在人工智能的发展历程中，自然语言处理（NLP）与计算机视觉（CV）的飞速发展已深刻重塑了人机互动的图景。如今，这股变革的浪潮正汹涌澎湃地涌入机器人技术领域，特别是具身智能的崭新篇章。清华大学的研究团队近期取得了具有里程碑意义的突破——揭示了data scaling laws的奥秘。这一发现不仅惊人地揭示了机器人领域与语言模型之间的深刻相似性，更为我们预测数据规模与模型性能之间的关系提供了坚实支撑。
</section><p> </p><p style="margin: 0px 0pt 16px;text-align: justify;font-family: 等线;font-size: 12pt;"><span style="font-family: Cambria;letter-spacing: 0pt;vertical-align: baseline;font-size: 15px;">研究团队借助便携式手持夹爪UMI，在丰富多样的真实环境中精心收集了超过4万条人类演示数据。这些数据广泛覆盖了火锅店、咖啡厅、公园等多种日常场景，更不乏喷泉旁、电梯内等独特环境，为模型训练提供了丰富的素材。</span><span lang="EN-US"><o:p></o:p></span></p></section><p style="margin: 16px 0pt 0px;text-align: left;font-family: 等线;font-size: 12pt;line-height: 1.6em;"><span style="font-size: 12px;color: rgb(136, 136, 136);"><span style="font-family: Cambria;letter-spacing: 0pt;vertical-align: baseline;">图2：使用UMI采集人类示教数据</span><span style="text-align: justify;font-family: mp-quote, -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.034em;"> </span></span></p><section style="text-align: center;margin-bottom: 16px;line-height: normal;margin-top: 16px;"><img src="https://mmbiz.qpic.cn/mmbiz_png/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwKuR1lEb43ic40gGibdZozxHzS9JpTbr5zbJ5cxBRUPe8m3uOibpp7FsL7g/640?wx_fmt=png&from=appmsg&random=0.95803&random=0.61091&random=0.029196&random=0.&random=0.32874&random=0.075147&random=0.22354&random=0.64447&random=0.01256&random=0.324&random=0.4444&random=0.0088659&random=0.42916&random=0.073211&random=0.&random=0.37663&random=0.05261&random=0.&random=0.99322&random=0.51954&random=0.21546&random=0.0" class="rich_pages wxw-img js_insertlocalimg" data-imgfileid="" data-ratio="1.4525" data-s="300,640" data-src="https://mmbiz.qpic.cn/mmbiz_png/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwKuR1lEb43ic40gGibdZozxHzS9JpTbr5zbJ5cxBRUPe8m3uOibpp7FsL7g/640?wx_fmt=png&from=appmsg&random=0.95803&random=0.61091&random=0.029196&random=0.&random=0.32874&random=0.075147&random=0.22354&random=0.64447&random=0.01256&random=0.324&random=0.4444&random=0.0088659&random=0.42916&random=0.073211&random=0.&random=0.37663&random=0.05261&random=0.&random=0.99322&random=0.51954&random=0.21546&random=0.0" data-type="png" data-w="579" style="height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;"></section><section style="margin: 0px 0pt;text-align: left;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="font-size: 12px;color: rgb(136, 136, 136);"><span style="font-family: Cambria;letter-spacing: 0pt;vertical-align: baseline;">图4：实验涉及的任务场景</span><span style="text-align: justify;font-family: mp-quote, -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.034em;"> </span></span></section><section style="margin: 0px 0pt;text-align: left;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="font-size: 12px;color: rgb(136, 136, 136);"><span style="text-align: justify;font-family: mp-quote, -apple-system-font, BlinkMacSystemFont, "Helvetica Neue", "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.034em;"><br /></span></span></section><ul class="list-paddingleft-1" style="list-style-type: square;"><li style="font-weight: bold;"><h4 style="text-align: justify;margin: 10pt 0pt 0px;"><section data-class="_mbEditor" data-id="16995"><section style="margin-top: 0px;line-height: normal;"><strong style="font-size: var(--articleFontsize);letter-spacing: 0.034em;"> 物体泛化</strong></section></section></h4></li></ul><section style="margin: 16px 0pt;text-align: justify;font-family: 等线;font-size: 12pt;line-height: 1.6em;"><span style="font-family: Cambria;vertical-align: baseline;letter-spacing: 1px;font-size: 15px;">在物体泛化实验中，研究者固定训练环境数量，逐步增加训练物体数量，细致观察模型在未见物体上的表现。此实验旨在揭示模型对新物体的泛化能力如何随训练物体数量的增加而提升。</span><o:p></o:p></section><section style="margin: 0pt 0pt 16px;text-align: justify;font-family: 等线;font-size: 12pt;"><o:p></o:p></section><p style="text-align: center;margin-top: 16px;margin-bottom: 16px;line-height: 1.6em;"><img src="https://mmbiz.qpic.cn/mmbiz_jpg/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwK2aeoiasFrcPXpAPxrtcPmCk1Td9iaQeCVUEdEoD8Sb7FXCo34rWOUGhg/640?wx_fmt=jpeg&from=appmsg&random=0.0&random=0.&random=0.20091&random=0.&random=0.&random=0.3453&random=0.57934&random=0.0881&random=0.39537&random=0.&random=0.66909&random=0.&random=0.&random=0.21436&random=0.6039&random=0.51065&random=0.04465&random=0.44598&random=0.06733&random=0.87477&random=0.92773&random=0.23833" class="rich_pages wxw-img js_insertlocalimg" data-imgfileid="" data-ratio="0.20359" data-s="300,640" data-src="https://mmbiz.qpic.cn/mmbiz_jpg/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwK2aeoiasFrcPXpAPxrtcPmCk1Td9iaQeCVUEdEoD8Sb7FXCo34rWOUGhg/640?wx_fmt=jpeg&from=appmsg&random=0.0&random=0.&random=0.20091&random=0.&random=0.&random=0.3453&random=0.57934&random=0.0881&random=0.39537&random=0.&random=0.66909&random=0.&random=0.&random=0.21436&random=0.6039&random=0.51065&random=0.04465&random=0.44598&random=0.06733&random=0.87477&random=0.92773&random=0.23833" data-type="jpeg" data-w="1002" style="height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;"></p><p style="margin: 0px 0pt;text-align: left;font-family: 等线;font-size: 12pt;line-height: 1.6em;"><span style="font-family: Cambria;letter-spacing: 0pt;vertical-align: baseline;font-size: 12px;color: rgb(136, 136, 136);">图5：对象泛化</span></p><p style="margin: 0px 0pt;text-align: left;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="color: rgb(136, 136, 136);font-family: Cambria;font-size: 12px;letter-spacing: 0pt;">每条曲线对应不同的演示使用量，显示的归一化分数是训练对象数量的函数。</span></p><p style="margin: 0px 0pt;text-align: left;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="color: rgb(136, 136, 136);font-family: Cambria;font-size: 12px;letter-spacing: 0pt;"><br /></span></p><h4 style="text-align: justify;margin: 10pt 0pt 0pt;"><ul class="list-paddingleft-1" style="list-style-type: square;"><li style="font-weight: bold;"><h4 style="text-align: justify;margin: 10pt 0pt 0pt;"><section data-class="_mbEditor" data-id="16995"><p style="line-height: 1.6em;"><strong>环境泛化</strong></p></section></h4></li></ul></h4><section style="margin: 16px 0pt;text-align: justify;font-family: 等线;font-size: 12pt;line-height: 1.6em;"><span style="font-family: Cambria;vertical-align: baseline;letter-spacing: 1px;font-size: 15px;">在环境泛化实验中，研究者固定训练物体数量，逐步增加训练环境数量，并深入评估模型在未见环境中的表现。此实验旨在探究模型对新环境的泛化能力如何随训练环境数量的增加而增强。</span><o:p></o:p></section><p style="text-align: center;margin-bottom: 16px;margin-top: 16px;"><img src="https://mmbiz.qpic.cn/mmbiz_png/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwKxWvHPeib6wjAibWx91xSLox9iczHUicpBvvOibV7MulSOCVWiaPhYdt59EoQ/640?wx_fmt=png&from=appmsg&random=0.32717&random=0.99482&random=0.&random=0.&random=0.&random=0.74044&random=0.01152&random=0.22575&random=0.57074&random=0.68694&random=0.094546&random=0.61626&random=0.12041&random=0.10727&random=0.90261&random=0.45784&random=0.57194&random=0.13763&random=0.4702&random=0.&random=0.6489&random=0.75781" class="rich_pages wxw-img js_insertlocalimg" data-imgfileid="" data-ratio="0." data-s="300,640" data-src="https://mmbiz.qpic.cn/mmbiz_png/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwKxWvHPeib6wjAibWx91xSLox9iczHUicpBvvOibV7MulSOCVWiaPhYdt59EoQ/640?wx_fmt=png&from=appmsg&random=0.32717&random=0.99482&random=0.&random=0.&random=0.&random=0.74044&random=0.01152&random=0.22575&random=0.57074&random=0.68694&random=0.094546&random=0.61626&random=0.12041&random=0.10727&random=0.90261&random=0.45784&random=0.57194&random=0.13763&random=0.4702&random=0.&random=0.6489&random=0.75781" data-type="png" data-w="1075" style="height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;"></p><p style="margin: 0px 0pt;text-align: justify;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="font-family: Cambria;letter-spacing: 0pt;text-align: left;font-size: 12px;color: rgb(136, 136, 136);">图6：环境泛化。</span></p><p style="margin: 0px 0pt;text-align: justify;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="color: rgb(136, 136, 136);font-family: Cambria;font-size: 12px;letter-spacing: 0pt;text-align: left;">每条曲线对应不同的演示使用率，显示的归一化分数是训练环境数量的函数。</span></p><p style="margin: 0px 0pt;text-align: justify;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="color: rgb(136, 136, 136);font-family: Cambria;font-size: 12px;letter-spacing: 0pt;text-align: left;"><br /></span></p><ul class="list-paddingleft-1" style="list-style-type: square;"><li style="font-weight: bold;"><h4 style="text-align: justify;margin: 10pt 0pt 0pt;"><section data-class="_mbEditor" data-id="16995"><p style="line-height: 1.6em;"><strong> </strong><span style="font-size: 16px;"><strong><span style="letter-spacing: 0.034em;">环境-物体组合泛化</span></strong></span></p></section></h4></li></ul><p style="margin-top: 16px;margin-bottom: 16px;line-height: 1.6em;"><span style="font-family: Cambria;font-size: 15px;letter-spacing: 1px;">在环境-物体组合泛化实验中，研究者同时调整训练环境和训练物体的数量，全面评估模型在未见环境-物体组合上的表现。此实验旨在深入剖析模型对新环境-物体组合的泛化能力如何随训练环境-物体对数量的增加而提升。</span></p><section style="text-align: center;margin-bottom: 16px;"><img src="https://mmbiz.qpic.cn/mmbiz_jpg/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwKUNSkout0eHSoq9miaL5lppDhXuTuesTkKliaprjoMaqfjIr4QSfafOVg/640?wx_fmt=jpeg&from=appmsg&random=0.49171&random=0.080469&random=0.0&random=0.07601&random=0.5498&random=0.30618&random=0.29819&random=0.84089&random=0.4313&random=0.&random=0.31761&random=0.01509&random=0.51295&random=0.22797&random=0.24105&random=0.&random=0.16683&random=0.97903&random=0.&random=0.&random=0.52149&random=0.24141" class="rich_pages wxw-img js_insertlocalimg" data-imgfileid="" data-ratio="0.85567" data-s="300,640" data-src="https://mmbiz.qpic.cn/mmbiz_jpg/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwKUNSkout0eHSoq9miaL5lppDhXuTuesTkKliaprjoMaqfjIr4QSfafOVg/640?wx_fmt=jpeg&from=appmsg&random=0.49171&random=0.080469&random=0.0&random=0.07601&random=0.5498&random=0.30618&random=0.29819&random=0.84089&random=0.4313&random=0.&random=0.31761&random=0.01509&random=0.51295&random=0.22797&random=0.24105&random=0.&random=0.16683&random=0.97903&random=0.&random=0.&random=0.52149&random=0.24141" data-type="jpeg" data-w="970" style="height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;"></section><p style="margin: 0px 0pt;text-align: justify;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="font-family: Cambria;letter-spacing: 0pt;text-align: left;font-size: 12px;color: rgb(136, 136, 136);">图7：跨环境和对象的泛化</span><o:p></o:p></p><p style="margin: 0px 0pt;text-align: left;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="font-family: Cambria;letter-spacing: 0pt;font-size: 12px;color: rgb(136, 136, 136);">每条曲线对应不同的每条曲线都对应于所使用的演示的不同比例，并以训练环境-对象对数的函数形式显示归一化分数。</span></p><ol class="list-paddingleft-1" style="list-style-type: decimal;"><li style="font-weight: bold;"></li><li style="font-weight: bold;"></li><li style="font-weight: bold;"></li></ol><p style="text-align: center;margin-bottom: 16px;"><img src="https://mmbiz.qpic.cn/mmbiz_png/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwKKeufEUicl1o9q8griaichcUj9RlyGV71WNU6O6iciaHIekvlMCXh0dmO6uw/640?wx_fmt=png&from=appmsg&random=0.&random=0.58477&random=0.45322&random=0.3758&random=0.60554&random=0.17509&random=0.67387&random=0.&random=0.&random=0.86947&random=0.&random=0.88681&random=0.58578&random=0.62856&random=0.57368&random=0.48782&random=0.12853&random=0.07077&random=0.084498&random=0.41193&random=0.&random=0.48262" class="rich_pages wxw-img js_insertlocalimg" data-imgfileid="" data-ratio="0.88888" data-s="300,640" data-src="https://mmbiz.qpic.cn/mmbiz_png/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwKKeufEUicl1o9q8griaichcUj9RlyGV71WNU6O6iciaHIekvlMCXh0dmO6uw/640?wx_fmt=png&from=appmsg&random=0.&random=0.58477&random=0.45322&random=0.3758&random=0.60554&random=0.17509&random=0.67387&random=0.&random=0.&random=0.86947&random=0.&random=0.88681&random=0.58578&random=0.62856&random=0.57368&random=0.48782&random=0.12853&random=0.07077&random=0.084498&random=0.41193&random=0.&random=0.48262" data-type="png" data-w="1080" style="height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;"></p><section style="margin: 0px 0pt;text-align: left;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="font-family: Cambria;letter-spacing: 0pt;vertical-align: baseline;font-size: 12px;color: rgb(136, 136, 136);">图8：跨环境和对象的泛化</span></section><section style="margin: 0px 0pt;text-align: left;font-family: 等线;font-size: 12pt;line-height: normal;"><span style="font-family: Cambria;letter-spacing: 0pt;vertical-align: baseline;font-size: 12px;color: rgb(136, 136, 136);">每条曲线对应不同的每条曲线都对应于所使用的演示的不同比例，并以训练环境-对象对数的函数形式显示归一化分数。</span><o:p></o:p></section><section style="margin: 9pt 0pt 8px;text-align: justify;font-family: 等线;font-size: 12pt;"><span style="font-family: Cambria;text-decoration: underline;background: rgb(255, 255, 0);vertical-align: baseline;letter-spacing: 1px;font-size: 15px;">这些发现表明，只要数据量足够庞大，机器人将能够自然而然地理解并适应物理世界的复杂多样性。</span></section><section data-class="_mbEditor" data-id=""><section data-class="_mbEditor" data-id="24863"><p> </p><section data-class="_mbEditor" data-id="24863"><section data-id="85855" style="border-width: 0px;border-style: none;border-color: initial;"><section style="text-align: center;margin:10px auto;"><section style="border-top: 2px solid;border-bottom: 2px solid;padding-top: 4px;padding-right: 10px;padding-bottom: 4px;display: inline-block;"><section style="display: inline-block;float: left;width:60px;background-color: rgb(254,254,254);margin-top:-8px;"><section style="display: table;width: 100%;color: inherit;border-color: rgb(72, 192, 163);" data-width="100%"><section style="display: table-cell;line-height:1em;"><span style="font-size: 24px;"><strong><span style="color:#0c5064;"><em>2</em></span></strong></span></section></section></section><section style="color: rgb(12, 80, 100);float: left;"><span style="font-size: 24px;"><strong>数据收集策略的突破</strong></span></section></section></section></section> </section><p> </p></section><p><span style="font-family: Cambria;font-size: 15px;letter-spacing: 1px;">研究团队还成功攻克了业界长期以来的一个难题：在给定操作任务的前提下，如何科学合理地选择环境数量、物体数量以及每个物体的演示次数？</span></p></section><section data-class="_mbEditor" data-id=""><p><span style="font-family: Cambria;font-size: 15px;letter-spacing: 1px;text-indent: -24pt;">除了数据规模，研究团队在模型规模化方面也取得了三项重要发现：</span></p></section><ol class="list-paddingleft-1" style="list-style-type: decimal;"><li style="letter-spacing: 1px;font-size: 15px;font-weight: bold;"></li><li style="letter-spacing: 1px;font-size: 15px;font-weight: bold;"></li><li style="letter-spacing: 1px;font-size: 15px;font-weight: bold;"><p><br /></p></li></ol><section data-class="_mbEditor" data-id=""><section data-class="_mbEditor" data-id="24863"><p><br /></p><section data-class="_mbEditor" data-id="24863"><section data-id="85855" style="border-width: 0px;border-style: none;border-color: initial;"><section style="text-align: center;margin:10px auto;"><section style="border-top: 2px solid;border-bottom: 2px solid;padding-top: 4px;padding-right: 10px;padding-bottom: 4px;display: inline-block;"><section style="display: inline-block;float: left;width:60px;background-color: rgb(254,254,254);margin-top:-8px;"><section style="display: table;width: 100%;color: inherit;border-color: rgb(72, 192, 163);" data-width="100%"><section style="display: table-cell;line-height:1em;"><span style="font-size: 24px;"><strong><span style="color:#0c5064;"><em>4</em></span></strong></span></section></section></section><section style="color: rgb(12, 80, 100);float: left;"><span style="font-size: 24px;"><strong>未来展望</strong></span></section></section></section></section> </section><p> </p></section></section><p><span style="font-family: Cambria;letter-spacing: 1px;font-size: 15px;">数据规模化正引领机器人技术步入一个全新的时代。但研究团队也提醒我们，盲目追求数据量的增长并非明智之举。相较于单纯增加数据量，提升数据质量可能更为关键。未来的挑战在于，如何准确识别出真正需要扩展的数据类型，并探索最高效的数据采集策略，以获取这些高质量的数据资源。</span></p><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;font-family: "PingFang SC", system-ui, -apple-system, BlinkMacSystemFont, "Helvetica Neue", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;background-color: rgb(255, 255, 255);text-align: left;line-height: 1.5em;"><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 12px;color: rgb(136, 136, 136);"><em style="-webkit-tap-highlight-color: transparent;outline: 0px;">Ref：</em></span></strong></section><p style="margin: 0px 0pt;text-align: justify;font-family: 等线;font-size: 12pt;"><span style="color: rgb(136, 136, 136);"><em><span style="font-family: Cambria;letter-spacing: 0pt;vertical-align: baseline;font-size: 12px;">Data Scaling Laws in Imitation Learning for Robotic Manipulation</span></em></span></p><p style="margin: 0px 0pt;text-align: justify;font-family: 等线;font-size: 12pt;"><span style="color: rgb(136, 136, 136);"><em><em style="font-family: "PingFang SC", system-ui, -apple-system, BlinkMacSystemFont, "Helvetica Neue", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;font-size: var(--articleFontsize);-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 12px;letter-spacing: 0.578px;text-decoration-style: solid;text-decoration-color: rgb(136, 136, 136);">编译｜sienna</span></em></em></span></p><p style="-webkit-tap-highlight-color: transparent;outline: 0px;font-family: "PingFang SC", system-ui, -apple-system, BlinkMacSystemFont, "Helvetica Neue", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;letter-spacing: 0.544px;background-color: rgb(255, 255, 255);margin-bottom: 0px;margin-top: 0px;"><span style="color: rgb(136, 136, 136);"><em><em style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 12px;letter-spacing: 0.578px;text-decoration-style: solid;text-decoration-color: rgb(136, 136, 136);">审核｜fafa</span></em></em><em style="-webkit-tap-highlight-color: transparent;outline: 0px;"></em></span></p><section style="margin: 9pt 0pt 8px;text-align: justify;font-family: 等线;font-size: 12pt;"><span style="font-family: Cambria;font-size: 12pt;letter-spacing: 1px;"><br /></span></section><p style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: left;line-height: normal;"><br /></p><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;text-decoration: underline;color: rgb(122, 68, 66);"><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 15px;">第二届线下自主机器人研讨会（ARTS）</span></strong></span></section><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;color: rgb(217, 33, 66);"><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 15px;">仅剩最后100名额！</span></strong></span></section><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;color: rgb(217, 33, 66);"><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 15px;">有意了解者请点击下方三个入口👇</span></strong></span></section><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;color: rgb(217, 33, 66);"><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 15px;"><br style="-webkit-tap-highlight-color: transparent;outline: 0px;" /></span></strong></span></section><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;color: rgb(107, 146, 214);font-size: 14px;"><em style="-webkit-tap-highlight-color: transparent;outline: 0px;">第二届ARTS报名入口及详情须知</em><em style="-webkit-tap-highlight-color: transparent;outline: 0px;"></em></span></section><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;color: rgb(107, 146, 214);font-size: 14px;"><em style="-webkit-tap-highlight-color: transparent;outline: 0px;">第二届ARTS奖学金通知</em><em style="-webkit-tap-highlight-color: transparent;outline: 0px;"></em></span></section><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;color: rgb(107, 146, 214);font-size: 14px;"><em style="-webkit-tap-highlight-color: transparent;outline: 0px;">ARTS 2024 学术辩论通知抢“鲜”发布</em><em style="-webkit-tap-highlight-color: transparent;outline: 0px;"></em></span></section><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><br /></section><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="color: rgb(107, 146, 214);font-size: 14px;"><em>扫码加入ARTS大会交流群：</em></span></section><p style="text-align: center;"><img src="https://mmbiz.qpic.cn/mmbiz_png/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwKkQor9ibfIEQPAlQYPib6mj5GDmUfKOQIW0fDziaFCB1eN98eEicicHic0ygw/640?wx_fmt=png&from=appmsg&random=0.&random=0.66692&random=0.51251&random=0.36597&random=0.&random=0.&random=0.0" class="rich_pages wxw-img" data-galleryid="" data-imgfileid="" data-ratio="1" data-s="300,640" data-src="https://mmbiz.qpic.cn/mmbiz_png/Nabxc8rdYrjEibv7vpoMMF6kgL2NXGWwKkQor9ibfIEQPAlQYPib6mj5GDmUfKOQIW0fDziaFCB1eN98eEicicHic0ygw/640?wx_fmt=png&from=appmsg&random=0.&random=0.66692&random=0.51251&random=0.36597&random=0.&random=0.&random=0.0" data-type="png" data-w="400" style="width: 164px;height: 164px;" /></p><hr style="-webkit-tap-highlight-color: transparent;outline: 0px;border-style: solid;border-right-width: 0px;border-bottom-width: 0px;border-left-width: 0px;border-color: rgba(0, 0, 0, 0.1);transform-origin: 0px 0px;transform: scale(1, 0.5);" /><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: left;line-height: 1.5em;"><br style="-webkit-tap-highlight-color: transparent;outline: 0px;" /></section><p style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;color: rgb(217, 33, 66);"><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 14px;">深蓝知识星球开启！！</span></strong></span></p><p style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 14px;color: rgb(107, 146, 214);">覆盖多个热门领域🔥</span></p><p style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 14px;color: rgb(107, 146, 214);">城市线下沙龙聚会🧑🎓</span></p><p style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 14px;color: rgb(107, 146, 214);">打破信息壁垒的技术交流社区👏</span></p><p style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 14px;color: rgb(121, 123, 170);"><br style="-webkit-tap-highlight-color: transparent;outline: 0px;" /></span></p><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: center;line-height: 1.5em;margin-top: 16px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 14px;color: rgb(121, 123, 170);"><img src="https://mmbiz.qpic.cn/mmbiz_jpg/Nabxc8rdYrjWria3g1eq2bVA5qgxEPgj0icPrCefQzmciaWY8qW0fjeewoIVMZ3C9CUWChcCSov60IsL4hGOb8YlA/640?wx_fmt=other&from=appmsg&random=0.0&wxfrom=5&wx_lazy=1&wx_co=1&random=0.54111&tp=webp&random=0.30016&random=0.79675&random=0.20272&random=0.71461&random=0.&random=0.97204&random=0.02784&random=0.079905&random=0.20434&random=0.&random=0.7022&random=0.04478&random=0.0266&random=0.042244&random=0.75239&random=0.64744&random=0.15966&random=0.3377&random=0.&random=0.72023&random=0.&random=0." class="rich_pages wxw-img" data-backh="212" data-backw="578" data-imgfileid="" data-ratio="0." data-src="https://mmbiz.qpic.cn/mmbiz_jpg/Nabxc8rdYrjWria3g1eq2bVA5qgxEPgj0icPrCefQzmciaWY8qW0fjeewoIVMZ3C9CUWChcCSov60IsL4hGOb8YlA/640?wx_fmt=other&from=appmsg&random=0.0&wxfrom=5&wx_lazy=1&wx_co=1&random=0.54111&tp=webp&random=0.30016&random=0.79675&random=0.20272&random=0.71461&random=0.&random=0.97204&random=0.02784&random=0.079905&random=0.20434&random=0.&random=0.7022&random=0.04478&random=0.0266&random=0.042244&random=0.75239&random=0.64744&random=0.15966&random=0.3377&random=0.&random=0.72023&random=0.&random=0." data-w="1080" style="-webkit-tap-highlight-color: transparent;outline: 0px;width: 677px !important;visibility: visible !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;height: auto !important;"></span></section><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: left;line-height: 1.5em;margin-top: 16px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 14px;color: rgb(136, 136, 136);"><em style="-webkit-tap-highlight-color: transparent;outline: 0px;"><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;">推荐阅读：</strong></em></span></section><section style="-webkit-tap-highlight-color: transparent;margin-bottom: 0px;outline: 0px;text-align: left;line-height: 1.5em;"><p style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="font-size: 14px;"><em>CoEdge: 面向自动驾驶的协作式边缘计算系统，实现分布式实时深度学习任务的高效调度与资源优化</em></span><br style="-webkit-tap-highlight-color: transparent;outline: 0px;" /></p><p style="-webkit-tap-highlight-color: transparent;outline: 0px;"><em><span style="outline: 0px;color: var(--weui-LINK);cursor: default;font-size: 14px;">燃爆！自主机器人技术研讨会 ——19 位青年报告嘉宾集结，火热报名中</span></em><span style="outline: 0px;color: var(--weui-LINK);cursor: default;font-size: 14px;"><em style="-webkit-tap-highlight-color: transparent;outline: 0px;"></em></span><em style="-webkit-tap-highlight-color: transparent;outline: 0px;"></em><br style="-webkit-tap-highlight-color: transparent;outline: 0px;" /></p><section style="-webkit-tap-highlight-color: transparent;outline: 0px;line-height: 1.5em;"><section style="-webkit-tap-highlight-color: transparent;outline: 0px;"><section data-class="_mbEditor" data-id="" style="-webkit-tap-highlight-color: transparent;outline: 0px;"><section style="-webkit-tap-highlight-color: transparent;outline: 0px;overflow: hidden;color: rgb(0, 0, 0);font-size: 16px;"><section style="-webkit-tap-highlight-color: transparent;margin: 10px auto;outline: 0px;display: flex;flex-direction: column;align-items: center;"><section style="-webkit-tap-highlight-color: transparent;padding: 25px 20px;outline: 0px;border-width: 1px;border-style: solid;border-color: rgb(0, 0, 0);"><section style="-webkit-tap-highlight-color: transparent;outline: 0px;white-space-collapse: preserve;font-size: 15px;color: rgb(51, 51, 51);line-height: 1.5em;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;color: rgb(107, 146, 214);"><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-family: "PingFang SC", system-ui, -apple-system, "system-ui", "Helvetica Neue", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;font-size: 14px;letter-spacing: 0.578px;">【深蓝AI】</span></strong></span><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 14px;letter-spacing: 0.578px;">的原创内容均由作者团队倾注个人心血制作而成，希望各位遵守原创规则珍惜作者们的劳动成果，转载请私信后台进行授权，发文时务必注明出自</span></strong><span style="-webkit-tap-highlight-color: transparent;outline: 0px;color: rgb(107, 146, 214);"><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-family: "PingFang SC", system-ui, -apple-system, "system-ui", "Helvetica Neue", "Hiragino Sans GB", "Microsoft YaHei UI", "Microsoft YaHei", Arial, sans-serif;font-size: 14px;letter-spacing: 0.578px;">【深蓝AI】</span></strong></span><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 14px;letter-spacing: 0.578px;">微信公众号，否则侵权必究⚠️⚠️</span></strong></section></section></section></section></section><p style="-webkit-tap-highlight-color: transparent;outline: 0px;text-align: right;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;color: rgb(255, 255, 255);background-color: rgb(107, 146, 214);"><em style="-webkit-tap-highlight-color: transparent;outline: 0px;"><strong style="-webkit-tap-highlight-color: transparent;outline: 0px;"><span style="-webkit-tap-highlight-color: transparent;outline: 0px;font-size: 14px;">*点击赞+在看收藏并推荐本文*</span></strong></em></span><br style="-webkit-tap-highlight-color: transparent;outline: 0px;" /></p></section></section><p style="-webkit-tap-highlight-color: transparent;outline: 0px;"><br /></p></section><p style="display: none;"><mp-style-type data-value="10000"></mp-style-type></p> 讯享网
版权声明：本文内容由互联网用户自发贡献，该文观点仅代表作者本人。本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如发现本站有涉嫌侵权/违法违规的内容，请联系我们，一经查实，本站将立刻删除。
如需转载请保留出处：https://51itzy.com/kjqy/196482.html