作者：百年不渡分类：未来幻想点击：287次下载：1519431次大小：70M 日期：2026-08-26

智联招聘

COP17 "China Corner" showcases China's achievements in combating desertification_我的网站

一 | 中新网上海7月22日电 (记者郑莹莹)在刚结束不久的2026世界人工智能大会暨人工智能全球治理高级别会议上，AI4S(Artificial Intelligence for Science，人工智能驱动的科学研究)领域备受关注，“科学数据的未来”引发行业人士深思。

People visit the exhibition area of the "China Corner" in the Blue Zone of the 17th Session of the Conference of the Parties (COP17) to the United Nations Convention to Combat Desertification (UNCCD) in Ulan Bator, Mongolia, Aug. 17, 2026. The "China Corner" themed exhibition made its debut on Monday, showcasing China's experience and achievements in combating desertification. (Xinhua/Bolortsetseg)The "China Corner" themed exhibition made its debut in the Blue Zone of the 17th Session of the Conference of the Parties (COP17) to the United Nations Convention to Combat Desertification (UNCCD) in Ulan Bator, capital of Mongolia, on Monday, showcasing China's experience and achievements in combating desertification.
Held under the overall theme of COP17, "Restoring Land, Restoring Hope," the exhibition presents China's efforts and achievements in combating desertification through five chapters across three dimensions.
Chapter 1: The Great Green Wall - Strengthening the Ecological Security Barrier in the North
Since 1978, China has launched a series of major ecological projects, including the Three-North Shelterbelt Forest Program (TSFP), widely recognized as the world's largest afforestation program. The TSFP was launched in 1978 and is scheduled to run until 2050. It is being implemented in eight phases, with the program currently in its sixth phase.
Since the implementation of the TSFP began, forest and grassland coverage in the project area has reached 40.76 percent, while the treatment rate of desertified land has reached 67.82 percent.
Chapter 2: Prairie Conservation and Restoration - Bringing Grasslands Back to Life
Since the 14th Five-Year Plan period, China has rehabilitated and restored nearly 15.33 million hectares of degraded grassland.
In addition, nearly 80.66 million hectares of grassland have been restored through measures such as grazing bans and enclosure, benefiting approximately 12 million farmers and herders.
Chapter 3: International Cooperation - Building a Green Community with a Shared Future
Upholding the concept of a community of life on Earth, China has established desertification cooperation platforms with Arab countries, Mongolia, and Central Asian countries. It has also supported Belt and Road countries in their efforts to combat desertification, sharing Chinese experience and solutions for global ecological governance.
Chapter 4: Scientific Prevention and Control - Combining Knowledge and Technology to Enhance Governance Efficiency
China adheres to an integrated approach to the protection and systematic management of mountains, rivers, forests, farmlands, lakes, grasslands, and deserts. It adapts measures to local conditions, prioritizes water-efficient approaches to ecological restoration, and implements targeted treatments in different areas.
In terms of technological advancement, the utilization rate of high-quality seeds has exceeded 75 percent, while the mechanization rate of sand control has reached 50 percent. The application of new equipment has increased the efficiency of sand-control operations by more than threefold.
Chapter 5: Coordinated Development of Ecological Industries - Achieving Green Prosperity Together
China has developed ecological industries and ecotourism initiatives tailored to local conditions, promoting the integrated development of ecological protection, industrial growth, and improvements in people's livelihoods. These efforts have helped transform desert and sandy areas into "green homes," enabling local communities to increase their incomes and achieve greater prosperity.
China has seen a continuous decline in the areas affected by desertification and land salinization, becoming the first country in the world to achieve "zero growth" in land degradation. Approximately 53 percent of manageable salinized land has been effectively treated, making China a global leader in ecological restoration.
During COP17, which runs through Aug. 28, the "China Corner" will also host a series of thematic side events to facilitate exchanges on ideas, technologies, and practices for combating desertification.
People visit the exhibition area of the China Corner in the Blue Zone of the 17th Session of the Conference of the Parties (COP17) to the United Nations Convention to Combat Desertification (UNCCD) in Ulan Bator, Mongolia, Aug. 17, 2026. The China Corner themed exhibition made its debut on Monday, showcasing China's experience and achievements in combating desertification. (Xinhua/Bolortsetseg)

Exhibits are displayed at the exhibition area of the "China Corner" in the Blue Zone of the 17th Session of the Conference of the Parties (COP17) to the United Nations Convention to Combat Desertification (UNCCD) in Ulan Bator, Mongolia, Aug. 17, 2026. The "China Corner" themed exhibition made its debut on Monday, showcasing China's experience and achievements in combating desertification. (Xinhua/Bolortsetseg)

A staff member speaks with a visitor at the exhibition area of the "China Corner" in the Blue Zone of the 17th Session of the Conference of the Parties (COP17) to the United Nations Convention to Combat Desertification (UNCCD) in Ulan Bator, Mongolia, Aug. 17, 2026. The "China Corner" themed exhibition made its debut on Monday, showcasing China's experience and achievements in combating desertification. (Xinhua/Bolortsetseg)
。近日，在2026世界人工智能大会“语料筑基智生时代”语料创新论坛圆桌环节，业界热议数据“进化”。中新网记者郑莹莹摄　　在“语料筑基智生时代”语料创新论坛上，上海库帕思科技有限公司首席运营官施佳樑指出，数据是AI的粮食，AI4S的“数据粮食”不仅要多，更要真、细、全、可闭环，还需要业界共同探索如何把原始湿数据转变成可训练、可验证、可流通的数据要素。　　上海创智学院青年研究员唐波昊从另一个视角看“数据进化”。他指出，这一年来人们开始重视实验过程内容的收集和处理，而痛点体现在负样本的收集上。“现在模型训练主要来源于公开的数据集以及公开文献，这些文献都是经过精心处理的，很少有人把失败的实验放进去。而科学发展是不断试错的过程，科研过程中的失败样本或异常样本非常重要。

二 | ” 　　唐波昊还表示，现在科研主要是以人为中心，人把任务拆分给AI，这个过程还需要人不断介入，效率并没有很高。如果科研范式发生转变、未来以AI为中心的话，人的角色将转为判断AI研究方向的可行性。　　上海人工智能实验室青年研究员赵润锴称，目前智能体欠缺自我反思和自我纠错的能力。“如果智能体不具备这样的能力，需要人来纠错，从人工成本维度考量，可能还不如靠人从头开始去执行科学实验。” 　　在上海科学智能研究院AI科学家魏龙看来，大模型的生成能力已经比较强，瓶颈在验证层面。“也就是说，难以验证结果是否有科学价值，甚至是否正确。这就好比导师带学生，导师提供的评估信号非常关键，好的信号才能引导学生往更高价值的方向去探索。事实上，目前这样的信号还难以提供。