美国国家工程院外籍院士、知名科技专家沈向洋在最新发言中指出，我们正步入一个由通用大模型主导的新时代，这些模型将广泛渗透至各个垂直行业。他表示，为了实现通用模型的高性能，模型的参数规模将达到前所未有的万卡乃至上万亿级别，这在科技领域引发了深刻的思考和变革。
沈向洋院士强调，未来最具潜力的研究和发展方向可能是个性化个人大模型。他认为，将每个人的个性化需求与大模型相结合，再辅以云计算和端设备的支持，将创造出具有重大价值的全新应用场景。这一构想预示着人机交互将更加智能化，同时为用户提供更为精准和个性化的服务。
沈向洋的观点揭示了人工智能技术发展的新趋势，即从通用性向个性化过渡，这不仅将改变科技行业的格局，也将对社会生活产生深远影响。随着大模型技术的不断进步，人机关系的重构将为各行各业带来前所未有的机遇和挑战。来源：财联社。
英语如下：
**News Title:** “Academician Sheng Xiangyang Envisions: The Era of Universal Large Models Will Bring Disruptive Changes to Human-Machine Relationships”
**Keywords:** Sheng Xiangyang, Large Models, Human-Machine Relationship
**News Content:**
Title: Academician Sheng Xiangyang Foresees the Era of Universal Large Models: Redefining Human-Machine Interactions, Personalization Takes Center Stage
Renowned technology expert and Foreign Member of the US National Academy of Engineering, Sheng Xiangyang, recently stated that we are entering a new era dominated by universal large models that will pervade various vertical industries. He emphasized that achieving high performance in these models would require parameter scales reaching an unprecedented trillion-level, sparking deep reflection and transformation in the tech sector.
Academician Sheng highlighted the potential of personalized individual large models as a key research and development direction for the future. He believes that combining each person’s unique needs with large models, supported by cloud computing and edge devices, will create new, highly valuable application scenarios. This vision suggests a more intelligent human-machine interaction and offers more precise, personalized services to users.
Sheng’s perspective reveals a new trend in AI technology development, transitioning from generality to personalization, which is set to reshape the tech landscape and have profound impacts on society. As large model technology advances, the reconfiguration of human-machine relationships will present unprecedented opportunities and challenges across various industries. Source: Sina Finance.
【来源】https://www.cls.cn/detail/1628008
Views: 1