CLOSED: Invitation to tender: Independent Technical Expert/s - evaluate proposed AI platform to predict classroom conditions in Tanzania
The FT Hub is looking to contract an individual/team to work closely with the FCDO team and World Bank leads from Global Facility for Disaster Reduction and Recovery (GFDRR) to respond to the question:
How many classrooms in Tanzania do we need to collect data in order to develop an AI model that can predict conditions in classrooms across Tanzania to a 95%+ accuracy?
The Invitation to Tender document can be found by clicking this LINK
Classrooms in Sub-Saharan Africa can often be uncomfortable places to be, let alone to teach and learn in. Sweltering temperatures have proven negative physiological impacts. A study showed classroom temperatures in Tanzania already regularly exceed 40℃. Poorly lit, or overly sun exposed, classrooms make it much more difficult to read, and introduces health risks (Ibhadode et al., 2019). Lack or excess of light excludes children with poor eyesight, introducing a layer of inequity. During the rainy season it becomes nearly impossible to teach given the sound of rain on corrugated iron.
Overall, these classroom conditions, all infrastructural in nature, are fundamentally part of the learner experience. Furthermore, the reliability of water supplies fluctuate impacting schooling. Natural disasters can derail the whole process of schooling. All of these challenges are expected to increase in the coming years due to climate change.
FCDO is currently funding work to measure classroom experience in Tanzania and has a separate study looking at what interventions, such as retrofitting, would be relevant for different conditions, and to prioritise these measures based on maximising impact to children..
To support this work, the FCDO would like to predict classroom experience in classrooms across Tanzania. Classroom experience is strongly related to factors such as sound and temperature, as well as access to light, levels of pollution, and disaster risk.
To predict these classroom conditions, an AI could be used to intelligently predict certain conditions. To create model/s, training data needs to be collected. The World Bank is collecting this data in 50 schools. However, we expect we will need more data to build/train an AI model.
An early indication of how big a sample is needed for the different measures is deemed important because not all schools can be surveyed.
The opportunity is to work to understand
What the potential for AI models could be to determine certain conditions, like temperature, light and sound, and
Understand for each metric how much data might need to be collected to produce an AI model with greater than 95% accuracy,
With this information, the team will use the answer to drive the next phase of data collection plans, and focus efforts on which measures show the greatest promise with regards to developing an AI model/s. This could help create a national model that maximises impact of investments made, and improve child education in Tanzania.
For those interested in applying, please note that the application deadline is 14 February 2024 @ 11:59pm UK time.
We look forward to receiving your submissions!
Please contact FTLenquiries@DT-Global.com for further enquiries regarding this Terms of Reference or for further information on the Frontier Technologies Programme.
