ExoQuest: Finding the Next Earth via Gamified Crowdsourcing

ExoQuest is an open-source project that combines citizen science and machine learning. It enables ordinary users to participate in exoplanet signal screening through a gamified interface, building a human-machine collaborative system for discovering habitable planets. This project aims to address the insufficient capacity of professional astronomers to process massive astronomical data (such as data from TESS and Gaia DR3 missions), using human intuition to judge ambiguous signals and improving model performance through an active learning loop.

Earth faces existential threats such as climate change and resource depletion, making the search for habitable exoplanets an urgent task in the astronomical community. Traditional detection relies on professional teams, but the data volume from TESS and Gaia DR3 exceeds the processing capacity of professionals; exoplanet signals have ambiguous zones (false positives may come from stellar activity, noise, etc.), and these boundary cases require human intuition to judge. ExoQuest attempts to convert citizen scientists' intuition into high-quality machine learning training data through gamification.

ExoQuest consists of three core components to form an ecosystem:

1. ExoQuest Scientific Pipeline: Includes Scout (screening K-type dwarf targets), Pulse (Wotan algorithm for detrending and noise reduction), and QuestX (TransitLeastSquares algorithm for transit search);
2. XQuest Gamified Interface: Includes Transit Toss (sliding annotation to judge signals), Mission HUD (narrative-driven character growth), and Leaderboard (global discovery ranking);
3. ExoReg Registry: Maintains a searchable database to archive all processed stellar data, ensuring scientific reproducibility.

The core innovation of ExoQuest is the active learning loop:

1. Screen ambiguous signals with medium machine confidence;
2. Users annotate through the game interface;
3. Retrain the Main Learning Model (MLM) after accumulating annotations;
4. Focus on more difficult boundary cases after model improvement. This loop forms positive feedback: more users → more training data → better model → processing larger datasets.

To maintain long-term user participation, ExoQuest adopts:

• Narrative packaging: Users play the role of "Galactic Architects" to give meaning to their work;
• Progress visualization: An 8-module growth path to unlock functions and storylines;
• Waiting optimization: Educational icons to display stellar knowledge;
• Timeout handling: A 7-second delay triggers a notification, allowing users to choose to continue or return later.

ExoQuest is an open-source project that collaborates via GitHub and welcomes contributors from all backgrounds, such as astronomers and developers. Open source ensures:

• Transparency: Auditable data processes to verify conclusions;
• Sustainability: The community can take over maintenance;
• Educational value: Students can learn astronomical data processing and machine learning applications through the source code.

Current limitations: The repository mainly contains XQuest front-end code, and the complete back-end pipeline may not be fully developed. Key success factors: Attracting enough active users, improving MLM accuracy, and addressing data quality issues. Future potential: Shifting from expert dependence to collective intelligence, real-time human-machine collaboration, and a paradigm shift in open crowdsourced astronomical research.