Paper Presentation & Student Travel Grant — IEEE TALE 2025

Co-authored and presented a peer-reviewed paper at the IEEE International Conference on Teaching, Assessment, and Learning for Engineering (TALE) 2025. The paper, titled "Adaptive Prompt-Based AES: A Teacher-AI Collaborative System for Improving Essay Scoring Accuracy over Time," introduces a novel framework for Automated Essay Scoring (AES) where human assessors and AI collaborate in grading student essays.

The core idea is an adaptive prompting approach — rather than relying on fixed few-shot examples or zero-shot rubric instructions, the system continuously expands its retrieval pool with newly teacher-graded essays. For each ungraded essay, semantically similar previously-graded examples are dynamically retrieved and provided as context to the LLM, allowing the system to improve its scoring alignment over time, much like a human grader refines their rubric interpretation through experience.

Simulations across seven diverse essay tasks from the ASAP 2.0 dataset showed that adaptive prompting progressively improved scoring agreement with human raters as the retrieval database grew. A pilot deployment with real instructors demonstrated a 39% reduction in manual grading workload while maintaining only a 10% discrepancy rate between human and AI scores.

This research directly informed the architecture of the CULI platform — the adaptive retrieval pipeline, the teacher-AI co-evaluation workflow, and the vector similarity infrastructure were all designed and validated through the experiments described in this paper.

Co-authored with Kasidech Aewsrisakul, Norasetha Somboon, Asst. Prof. Dr. Chulaporn Kongkeo (CULI), and Asst. Prof. Dr. Dittaya Wanvarie (Faculty of Science, Chulalongkorn University). Received an IEEE Student Award Travel Grant to present this work at the conference.

Presentation & Award

Presenting at IEEE TALE 2025

Student Travel Grant award

Certificate