Summary of Zhejiang Shuju Culture Conference Call Company Overview - **Company**: Zhejiang Shuju Culture - **Period**: First three quarters of 2025 Financial Performance - **Revenue**: 2.152 billion CNY, a slight decrease of 0.79% year-on-year [2][3] - **Net Profit**: 535 million CNY, an increase of 12.65% year-on-year [3] - **Net Profit (Excluding Non-recurring Items)**: 340 million CNY, a growth of 6% year-on-year [3] - **Operating Cash Flow**: Improved from a negative 182 million CNY to a positive 324 million CNY [3] - **R&D Investment**: Increased by 20 million CNY to 268 million CNY, representing 12.5% of revenue [2][3] Business Segments Gaming Sector - **Peak Games**: Maintained steady growth despite no new game licenses; strong performance from popular games like "Doudizhu" [2][4] - **Innovation Team**: Established to explore mobile games and other categories [4] - **Profit Growth**: Driven by enhanced traffic effects from premium games [6] Digital Marketing - **Contribution to Profit**: Digital marketing segments like Jiutian Interactive and Taotian Media contributed to net profit growth [6] IP Economy - **Focus Area**: IP economy is a key growth area with multiple products launched in September and October, including toys and collectibles [7] - **Future Plans**: More products expected in Q4 and next year, with resource integration from previous investments anticipated to positively impact financials [7] Digital Technology (AIDC) - **Stability**: The AIDC segment remains stable, with plans for collaboration with leading computing card manufacturers in Beijing [8] - **Growth Potential**: The Dajiangdong area shows potential for growth, supported by Alibaba's significant investment in AI [8] Strategic Partnerships - **Collaboration with Alibaba Cloud**: A strategic framework agreement signed to upgrade computing power at the Fuyang base and deepen cooperation on the Dajiangdong project [9] - **AI Infrastructure**: Plans to enhance AI industry infrastructure through collaboration [9] AI Applications - **AI Development**: Formation of AI models and algorithms aimed at various sectors including smart cities and digital media [10] - **Commercialization**: Existing R&D outcomes are being transformed into competitive products for commercial value [10] Data Trading Center - **Growth in Trading Volume**: Expected to double in 2025, surpassing 100 million CNY [11] - **Future Potential**: Current trading volume is only 2%, with significant growth potential as national policies promote data trading [12] Investment Activities - **IPO Projects**: Investments in companies like Haima Cloud and Tongshifu are expected to yield returns upon their IPOs [13] - **Stock Holdings**: Company holds shares in Huatuo, with partial reductions noted; further details pending in the upcoming quarterly report [13] Collaboration with Alibaba - **Equity and Business Cooperation**: Includes joint ventures and strategic projects aimed at leveraging shared resources for mutual growth [14] - **Investment in Media**: Significant investment in media ventures to enhance collaborative business models [14] This summary encapsulates the key points from the conference call, highlighting the financial performance, business segments, strategic partnerships, and future outlook for Zhejiang Shuju Culture.

Core Viewpoint - The research indicates that large models can effectively play various card games, demonstrating their capabilities in complex decision-making scenarios [2][4][52]. Group 1: Model Performance - Different models exhibit varying performance across different card games, with fine-tuned models showing superior results compared to API-based and base models [3][40]. - Among the API-based models, GPT-4o performs the best overall, while GLM-4 demonstrates strong capabilities in games like DouDizhu and GuanDan [39][40]. - Fine-tuned models, particularly GLM4-9B-Chat-mix, excel in multiple games, including DouDizhu, GuanDan, and Uno, indicating their versatility [42][40]. Group 2: Game Selection and Learning Methodology - The research team selected eight popular card games based on their complexity and the availability of high-quality models and data [8]. - The learning process involved generating high-quality interaction data through teacher models and opponents, allowing the large language models to learn effectively [14][16]. - The complexity of the games influenced the number of training instances collected, with more complex games like DouDizhu and GuanDan requiring larger datasets [20][21]. Group 3: Inter-Game Influence - The study found that models trained on similar games can enhance each other's performance, while those trained on games with significant rule differences may experience performance conflicts [52][49]. - For instance, models trained on GuanDan showed good performance in DouDizhu, suggesting a positive transfer of skills between these games [45]. Group 4: Generalization and Capability - The research indicates that while training on card games, the general capabilities of the models may decline, but this can be mitigated by incorporating general data into the training process [56][54]. - The mixed training approach allowed for some recovery of general capabilities, demonstrating the balance between specialized game skills and broader knowledge [56].