IdeaSynth: Iterative Research Idea Development Through Evolving and Composing Idea Facets with Literature-Grounded Feedback

Kevin Pu·K. Feng·Tovi Grossman...Joseph Chee Chang...
CHI·2025·41 citations

TLDRIt is demonstrated that participants effectively used IdeaSynth for real-world research projects at various ideation stages from developing initial ideas to revising framings of mature manuscripts, highlighting the possibilities to adopt IdeaSynth in researcher’s workflows.

How do people cite this paper?

(generated 20 days ago)

IdeaSynth's canvas-based, node-link interface for decomposing research ideas into modular facets has informed the design of subsequent AI-assisted ideation tools — with systems adopting similar spatial ideation boards, tree-structured exploration views, and node-link diagrams for evolving research intents — while its approach to iterative, literature-grounded idea refinement through facet decomposition and recombination has been situated as a representative method in systematic reviews of LLM-assisted ideation, referenced as a model for human-LLM collaborative research workflows, and used to motivate or contrast with new systems for agentic ideation pipelines, structured idea generation, and end-to-end proposal writing, with specific elements such as its evaluation criteria (novelty, feasibility, impact) and identified user needs being directly adopted by follow-on tools.

Loading PDF reader...

IdeaSynth: Iterative Research Idea Development Through Evolving and Composing Idea Facets with Literature-Grounded Feedback

Kevin Pu·K. Feng·Tovi Grossman...Joseph Chee Chang...
CHI·2025·41 citations

TLDRIt is demonstrated that participants effectively used IdeaSynth for real-world research projects at various ideation stages from developing initial ideas to revising framings of mature manuscripts, highlighting the possibilities to adopt IdeaSynth in researcher’s workflows.

How do people cite this paper?

(generated 20 days ago)

IdeaSynth's canvas-based, node-link interface for decomposing research ideas into modular facets has informed the design of subsequent AI-assisted ideation tools — with systems adopting similar spatial ideation boards, tree-structured exploration views, and node-link diagrams for evolving research intents — while its approach to iterative, literature-grounded idea refinement through facet decomposition and recombination has been situated as a representative method in systematic reviews of LLM-assisted ideation, referenced as a model for human-LLM collaborative research workflows, and used to motivate or contrast with new systems for agentic ideation pipelines, structured idea generation, and end-to-end proposal writing, with specific elements such as its evaluation criteria (novelty, feasibility, impact) and identified user needs being directly adopted by follow-on tools.

Paper

Loading PDF reader...