Spatial Reasoning Task- Mathematics
As part of the IMAGE-QA campaign to train an AI model, I designed and hand-crafted mathematically rigorous problem sets that guide the AI from simple 2-D translations to intricate 3-D transformations, topology, and path planning. Each problem is presented with clear instructions, sequenced to increase difficulty without cognitive jumps, and audited to eliminate hidden shortcuts or biases that might let the model cheat. Together, these tasks form a benchmark-quality suite that accelerates deep search while ensuring the AI’s spatial intuition is comprehensive, robust, and ready for practical deployment.
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi
@copyright Shubhangi Dwivedi