EduFlow: Advancing MLLMs' Problem-Solving Proficiency through Multi-Stage, Multi-Perspective Critique

2025-07-12

AI Summary

Multimodal large language models (MLLMs) exhibit limited performance on multi-step, interpretable scientific reasoning tasks due to the absence of domain-specific scientific reasoning patterns, lack of global coherence in reasoning paths, and no reflective self-correction capability. To address these challenges, we propose EduFlow—the first end-to-end scientific reasoning framework tailored for educational settings. Our method introduces (1) EduPRM, a process-aware reward model enabling fine-grained assessment of reasoning quality, and (2) EduMCTS, a domain-adapted Monte Carlo Tree Search mechanism supporting multi-stage critical feedback and reflective self-correction. Leveraging MCTS-guided trajectory generation, curriculum learning, controlled error injection, teacher-student dialogue supervision, and self-consistency optimization, EduFlow significantly improves both reasoning accuracy and consistency. We release EduMCTS-160K, a large-scale dataset of scientific reasoning trajectories, alongside open-source code, data, and models.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Multiagent Systems: Multiagent Learning

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work

Abstract

Multimodal large language models (MLLMs) still perform poorly on scientific tasks, particularly those requiring multi-step and interpretable reasoning. Their limitations include insufficient scientific reasoning patterns, lack of global coherence in multi-step inference, and the absence of reflective self-correction, making them unreliable in structured scientific contexts. We introduce EduFlow, the first end-to-end framework that covers the full pipeline of educational scientific reasoning, including data selection, MCTS-based trajectory construction, model training, and output optimization. At its core is EduPRM, a process-aware reward model that critiques reasoning steps with tags and justifications. EduPRM is trained via curriculum learning on three complementary supervision sources: MCTS-guided trajectories, error-injected critiques, and teacher-student dialogues, enabling dynamic adaptation to multi-stage problem solving and iterative refinement during inference. We further propose EduMCTS, a domain-adapted search framework that introduces bootstrapping actions specifically designed for educational reasoning, such as a self-reflection mechanism that promotes reflective error correction. It further leverages EduPRM's fine-grained feedback to guide the search toward higher-quality reasoning trajectories. By applying self-consistency and rejection sampling, we constructed EduMCTS-160K, a large-scale dataset of educational reasoning trajectories. Extensive experiments demonstrate that EduFlow enhances reasoning consistency and coherence. Code, data, and models will be released.

Problem

• Improving MLLMs' scientific reasoning with multi-step critique
• Enhancing global coherence in multi-step scientific inference
• Introducing reflective self-correction for structured scientific tasks

Innovation

• EduFlow: end-to-end educational reasoning framework
• EduPRM: process-aware reward model with curriculum learning
• EduMCTS: domain-adapted search with self-reflection mechanism