Cedar Initiative by NoxusDynamics gives students hands-on mentorship from collecting available data to training a real AI model, deployed against a real problem in the agriculture sector.
Instead of traditional lectures, students gather data and train a real AI model, grasping the core concepts
independently. A dedicated mentor clears doubts and guides the process, but the thinking is up to the
students.
"The mentorship is in the thinking, not the tooling."
Grasp concepts independently. There are no passive lectures instead you learn by researching, experimenting, and building on your own.
Work as a team of 4 to submit one cohesive project report, ensuring every member contributes actively to the presentations.
Defend choices and share progress in weekly team presentations, receiving direct feedback and course correction from your mentor.
Learning AI is only valuable when applied to real-world datasets. Students focus on solving daily agricultural challenges by training AI models that translate raw data into clear, actionable classifications and decission making models.
This month's Cader Initiative focus is on:
Mentor allots teams of 4 with an assigned Lead. Teams set up their workspace and collect available datasets.
Design and train a real AI model on the allotted agricultural problem statement.
Conduct the weekly team presentation to review choices, optimize the model, and clear doubts with the mentor.
Deliver the final team presentation and submit a single, collaborative project report for certification.
Binary classification (Ripened vs. Unripened) using skin color analysis for local orchards.
Predicting 'Mature' vs. 'Tender' status based on weight and color metrics for the coconut oil industry.
Detecting Sigatoka disease in Nendran banana varieties using leaf image classification.
Classifying soil suitability for specific agricultural crops based on visual texture or pH/moisture data.
Determining harvest readiness for jackfruit using external spine density and color.
Identifying 'Blast' infection in rice crops to protect yields and fields.
Binary identification of crops vs. weeds to assist organic farming initiatives.
Analyzing gill and eye color to classify fish as 'Fresh' or 'Stale' for local seafood markets.
Sorting pineapples into 'Market Ready' or 'Unripe' specifically for commercial growers.
Detecting fruit split in nutmeg to signal optimal harvest timing for backyard growers.
Predicting the best crop to plant based on Nitrogen-Phosphorous-Potassium (NPK) soil values.
Detecting Early Blight vs. Healthy leaves using the standard PlantVillage dataset framework.
Predicting 'Heavy' vs. 'Light' rainfall to help farmers plan harvest and planting cycles.
Identifying common rust infections in maize leaves to support the local fodder crop sector.
Monitoring potato health in high-altitude agricultural regions through leaf classification.
If you face or find a real-world agricultural challenge that could be addressed using machine learning, submit your problem statement. Our cohorts will analyze the problem, collect datasets, and train custom AI models to solve it.