Date
Publisher
arXiv
We propose AnveshanaAI, an application-based learning platform for artificial
intelligence. With AnveshanaAI, learners are presented with a personalized
dashboard featuring streaks, levels, badges, and structured navigation across
domains such as data science, machine learning, deep learning, transformers,
generative AI, large language models, and multimodal AI, with scope to include
more in the future. The platform incorporates gamified tracking with points and
achievements to enhance engagement and learning, while switching between
Playground, Challenges, Simulator, Dashboard, and Community supports
exploration and collaboration. Unlike static question repositories used in
existing platforms, AnveshanaAI ensures balanced learning progression through a
dataset grounded in Bloom's taxonomy, with semantic similarity checks and
explainable AI techniques improving transparency and reliability. Adaptive,
automated, and domain-aware assessment methods are also employed. Experiments
demonstrate broad dataset coverage, stable fine-tuning with reduced perplexity,
and measurable gains in learner engagement. Together, these features illustrate
how AnveshanaAI integrates adaptivity, gamification, interactivity, and
explainability to support next-generation AI education.
Study at a glance
- Question: How can an integrated platform generate diverse AI challenges, support adaptive practice, and sustain learner engagement?
- Setting and data: A curated dataset of 10,845 AI/ML question-answer pairs spanning 26 subject categories across beginner to expert difficulty levels.
- Method: System development and prototype evaluation, including fine-tuning an open-source language model, dataset statistical analysis, and explainability assessment.
Why it matters
AI and machine learning education requires conceptual reasoning, experimentation, and dynamic interpretation, yet existing learning platforms rely on static repositories and lack integrated systems for automated question generation, adaptive assessment, and long-term learner engagement.
Key findings
- Fine-tuning Mistral-7B with 4-bit quantization achieved stable convergence, a low validation perplexity of 2.04, and a BERTScore F1 of 0.427.
- The 10,845-entry dataset achieved 97% cognitive completeness across Bloom's taxonomy levels, with question-answer semantic similarities clustering between 0.6 and 0.8.
- Explainable AI token attribution revealed the model consistently prioritized pedagogically meaningful tokens such as explanation, research-quality, and Bloom.
Summary drafted with AI from the full paper and may contain errors. See the full paper for details.
What is the application?
Who age?
Why use AI?
Study design
