The G-CAST system is deployed on e-commerce platforms
The story so far
qwen proposed the G-CAST system, designed to address the challenges faced by e-commerce agents under changing demands and capabilities. The system transforms daily production evidence into reward-guided, entropy-priority training data and governs the evolution of rewards, enabling the agent to learn new behaviors without sacrificing existing capabilities. G-CAST controls model updates under cumulative evaluation and outperforms strong self-play baselines in real e-commerce agent experiments. Currently, G-CAST is deployed on Yuancai, the B2B procurement agent on Alibaba's 1688 platform.
AI-written from the reports · Updated 19 minutes ago
Coverage timeline
Follow the coverage to see every side of the story.
- Hugging Face Daily PapersG-CAST: Governed Continual Self-Training for E-Commerce Agents under Evolving Demands
G-CAST addresses the challenges of e-commerce agents under changing demands and capabilities by transforming daily production evidence into reward-guided, entropy-priority training data and governance reward evolution. The system controls model updates under cumulative evaluation, enabling the agent to learn new behaviors without sacrificing existing capabilities. Real e-commerce agent experiments show that G-CAST outperforms strong self-supervised post-training baselines in new demand adaptation and old behavior retention, and that reward routing, entropy-priority data orchestration, and governance reward evolution are all crucial for performance improvement.
Heat over time
Not enough data yet to draw a trend.