The Diagram is like Guardrails: Structuring GenAI-assisted Hypotheses Exploration with an Interactive Shared Representation
An interactive shared representation that structures GenAI-assisted hypotheses exploration using diagram-based guardrails.
Zijian "Jason" Ding
PhD Candidate
5th-year PhD candidate at University of Maryland, College Park.
8+ years of research in Human-Centered AI, 5+ years in Generative Models.
20+ publications at CHI, C&C, EMNLP, CSCW, UIST, IUI.
Can AI Prompt Humans? Multimodal Agents Prompt Players' Game Design Ideas
Multimodal agents that prompt players to generate game design ideas, exploring whether AI can effectively prompt humans.
Structuring GenAI-assisted Hypotheses Exploration with an Interactive Shared Representation
Formulating or Fixating: Effects of Examples on Problem Solving Vary as a Function of Example Presentation Interface Design
An interaction-oriented theoretical framework for understanding creative inspiration from examples.
joel-chan
How Diverse Initial Samples Help and Hurt Bayesian Optimizers
eesh-kamrah
Harnessing the Power of LLMs: Evaluating Human-AI Interaction through the Lens of News Headline Generation
Evaluating human-AI interaction with LLMs through the task of news headline generation.
Fluid Transformers and Creative Analogies: Exploring Large Language Models' Capacity for Augmenting Cross-Domain Analogical Creativity
Exploring LLMs' capacity for augmenting cross-domain analogical creativity. Best Paper Honorable Mention.
Mapping the Design Space of Interactions in Human-AI Text Co-creation Tasks