LoRA tuning improves control systems course answers in open language models
LoRA Fine-Tuned Models for Control Systems Course Q\&A: A Multidimensional Evaluation of Model Scale and Rank Effects
Artificial Intelligence
Summary
Large language models (LLMs) sometimes struggle to give clear and well-structured answers for control systems coursework. The authors created a dataset of questions and ideal answers from a linear control systems class and fine-tuned two LLMs using a technique called LoRA. This tuning helped the models produce answers that more closely matched course solutions, both in wording and organization. However, the measures used mostly check text similarity and formatting, not whether the math or reasoning is correct, which still needs expert checking.
What this means in practice
- •For educational content developers: Improve automated answer generation to match course-specific terminology and formatting in control systems training materials.
- •For technical support teams: Enhance chatbot responses for technical questions by fine-tuning open models on specialized domain Q&A datasets.
Tested on one dataset.
Authors
Shaowen Lu, Chengxu Liu, Ping Zhou, Tao Yang
Abstract
Large language models (LLMs) are increasingly used in specialized university courses, but control-systems questions require coordinated terminology, notation, derivations, and stepwise explanations. Direct general-purpose responses may be inconsistently structured and hard to verify. Using exercises and reference solutions from a Linear Control Systems course, we built a supervised fine-tuning dataset of 360 system-user-assistant conversations. We applied LoRA to Qwen2.5-3B-Instruct and Qwen2.5-7B-Instruct. With identical data splits, inference settings, and evaluation protocols, we compared base and fine-tuned models and tested LoRA ranks r=4, 8, and 16. Evaluation used ROUGE, BERTScore, and structured-output features to measure reference-answer similarity and stability of the Solution-Method-Teaching Points format. LoRA improved both similarity and structured-output stability at both sizes. On the current test set, 7B-r16 achieved the highest ROUGE-L (0.4093) and BERTScore-F1 (0.8643), while r=8 offered a better balance between performance and parameter efficiency. Bootstrap resampling showed ROUGE-L gains of 0.0764 [0.0613, 0.0915] for 3B-r16 and 0.0874 [0.0687, 0.1042] for 7B-r16; both intervals exceeded zero, indicating stable textual-similarity improvements on the current test set. These results suggest LoRA can align open-source instruction-tuned models more closely with the language and pedagogical organization of course reference answers. However, the metrics mainly capture textual similarity and formatting consistency, not domain-specific reasoning or mathematical correctness, which require expert assessment and task-specific rubrics.