Literature Review Workflow Enhancements

[movinam]

Feature Type

Enhancement to existing feature

Problem Statement

The pipeline’s current literature review agent implementation using STORM requires several improvements. This thread is to keep track of what changes can be made to this workflow to compare with other existing article generation methods such as Co-STORM and WikiAutoGen.

Proposed Solution

• Independent Nodes: Detach the nodes that are part of the sequential workflow and make them reusable (for example, the research interviews, the outline generation, etc.)
• Integrate Multi-Modal Generation: Enable the agent to incorporate information and generate formats such as text, images, tables, and graphs.
• Format Outputs: Add functionality to format the output to a user-defined outline instead of one generated by the LLM.
• HITL: Add human in the loop to allow the user to interrupt the workflow.
• Enable Incremental Updates: Allow the agent to update or expand existing literature reviews based on new input or discoveries, rather than regenerating the entire document.

Alternative Solutions

No response

User Benefits

No response

Implementation Ideas

No response

Contribution

• I'm willing to submit a PR for this feature
• I'm willing to test this feature
• I'm willing to help document this feature

Additional Context

No response