You are viewing the 2025 archive. Submissions for this year are closed.Go to 2026 →

Mini Challenge — 2025

The 2025 mini challenge encouraged the exploration of how agentic AI can automate data visualization and visual analytics. We provided the dataset to visualize (VisPubData), alongside a Mini Challenge Template capable of inputting and analyzing dataset, and producing concise data visualization reports. The template was to help participants get started quickly.

Evaluation

After the server closed, we applied a review process similar to that of paper submissions. Evaluation criteria included:
  • Agent-generated report — clarity, coherence, and insightfulness;
  • Technical report — explanation of key decisions, challenges faced, and lessons learned.
    • Accepted submissions were invited to present their works.

Submission

The template as baseline is available at Mini Challenge Template with README to explain the file structure and how to use it. Participants could submit multiple times, up to 10 times per day, and each submission was at most 1M tokens in total. By the submission deadline, participants submitted a technical report (VGTC short paper format, up to 2 pages via the PCS system) that (1) describes how their best-performing agent was developed and (2) includes a link to its best generated report.
PCS TrackPCS Category
GuideHow to Start with Template & Submit for Challenge?

Leaderboards

In this competition, we offered both public and private leaderboards:
  • Public leaderboard: This determined the mini-challenge awards. All the Finalized Submissions were shown for reviewers to evaluate.
  • Private leaderboard: Displayed all previous submissions. Participants could review them and mark only one as their finalized submission.

FAQ

  • Did I need an LLM's API key?
    Participants needed an API key to test their agent locally. However, it was not required for submissions — the evaluation server handled all LLM API calls. Multiple LLMs were supported, including GPT-4o, O1, and O4-mini.
  • Did the mini-challenge count as a publication?
    Yes, awarded submissions were published in the same format as the short paper submissions.
  • Were frameworks other than LangGraph supported?
    Yes. As long as the implementation followed the template instructions and the interface in agent.py with the class agent, producing an output in the root directory during evaluation, participants were free to use other frameworks.

Timeline

  • Submission Site Opens: June 1st, 2025
  • Challenge closes: Aug 29th (Friday), 2025
  • Author Notification: Sep 10th, 2025
  • Camera-Ready Deadline: Oct 1st, 2025
  • Workshop Day: Nov 3rd, 2025

Awards

Challenge finalists teams received a share of $6,000 in AWS cloud credits:

  • $3K for the Top1 Winner
  • $1.5K for the Runner-Up
  • $500 for 3 Honorable Mentions

AWS