在全球品牌竞争从"流量争夺"转向"情感连接"的今天，如何让海外消费者真正认同品牌，成为中国企业出海面临的核心挑战。 12月26日，在2025第三届全国人工智能应用场景创新挑战赛(CICAS)姑苏专项晋级赛中，明略科技(2718.HK)联合北京大学的参赛项目《基于多模态大模 型的品牌出海创意生成与情感链接智能平台》从70余个参赛团队中脱颖而出，斩获"特等奖"殊荣，成功晋级全国总决赛。 全国人工智能应用场景创新挑战赛(CICAS)是在科学技术部战略规划司指导支持下，由中国人工智能学会与科技部新一代人工智能发展研究中心联合主办 的综合性年度赛事。自2023年办赛以来，共吸引4800余个国内外优秀科技人才团队和优质创新创业项目参与，目前已成为推动AI与实体经济深度融合的 重要平台。 本届大赛以"场景驱动·数智强国"为主题，由中国人工智能学会、苏州市姑苏区人民政府、苏州大学共同主办，全国人工智能应用场景创新挑战赛组委 会、姑苏区经济和科技局联合承办，长三角数字经济双创中心提供支持，共设立49个场景应用专题，明略科技联合北京大学的获奖方案正是"AI+营销"场 景的一次创新突破。 品牌出海：情感共鸣成为新赛点 随着中国品牌 ...

Core Viewpoint - The article highlights the journey of Wu Minghui, the founder of Minglue Technology, emphasizing his technical background, entrepreneurial challenges, and the evolution of his company towards AI-driven solutions, particularly focusing on trust and data credibility in business decision-making. Group 1: Entrepreneurial Journey - Wu Minghui is portrayed as a typical "scholar-type" entrepreneur with a strong technical background, having excelled in mathematics and computer science [1][7] - The company experienced significant ups and downs, including a dramatic downturn where it struggled to pay severance to employees, leading to negative public perception [1][39][46] - After nearly two decades of exploration in the business world, Wu has focused on the core question of what constitutes trustworthy data [3][24] Group 2: Product Development and Innovation - Minglue Technology recently launched the multi-modal foundational model web GUI intelligent agent, Mano, which achieved state-of-the-art performance in international benchmarks [1][2] - The proprietary large model product line, DeepMiner, aims to address the challenge of making AI agents trustworthy, explainable, and traceable in enterprise decision-making [2][68] - DeepMiner is designed to connect credible data sources, enabling businesses to make informed decisions based on reliable data analysis [68][69] Group 3: Strategic Insights and Reflections - Wu reflects on the importance of trust in data and the need for AI to act as a gatekeeper in business decisions [4][66] - The article discusses the strategic errors made during the company's rapid expansion, emphasizing the need for a controlled strategic pace [50][51] - Wu acknowledges the lessons learned from past failures, particularly the necessity of aligning team goals and maintaining trust within the organization [54][57]

Core Viewpoint - Minglue Technology's proprietary GUI model, Mano, has achieved record-breaking SOTA results in the recognized benchmarks Mind2Web and OSWorld, establishing a new paradigm for GUI intelligent agents through innovations in online reinforcement learning and automatic data collection [1][14][23]. Group 1: Performance Achievements - Mano achieved a success rate of 40.1% in the OSWorld-Verified benchmark, surpassing other models such as qwen and GUI-Owl [10][19]. - In the Mind2Web benchmark, Mano demonstrated superior performance across various metrics, including element accuracy and step success rate, significantly outperforming all other SOTA methods [18][15]. - The model's success rate in OSWorld-Verified reached 41.6±0.7%, marking an approximate 7 percentage point improvement over competitors [21][19]. Group 2: Innovations and Methodology - Mano introduces online reinforcement learning as a novel training paradigm in the GUI interaction field, enhancing its performance in dynamic environments [22][23]. - The model's architecture consists of three main components: exploration module, processing flow, and optimization process, which collectively improve its reasoning and adaptability [25][26]. - The automatic data collection method developed by the technical team significantly enhances the efficiency and accuracy of data acquisition, allowing for the generation of high-quality interaction trajectory data [48][49]. Group 3: Market Context and Future Directions - The demand for AI agents is expected to surge by 2025, positioning Mano as a key player in differentiated competition by accessing data sources that other agents cannot reach [59][63]. - Minglue Technology plans to continue exploring areas such as data collection, training integration, and CAPTCHA handling to further optimize Mano for real-world applications [66].