TACT: Taxonomy-Aligned Post-Training for Pedagogically Adaptive English Tutoring
2026-08-04 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors created a new system called TACT to improve how AI tutors help people learning English by making sure the tutor picks teaching strategies based on what the learner does and the conversation's context. They designed categories to describe both tutor actions and student behaviors and used these to label a large set of real teacher-student chats. By training an AI model with this labeled data and special optimization, their AI tutor, TACTutor, better supports learning than previous models and other tested tutors. They also tested it with real learners, who rated it highly. The authors provide the data and models openly to help future research in adaptive ESL tutoring.
Large Language ModelsESL tutoringPedagogical adaptationTutor strategy taxonomyStudent move taxonomySupervised fine-tuningPolicy optimizationConversational AIScaffoldingBenchmark evaluation
Authors
Dongjie Yang, Siyan Lin, Leixian Shen, Rui Sheng, Huamin Qu, Zixin Chen
Abstract
Large language models (LLMs) are increasingly used to provide conversational practice for English-as-a-second-language (ESL) learners. Effective ESL tutoring, however, requires more than fluent response generation: a tutor must select an appropriate pedagogical action based on learner behavior and dialogue context. Human-tutoring research offers principles for adaptive support, but they are often task-specific and remain insufficiently integrated into LLM-based ESL tutor training and evaluation. We present TACT (Taxonomy-Aligned Conversational Tutor), a human-grounded framework for post-training and evaluating pedagogically adaptive ESL tutors. Drawing on established literature, we develop two complementary taxonomies: the Tutor-Strategy Taxonomy with 13 tutor response strategies and the Student-Move Taxonomy characterizing learner behavior by move type and status. Using these taxonomies, we construct TACTCorpus, which enriches 260 authentic teacher-student conversations with 32,379 annotations and quality-controlled augmented training data. We then post-train Qwen3.5-4B through supervised fine-tuning followed by taxonomy-aligned Group Relative Policy Optimization, producing TACTutor and optimizing it for scaffolding quality rather than reference imitation alone. On TACTBench, a strategy-balanced diagnostic benchmark comprising 78 authentic tutoring contexts, TACTutor improves over its backbone by 20.30% and outperforms all evaluated proprietary baselines under the same protocol, while maintaining backbone performance on established external educational benchmarks; in a blinded study with 50 learners, it also receives the highest overall mean rating among the evaluated tutors. We release the data, benchmark, and model weights, providing an open foundation for developing pedagogically adaptive ESL tutors.