Britain gains access to Ukraine battlefield data - FT中文网
登录×
电子邮件/用户名
密码
记住我
请输入邮箱和密码进行绑定操作:
请输入手机号码,通过短信验证(目前仅支持中国大陆地区的手机号)。如果该手机号尚未注册,将自动创建 FT中文网账号。短信可能需要几分钟送达,验证码15分钟内有效,请耐心等待:
请您阅读我们的用户注册协议和隐私权保护政策,点击下方按钮即视为您接受。
商业快报

Britain gains access to Ukraine battlefield data

Trove of combat imagery used to train AI models to identify and strike Russian targets
00:00

{"text":[[{"start":10.48,"text":"Britain will become the first foreign country to gain access to a vast trove of Ukrainian battlefield data used to train AI models to identify and strike Russian targets under a broad technology-sharing partnership."}],[{"start":24.32,"text":"Prime Minister Andy Burnham and Ukrainian President Volodymyr Zelenskyy signed the AI partnership on Monday in Kyiv, opening Ukraine’s Avengers Labs data platform to British researchers and technology companies."}],[{"start":37.36,"text":"The database draws on millions of observations collected by thousands of cameras and sensors deployed across Ukrainian battlefields."}],[{"start":45.52,"text":"The announcement came on Burnham’s first overseas trip as prime minister, during which he vowed to maintain the UK’s full support for Kyiv and denounced “outrageous” threats against Britain by Moscow."}],[{"start":56.64,"text":"Burnham, speaking on the 35th anniversary of Ukraine’s independence, shrugged off recent warnings from Moscow and said: “Despite Russia’s outrageous threats to my country, we will continue our support for Ukraine.”"}],[{"start":69.2,"text":"Ukraine’s Ministry of Defence says Avengers Labs now processes more than 100,000 drone video streams a month and detects about 70 per cent of enemy targets appearing in the imagery. Earlier versions were reportedly identifying about 12,000 pieces of enemy equipment per week."}],[{"start":85.96,"text":"The data comes from what Ukraine’s MoD refers to as its Universal Military Dataset: a huge collection of real combat imagery, mostly from drones and acoustic sensors, that has been organised and manually labelled."}],[{"start":98.8,"text":"It is used to train machine-learning models that can recognise and classify battlefield targets such as tanks, drones and artillery pieces, part of Ukraine’s rapid development of increasingly autonomous weapons during more than four years of war."}],[{"start":112.28,"text":"“High-quality, labelled battlefield data is one of the biggest constraints on developing reliable AI for autonomous systems,” said Misha Nestor, chief product officer for Swarmer, the Ukrainian drone software company. “Ukraine has accumulated something that is extremely difficult to replicate anywhere else: large volumes of data generated under real combat conditions.”"}],[{"start":133.96,"text":"Until now this data has been shared only with domestic companies, giving Ukrainian drone makers a competitive edge over foreign rivals, whose computer vision models were often trained using synthetic data and did not perform so well in combat."}],[{"start":147,"text":"Britain’s military is understood to be particularly interested in data gathered by acoustic sensors placed throughout Ukraine and used to identify incoming Russian drones. Properly trained, these sensors can be far more accurate than radar and could be used to defend military installations and critical infrastructure."}],[{"start":165.04,"text":"Kyiv first announced it would allow international partners access to the dataset in March, offering “the opportunity to train their AI models on real data from modern warfare”, according to Mykhailo Fedorov, Ukraine’s defence minister at the time. In return, Ukraine would gain “faster progress in the development of autonomous systems and new technological solutions for the front”, he said. Zelenskyy sacked Fedorov in July."}],[{"start":191.56,"text":"The agreement with the UK is part of a wider technology swap intended to marry Ukraine’s battlefield experience and unique wartime data sets with Britain’s universities, AI researchers and technology companies."}],[{"start":204.28,"text":"Burnham said the deal “will bring together Ukraine’s unrivalled operational experience with Britain’s world-class AI ecosystem”."}],[{"start":211.6,"text":"It follows Burnham’s announcement that European missile joint venture MBDA would be permitted to release classified information on British components of advanced missiles to help establish assembly lines in Ukraine."}],[{"start":223.52,"text":"Kremlin spokesman Dmitry Peskov said the decision to share technology on the Scalp missile with Kyiv had been received “extremely negatively” by Moscow, which last week warned of unspecified “consequences” for Britain over its support for Ukraine’s long-range strikes on Russia."}],[{"start":239.54,"text":"The UK “methodically and regularly adds fuel to the fire and hatches schemes to prevent even the slightest progress in the peaceful settlement process”, Peskov told reporters on Monday, according to the Interfax state news agency."}],[{"start":253.56,"text":"Speaking at the end of his visit to Kyiv on Monday, Burnham was asked by ITV News whether he would seek to persuade countries such as Spain and Greece to hand over stocks of Patriot missiles to help defend Ukraine."}],[{"start":264.28,"text":"“It’s not for us to say that,” he said. “It’s a matter for those countries first and foremost, but obviously we will use our influence.”"}],[{"start":273.08,"text":"Asked if people in Britain should be worried that the UK was being dragged into the conflict, Burnham replied: “No, we’re not doing anything different. This is continuity.”"}],[{"start":287.88,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1787635872_2923.mp3"}

版权声明:本文版权归FT中文网所有,未经允许任何单位或个人不得转载,复制或以任何其他方式使用本文全部或部分,侵权必究。

民主党横扫中期选举,华尔街大金主押注失利

顶级对冲基金经理和私募股权投资公司高管为共和党竞选活动投入逾3亿美元,却几乎未见成效。
2小时前

美国的清晨,市场的黄昏

罗奇:美国中期选举结果再令人震惊,也阻止不了经济金融周期的转向。

共和党为特朗普的失败政策付出代价

特朗普承诺增加制造业就业岗位、提高美国本土出生人口的就业率,但这两项承诺都未兑现。

“鲁珀特不想和失败者为伍”:默多克的媒体帝国对特朗普态度转冷

共和党遭遇失利后,福克斯新闻主持人质疑对伊朗战争的战略,《纽约邮报》则在头版刊出“蓝色海啸”。

Lex专栏:美国选民给数据中心建设热潮降温,未必是坏事

美国中期选举结果可能影响45%的数据中心扩建计划,但对投资者而言,这未尝不是好消息。

英国移民改革:有哪些变化?谁受益?

根据周一公布的改革计划,高收入者及其雇主可能从中受益,但英国也可能因此吓退国际学生。
设置字号×
最小
较小
默认
较大
最大
分享×