LitLinker: Supporting the Ideation of Interdisciplinary Contexts with Large Language Models for Teaching Literature in Elementary Schools

Research Background and Issues

• Identified Problems or Challenges:

  1. Connecting interdisciplinary content (e.g., science, art, etc.) with reading materials in literature teaching has been shown in previous studies to enhance elementary school students' learning outcomes. However, designing suitable interdisciplinary contexts remains highly challenging.
  2. Literature teachers often focus on a single domain, lacking interdisciplinary expertise and the time to comprehensively gather and evaluate information.
  3. Existing tools (e.g., ChatGPT) possess the ability to understand long texts, but the content they generate often requires additional adjustments by teachers and may not align with the cognitive and teaching needs of elementary education.
  4. There is a lack of dedicated interactive tools to assist teachers in designing interdisciplinary teaching contexts.

• Significance: Interdisciplinary education helps cultivate students' critical thinking, creativity, and problem-solving abilities, while also increasing their learning motivation. Developing innovative tools to support teachers in designing interdisciplinary curricula is crucial to addressing modern educational needs.

• Research Motivation and Related Work:

  1. In the academic domain, existing studies have demonstrated that large language models (LLMs) can support activities such as information exploration and curriculum design across different disciplines.
  2. Related research has primarily focused on STEM fields, while interdisciplinary integration in elementary literature teaching remains underexplored, particularly the application of technology in diverse teaching environments.

Solution

• Methods and Solutions:

  1. This paper develops an interactive system called LitLinker, which leverages the analytical capabilities of LLMs to support teachers in designing diverse interdisciplinary contexts for literature courses.
  2. LitLinker analyzes the relationships between texts and themes, recommends interdisciplinary topics, and generates detailed lesson plans and teaching activities.

• Innovations:

  1. Provides a workflow suitable for teachers' "reverse thinking," simulating human team collaboration in designing interdisciplinary contexts.
  2. Combines LLMs' analytical capabilities to support teachers in deeply exploring the connections between reading materials and interdisciplinary themes.
  3. Generates structured teaching plans (including course introductions and recommended classroom activities), reducing teachers' workload.

• Implementation Steps and Key Technologies:

  1. Key Technologies: Utilizes GLM-4 LLMs for retrieval-augmented generation (RAG) to optimize content.
  2. Functional Areas: Includes three main interfaces—"Theme Exploration View," "Text Exploration View," and "Collection View"—to provide users with exploration and management functionalities.
  3. User Interaction Process:
     • Select relevant themes from recommended interdisciplinary topics;
     • Analyze the relationships between different texts and the selected themes, obtaining in-depth analysis through LLMs;
     • Generate a lesson plan and related activity suggestions that can be implemented in actual teaching.

Research Outcomes

• Specific Outcomes:

  1. LitLinker significantly enhances the depth of interdisciplinary theme integration while reducing teachers' workload during the design process.
  2. Experimental evaluations indicate that LitLinker's generated content is more comprehensive and effective, particularly in designing classroom activities.

• Advantages Over Existing Solutions:

  1. LitLinker's high degree of automation and structured output expands teachers' exploration of interdisciplinary contexts while retaining space for user interaction and critical thinking.
  2. Compared to single-model generated conversational responses, LitLinker's modular design and layered information processing are more reasonable and user-centered.

• Experimental or Evaluation Results:

  1. In a "participatory experiment" involving 16 teacher participants, LitLinker was shown to significantly reduce teachers' workload while enhancing task satisfaction.
  2. Expert interviews with 9 classroom teachers highlighted LitLinker's advantages in broadening teachers' perspectives and generating actionable classroom activity suggestions.
  3. However, experts also noted that some interdisciplinary theme recommendations were repetitive, and certain content failed to align closely with curriculum standards.

• Limitations and Future Directions:

  1. The current theme data sources lack diversity. While some "annual buzzwords" enhance engagement, they sometimes do not fully align with educational objectives.
  2. Some of LitLinker's recommended content deviates from core teaching goals, particularly the specific objectives of literature education.
  3. Future work should expand the theme library through diverse datasets, optimize LLMs' generation logic (e.g., incorporating more advanced RAG techniques), and integrate multimodal content (e.g., images or videos) to enhance educational outcomes.

Through the research outcomes presented in this paper, LitLinker demonstrates the significant potential of LLMs in designing interdisciplinary teaching for elementary literature education, offering a new perspective on optimizing educational processes through human-computer collaboration.