{"id":91367,"topic":"ai","source":"新浪财经","title":"人工智能可在二十毫秒内预测和控制聚变等离子体|普林斯顿大学|核聚变|高温|美国|研究人员_手机新浪网 - 新浪财经","url":"https://finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html?vt=4","url_hash":"69bba455b30f1d98cc8641f844596d9748f6961e","author":"","summary":"<a href=\"https://news.google.com/rss/articles/CBMieEFVX3lxTFA5UFlOR3l1ZVFfQ0JNSjRRUGlHVWJqc3pyVi1WZ2p5TnNONEVDcUgteTJkVDhBa09qajNBUWdzZS1GYjI3S2x3bjNRaUgyYy1OV2dOWnM1R0NVQ0RsYUZoSmY2bGVfZ2swT0lCT0ZYdlBlRXNZZElrVg?oc=5\" target=\"_blank\">人工智能可在二十毫秒内预测和控制聚变等离子体|普林斯顿大学|核聚变|高温|美国|研究人员_手机新浪网</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">新浪财经</font>","content":"人工智能可在 二十毫秒内预测和控制聚变等离子体\n●张佳欣\n美国能源部普林斯顿等离子体物理实验室和普林斯顿大学研究人员开发出一种人工智能（AI）软件框架，可在约20毫秒内完成一次对聚变等离子体的预测与控制，快速调整其状态，同时受到严格的硬件安全限制。相关成果发表于新一期《核聚变》杂志。\n科学家正在利用托卡马克等装置探索可控核聚变。托卡马克装置通过强磁场把高温等离子体约束在其内部。要让聚变反应持续进行，等离子体需要保持高温、高密度和稳定状态。但等离子体中的一些不稳定现象发展得非常快，可能在几毫秒内迅速增强，人工操作员往往来不及作出反应。\n预测等离子体接下来会发生什么，是控制聚变反应的关键。传统计算机模拟一次可能需要数天甚至数月，无法满足实时控制要求。机器学习模型则能利用实验数据快速判断等离子体状态，并预测其变化。\n此次，研究人员开发的这套系统名为PACMAN，即“利用机器学习进行预测与控制”。它可以同时调用多个AI模型，根据预测结果调整加热设备、磁场等相关参数。系统还会检查AI给出的控制方案，并设置硬件安全限制，确保设备不会执行超出安全范围的操作。\n在美国能源部DIII-D国家聚变设施的托卡马克上开展的5项实验结果显示，PACMAN能够利用AI控制等离子体加热系统，预测等离子体边缘的能量爆发，探测并控制由高速粒子驱动的等离子体波，还能根据预设目标调整等离子体密度和旋转状态。\n研究人员还利用PACMAN同时控制DIII-D装置的6个回旋管。回旋管通过强力微波束向等离子体提供能量。实验中，AI能够同时调整6个回旋管的功率和反射镜位置，让这些设备协同工作，以达到研究人员设定的目标。据《科技日报》","image_url":"//n.sinaimg.cn/default/2fb77759/20151125/320X320.png","lang":"zh","published_at":"2026-09-27T19:55:15+00:00","fetched_at":"2026-09-27T20:15:07+00:00","status":"read","starred":0,"extract_state":"ok","summary_auto":"人工智能可在…","cluster_id":null,"extract_retries":0,"extract_error":null,"contract_version":"news_item.v1","format_contract_version":"news_item_formats.v1","dedup_url":"https://finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html?vt=4","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 704 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":704,"summary_length":7,"usable_text_length":704,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":704,"summary_length":7}},"news_item":{"id":91367,"canonical_url":"https://finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html?vt=4","source_url":"https://finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html?vt=4","title":"人工智能可在二十毫秒内预测和控制聚变等离子体|普林斯顿大学|核聚变|高温|美国|研究人员_手机新浪网 - 新浪财经","source_name":"新浪财经","author":null,"published_at":"2026-09-27T19:55:15+00:00","locale":"zh","topic":"ai","tags":[],"rss_summary":"<a href=\"https://news.google.com/rss/articles/CBMieEFVX3lxTFA5UFlOR3l1ZVFfQ0JNSjRRUGlHVWJqc3pyVi1WZ2p5TnNONEVDcUgteTJkVDhBa09qajNBUWdzZS1GYjI3S2x3bjNRaUgyYy1OV2dOWnM1R0NVQ0RsYUZoSmY2bGVfZ2swT0lCT0ZYdlBlRXNZZElrVg?oc=5\" target=\"_blank\">人工智能可在二十毫秒内预测和控制聚变等离子体|普林斯顿大学|核聚变|高温|美国|研究人员_手机新浪网</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">新浪财经</font>","full_text":"人工智能可在 二十毫秒内预测和控制聚变等离子体\n●张佳欣\n美国能源部普林斯顿等离子体物理实验室和普林斯顿大学研究人员开发出一种人工智能（AI）软件框架，可在约20毫秒内完成一次对聚变等离子体的预测与控制，快速调整其状态，同时受到严格的硬件安全限制。相关成果发表于新一期《核聚变》杂志。\n科学家正在利用托卡马克等装置探索可控核聚变。托卡马克装置通过强磁场把高温等离子体约束在其内部。要让聚变反应持续进行，等离子体需要保持高温、高密度和稳定状态。但等离子体中的一些不稳定现象发展得非常快，可能在几毫秒内迅速增强，人工操作员往往来不及作出反应。\n预测等离子体接下来会发生什么，是控制聚变反应的关键。传统计算机模拟一次可能需要数天甚至数月，无法满足实时控制要求。机器学习模型则能利用实验数据快速判断等离子体状态，并预测其变化。\n此次，研究人员开发的这套系统名为PACMAN，即“利用机器学习进行预测与控制”。它可以同时调用多个AI模型，根据预测结果调整加热设备、磁场等相关参数。系统还会检查AI给出的控制方案，并设置硬件安全限制，确保设备不会执行超出安全范围的操作。\n在美国能源部DIII-D国家聚变设施的托卡马克上开展的5项实验结果显示，PACMAN能够利用AI控制等离子体加热系统，预测等离子体边缘的能量爆发，探测并控制由高速粒子驱动的等离子体波，还能根据预设目标调整等离子体密度和旋转状态。\n研究人员还利用PACMAN同时控制DIII-D装置的6个回旋管。回旋管通过强力微波束向等离子体提供能量。实验中，AI能够同时调整6个回旋管的功率和反射镜位置，让这些设备协同工作，以达到研究人员设定的目标。据《科技日报》","excerpt":"人工智能可在…","extraction":{"state":"ok","confidence":0.9,"error":null,"explanation":"High confidence: full text extraction produced 704 characters.","diagnostics_url":"/api/diagnose?url=https%3A//finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html%3Fvt%3D4","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 704 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":704,"summary_length":7,"usable_text_length":704,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":704,"summary_length":7}}},"display_formats":["compact","card","full","digest_section","json"]},"daily_stack_record":{"title":"人工智能可在二十毫秒内预测和控制聚变等离子体|普林斯顿大学|核聚变|高温|美国|研究人员_手机新浪网 - 新浪财经","url":"https://finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html?vt=4","summary":"人工智能可在…","source":"新浪财经","date":"2026-09-27T19:55:15+00:00","content":"人工智能可在 二十毫秒内预测和控制聚变等离子体\n●张佳欣\n美国能源部普林斯顿等离子体物理实验室和普林斯顿大学研究人员开发出一种人工智能（AI）软件框架，可在约20毫秒内完成一次对聚变等离子体的预测与控制，快速调整其状态，同时受到严格的硬件安全限制。相关成果发表于新一期《核聚变》杂志。\n科学家正在利用托卡马克等装置探索可控核聚变。托卡马克装置通过强磁场把高温等离子体约束在其内部。要让聚变反应持续进行，等离子体需要保持高温、高密度和稳定状态。但等离子体中的一些不稳定现象发展得非常快，可能在几毫秒内迅速增强，人工操作员往往来不及作出反应。\n预测等离子体接下来会发生什么，是控制聚变反应的关键。传统计算机模拟一次可能需要数天甚至数月，无法满足实时控制要求。机器学习模型则能利用实验数据快速判断等离子体状态，并预测其变化。\n此次，研究人员开发的这套系统名为PACMAN，即“利用机器学习进行预测与控制”。它可以同时调用多个AI模型，根据预测结果调整加热设备、磁场等相关参数。系统还会检查AI给出的控制方案，并设置硬件安全限制，确保设备不会执行超出安全范围的操作。\n在美国能源部DIII-D国家聚变设施的托卡马克上开展的5项实验结果显示，PACMAN能够利用AI控制等离子体加热系统，预测等离子体边缘的能量爆发，探测并控制由高速粒子驱动的等离子体波，还能根据预设目标调整等离子体密度和旋转状态。\n研究人员还利用PACMAN同时控制DIII-D装置的6个回旋管。回旋管通过强力微波束向等离子体提供能量。实验中，AI能够同时调整6个回旋管的功率和反射镜位置，让这些设备协同工作，以达到研究人员设定的目标。据《科技日报》","confidence":0.9,"diagnostics_url":"/api/diagnose?url=https%3A//finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html%3Fvt%3D4","quality_bucket":"high","failure_kind":"none","retryable":false,"quality_reason":"High confidence: full text extraction produced 704 characters.","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 704 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":704,"summary_length":7,"usable_text_length":704,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":704,"summary_length":7}},"tags":[]},"fallback_formats":["markdown","json","html"],"actions":{"read":"/item/91367","export_markdown":"/api/items/91367/export?format=markdown","export_json":"/api/items/91367/export?format=json","diagnose":"/api/diagnose?url=https%3A//finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html%3Fvt%3D4"},"formats":{"full":{"id":91367,"title":"人工智能可在二十毫秒内预测和控制聚变等离子体|普林斯顿大学|核聚变|高温|美国|研究人员_手机新浪网 - 新浪财经","url":"https://finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html?vt=4","source":"新浪财经","author":null,"published_at":"2026-09-27T19:55:15+00:00","locale":"zh","topic":"ai","tags":[],"excerpt":"人工智能可在…","full_text":"人工智能可在 二十毫秒内预测和控制聚变等离子体\n●张佳欣\n美国能源部普林斯顿等离子体物理实验室和普林斯顿大学研究人员开发出一种人工智能（AI）软件框架，可在约20毫秒内完成一次对聚变等离子体的预测与控制，快速调整其状态，同时受到严格的硬件安全限制。相关成果发表于新一期《核聚变》杂志。\n科学家正在利用托卡马克等装置探索可控核聚变。托卡马克装置通过强磁场把高温等离子体约束在其内部。要让聚变反应持续进行，等离子体需要保持高温、高密度和稳定状态。但等离子体中的一些不稳定现象发展得非常快，可能在几毫秒内迅速增强，人工操作员往往来不及作出反应。\n预测等离子体接下来会发生什么，是控制聚变反应的关键。传统计算机模拟一次可能需要数天甚至数月，无法满足实时控制要求。机器学习模型则能利用实验数据快速判断等离子体状态，并预测其变化。\n此次，研究人员开发的这套系统名为PACMAN，即“利用机器学习进行预测与控制”。它可以同时调用多个AI模型，根据预测结果调整加热设备、磁场等相关参数。系统还会检查AI给出的控制方案，并设置硬件安全限制，确保设备不会执行超出安全范围的操作。\n在美国能源部DIII-D国家聚变设施的托卡马克上开展的5项实验结果显示，PACMAN能够利用AI控制等离子体加热系统，预测等离子体边缘的能量爆发，探测并控制由高速粒子驱动的等离子体波，还能根据预设目标调整等离子体密度和旋转状态。\n研究人员还利用PACMAN同时控制DIII-D装置的6个回旋管。回旋管通过强力微波束向等离子体提供能量。实验中，AI能够同时调整6个回旋管的功率和反射镜位置，让这些设备协同工作，以达到研究人员设定的目标。据《科技日报》","reading_time_min":1,"extraction":{"state":"ok","confidence":0.9,"error":null,"explanation":"High confidence: full text extraction produced 704 characters.","diagnostics_url":"/api/diagnose?url=https%3A//finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html%3Fvt%3D4","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 704 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":704,"summary_length":7,"usable_text_length":704,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":704,"summary_length":7}}},"quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 704 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":704,"summary_length":7,"usable_text_length":704,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":704,"summary_length":7}},"actions":{"read":"/item/91367","export_markdown":"/api/items/91367/export?format=markdown","export_json":"/api/items/91367/export?format=json","diagnose":"/api/diagnose?url=https%3A//finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html%3Fvt%3D4"}},"digest":{"id":91367,"title":"人工智能可在二十毫秒内预测和控制聚变等离子体|普林斯顿大学|核聚变|高温|美国|研究人员_手机新浪网 - 新浪财经","url":"https://finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html?vt=4","source":"新浪财经","topic":"ai","published_at":"2026-09-27T19:55:15+00:00","excerpt":"人工智能可在…","quality_bucket":"high","quality_reason":"High confidence: full text extraction produced 704 characters.","reading_time_min":1,"cluster_id":null},"card":{"display_title":"人工智能可在二十毫秒内预测和控制聚变等离子体|普林斯顿大学|核聚变|高温|美国|研究人员_手机新浪网 - 新浪财经","subtitle":"新浪财经 · 2026-09-27","summary":"人工智能可在…","badges":["quality:high"],"links":{"read":"/item/91367","original":"https://finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html?vt=4","diagnose":"/api/diagnose?url=https%3A//finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html%3Fvt%3D4"},"quality_warning":null},"export":{"title":"人工智能可在二十毫秒内预测和控制聚变等离子体|普林斯顿大学|核聚变|高温|美国|研究人员_手机新浪网 - 新浪财经","url":"https://finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html?vt=4","summary":"人工智能可在…","source":"新浪财经","date":"2026-09-27T19:55:15+00:00","content":"人工智能可在 二十毫秒内预测和控制聚变等离子体\n●张佳欣\n美国能源部普林斯顿等离子体物理实验室和普林斯顿大学研究人员开发出一种人工智能（AI）软件框架，可在约20毫秒内完成一次对聚变等离子体的预测与控制，快速调整其状态，同时受到严格的硬件安全限制。相关成果发表于新一期《核聚变》杂志。\n科学家正在利用托卡马克等装置探索可控核聚变。托卡马克装置通过强磁场把高温等离子体约束在其内部。要让聚变反应持续进行，等离子体需要保持高温、高密度和稳定状态。但等离子体中的一些不稳定现象发展得非常快，可能在几毫秒内迅速增强，人工操作员往往来不及作出反应。\n预测等离子体接下来会发生什么，是控制聚变反应的关键。传统计算机模拟一次可能需要数天甚至数月，无法满足实时控制要求。机器学习模型则能利用实验数据快速判断等离子体状态，并预测其变化。\n此次，研究人员开发的这套系统名为PACMAN，即“利用机器学习进行预测与控制”。它可以同时调用多个AI模型，根据预测结果调整加热设备、磁场等相关参数。系统还会检查AI给出的控制方案，并设置硬件安全限制，确保设备不会执行超出安全范围的操作。\n在美国能源部DIII-D国家聚变设施的托卡马克上开展的5项实验结果显示，PACMAN能够利用AI控制等离子体加热系统，预测等离子体边缘的能量爆发，探测并控制由高速粒子驱动的等离子体波，还能根据预设目标调整等离子体密度和旋转状态。\n研究人员还利用PACMAN同时控制DIII-D装置的6个回旋管。回旋管通过强力微波束向等离子体提供能量。实验中，AI能够同时调整6个回旋管的功率和反射镜位置，让这些设备协同工作，以达到研究人员设定的目标。据《科技日报》","confidence":0.9,"diagnostics_url":"/api/diagnose?url=https%3A//finance.sina.cn/2026-09-28/detail-initifmt9866647.d.html%3Fvt%3D4","quality_bucket":"high","failure_kind":"none","retryable":false,"quality_reason":"High confidence: full text extraction produced 704 characters.","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 704 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":704,"summary_length":7,"usable_text_length":704,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":704,"summary_length":7}},"tags":[],"format_contract_version":"news_item_formats.v1"}}}