Dashboard2Code: Evaluating Multimodal Models on Reconstructing Interactive Dashboards
Learn how to evaluate multimodal models on reconstructing interactive dashboards with Dashboard2Code, a novel task that assesses a model's ability to proactively explore and integrate feedback from interactions.
- Implement a multimodal model that can generate interactive dashboards
- Use Dashboard2Code to evaluate the model's ability to proactively explore and integrate feedback from interactions
- Configure the model to acquire and integrate feedback from its own interactions, such as clicking and filtering
- Test the model's performance on reconstructing interactive dashboards
- Compare the results with existing models and techniques
Data scientists and AI engineers working on multimodal models and interactive data visualization can benefit from this task, as it evaluates a model's ability to generate interactive dashboards that can be used for real-world data exploration.
💡 Multimodal models can be evaluated on their ability to generate interactive dashboards that can be used for real-world data exploration, using tasks like Dashboard2Code.
📊 Introducing Dashboard2Code: a novel task for evaluating multimodal models on reconstructing interactive dashboards! 🚀
Key Takeaways
Learn how to evaluate multimodal models on reconstructing interactive dashboards with Dashboard2Code, a novel task that assesses a model's ability to proactively explore and integrate feedback from interactions.
Full Article
Abstract:
arXiv:2607.04727v1 Announce Type: cross Abstract: Automatic data visualization generation has advanced rapidly with multi-modal large language models, yet existing efforts largely focus on static charts and overlook the interactive dashboards commonly used for real-world data exploration. We introduce Dashboard2Code, a novel task that requires a model to proactively explore an interactive dashboard, acquire and integrate feedback from its own interactions (e.g., clicking and filtering), and gene
DeepCamp AI