful. We try to fly closer to midday to limit this. And beyond light, things like wind can be a challenge, with its impact on getting stable, non-blurry imagery, especially with smaller UAVs.  The ideal day for us is still and cloudy. However, our field teams are very busy and often travel several hours to get to trial locations, so we must be able to deal with imagery they can get us, even if it is from suboptimal conditions.

7. Historically, collecting this data was very labor-intensive, which limited the insights that scientists could provide. How did farmers reacted to this new way of doing agriculture? Were they all happy about it?  Or did they have some worries and fears as well? 

We are collecting data in more new ways than ever before in our R&D, commercial trialing, and manufacturing operations – this enables us to make better product recommendations and create new business models and value for farmers. In some cases – such as Climate Fieldview enabling a central place for farmers to collect their planting, yield, and other data – we are also enabling farmers to benefit from the advances in science and technology. But the biggest way this technology is making a difference today is in how it enables us to create and recommend better products tailored to farmers specific needs on every field. And when those products show real results, that is what makes farmers happy.

8. How do you communicate with farmers about the use of new technology? Are they involved in the development of technologies at Bayer Crop Science?  

At Bayer, farmers are at the center of everything we do. So, communication is a two-way street, and we do a lot of listening to make sure solutions are tailored to their emerging needs, and then create access, knowledge sharing and tools that meet those needs. Here are a few examples of how we’re doing this: 

• Collaboration in the field: We have key partnerships with farmers around the world for contracting and growing our R&D and commercial trials and seed manufacturing operations – from seed to software development. 
• Better Life Farming: Through the Better Life Farming Alliance, we support smallholders with education and ambassador programs, digital in-field services like seed planting, precision irrigation and crop protection advice, as well as business services like banking, financing, and insurance solutions.
• Global Farmer Survey: Earlier this year, Bayer commissioned the 2023 Farmer Voice Survey – reaching 800 farmers in 8 countries – to talk to farmers on the ground about what they need to succeed today and in the future. In this survey (full report here), they told us about their hopes, challenges, and frontiers, what they want the world to understand about farming during this critical time, and how Bayer can use its position as a leader in agriculture to provide solutions to feed the growing world sustainably, regenerate vital resources and help future generations succeed.   
• Global Farmer Network: We also have a strategic partnership with the Global Farmer Network, which amplifies the farmers’ voice in promoting trade, technology, sustainable farming, economic growth, and food security.

9. Anything else you wish to add? 

AI is a core technology in powering the digital transformation of agriculture, which helps to tackle climate change and ensure food security across the world. We firmly believe that this is a possible mission. We need to re-think agriculture and how we grow and produce our food – and increase yields with less land, less water, less energy, less crop protection, and a lower climate footprint. This can only be achieved with more innovation – in the lab and in the field – and with a willingness to adopt new technologies, new approaches, as well as digital and data-driven solutions.

………………………………………….

Amanda McClerren is the CIO and Head of Digital Transformation & Information Technology at Bayer Crop Science.

She is a member of the Crop Science Executive Leadership and Global Information Technology Leadership teams, and responsible for the digital transformation of all major functions within the Crop Science organization to enable business and digital strategy objectives.

In her last role as Head of Crop Science R&D Digital Transformation & Information Technology, Amanda was responsible for driving the digital road map across multiple technical teams to deliver business outcomes and an industry leading field-testing platform across Crop Science.

Previously, Amanda served as Head of Trait and Pipeline Delivery in R&D, where she delivered the global breeding pipeline and developed the global protected culture strategy to build fully automated and digital greenhouse facilities. She also played critical roles in Biotech and Breeding, developing the Genotyping by Sequencing (GBS) platform to enable Genome wide selection at-scale, developing trait product concepts and strategies to support deregulation.

Amanda has a bachelor’s degree from the University of Illinois at Urbana-Champaign and a Ph.D. in Biochemistry from Duke University. She is actively involved in the St. Louis community as a board member in the Guardian Angel Settlement Association and is also active in the Greater Missouri Leadership Foundation and GROW, Bayer’s Business Resource Group for women’s advancement.

Follow Amanda on LinkedIn: Amanda McClerren | LinkedIn