<p>北京，2026 年 1 月 2 日（全球新闻通讯社）——WiMi Hologram Cloud Inc.（纳斯达克：WiMi）（&#34;WiMi&#34;或&#34;公司&#34;），一家全球领先的全息增强现实（&#34;AR&#34;）技术提供商，发布了一项突破性成就——基于参数高效量子模块的混合量子 - 经典深度学习技术 QB-Net（量子瓶颈网络）。该技术通过将轻量级量子计算模块嵌入经典 U-Net 深度学习架构，实现了重大突破，瓶颈层的参数数量减少了多达 30 倍，同时保持了与经典 U-Net 相当的性能。这一研发成果不仅展示了混合量子 - 经典人工智能的前沿潜力，还为传统深度学习架构提供了一种全新的优化范式。</p>
<p>量子计算的核心优势在于其能够通过量子比特的叠加态表达高维信息，并在指数维空间中执行线性操作，使其具备超越经典架构的表达和变革能力。然而，在当前阶段，量子硬件仍无法支持大规模量子神经网络或构建完整的量子 U-Net 或量子 Transformer。</p>
<p>因此，WiMi 采取了一条完全不同的路径：不是构建完全量子化的人工智能模型，而是构建量子增强模块。</p>
<p>这一概念源于一个关键观察：深度网络的瓶颈层本质上是高维特征的高密度表达问题，而量子态自然适合表达极高维的向量空间。</p>
<p>当一个经典网络需要数万个参数来完成映射任务时，理论上一个量子态只需几十个量子比特就可以实现相同甚至更高的表达能力。这意味着，只要经典特征能够映射到量子态并通过量子电路进行转换，就有可能以极低的参数数量实现等效能力。</p>
<p>基于这一理念，WiMi 设计了一个可插拔的量子瓶颈模块。该模块以最小的参数数量、结构稳定性、可训练性以及能够集成到经典网络中的能力为核心目标，并已嵌入经典 U-Net，形成 QB-Net。</p>
<p>QB-Net 保留了 U-Net 的整体结构，包括编码器、上采样路径和跳跃连接。然而，在瓶颈层位置，传统的多个卷积层被替换为量子特征压缩 - 转换 - 重建模块。该模块由三个关键步骤组成：</p>
<p>第一步是将经典特征编码为量子态。编码模块使用线性投影或幅度编码等技术，将经典特征张量映射为适合进入量子电路的紧凑向量形式。编码策略的设计遵循两个主要原则：尽可能最小化量子比特的数量，同时保留特征的关键信息而不丢失。</p>
<p>第二步是通过量子电路进行特征转换，这是整个系统的核心环节，也是参数效率的关键。传统的卷积瓶颈层可能包含数十万甚至数百万个参数，而量子电路只需几十到几百个可调旋转参数即可实现等效的表达转换。</p>
<p>WiMi 使用参数化量子电路（PQC），通过层叠构建一个深度可控的量子态变换器。量子电路包括纠缠结构，以确保量子比特之间的信息流动充足，形成比经典线性变换更高维的表示能力。</p>
<p>第三步是将量子态解码回经典张量。通过经典集成和校正模块重建从量子测量中获得的结果，最终返回到经典 U-Net 的解码路径。通过量子瓶颈压缩的特征保留了表达能力，同时以极低的参数数量完成了高维信息的过滤和抽象。整个过程可以直接嵌入现有模型中，而无需修改 U-Net 架构或改变训练范式，实现真正的 “即插即用量子增强”。</p>
<p>WiMi 的 QB-Net 的发布标志着我们公司在量子人工智能技术道路上的关键一步。它不仅证明了量子计算现在可以提供真正的价值，还展示了量子技术与深度学习之间深度融合的巨大潜力。在未来，混合量子 - 经典架构将不再被视为过渡技术，而将成为长期以来人工智能的主流形式之一。</p>
<p>QB-Net 代表了一种全新的思维方式：让量子计算成为人工智能中最有价值的部分，而不是全部。基于参数高效量子模块的混合深度学习框架将为全球人工智能产业带来全新的结构优化范式，并为企业级智能系统提供全新的性能提升路径。</p>
<p><strong>关于 WiMi Hologram Cloud</strong></p>
<p>WiMi Hologram Cloud Inc.（纳斯达克：WiMi）专注于全息云服务，主要集中在专业领域，如车载增强现实全息抬头显示器、3D 全息脉冲激光雷达、头戴式光场全息设备、全息半导体、全息云软件、全息汽车导航、元宇宙全息增强现实/虚拟现实设备以及元宇宙全息云软件。它涵盖了全息增强现实技术的多个方面，包括车载全息增强现实技术、3D 全息脉冲激光雷达技术、全息视觉半导体技术、全息软件开发、全息增强现实虚拟广告技术、全息增强现实虚拟娱乐技术、全息增强现实 SDK 支付、互动全息虚拟通信、元宇宙全息增强现实技术和元宇宙虚拟云服务。WiMi 是一个综合性的全息云技术解决方案提供商。有关更多信息，请访问 http://ir.wimiar.com。</p>
<p><strong>翻译免责声明</strong></p>
<p>本公告的原始版本是官方授权的唯一具有法律约束力的版本。如果中文翻译与原始版本之间存在任何不一致或意义上的差异，以原始版本为准。WiMi Hologram Cloud Inc. 及相关机构和个人对翻译版本不作任何保证，并对因翻译不准确而导致的任何直接或间接损失不承担责任。</p>
<p><strong>投资者咨询，请联系：</strong></p>
<p>WIMI Hologram Cloud Inc.</p>
<p>电子邮件：pr@wimiar.com</p>
<p>ICR, LLC</p>
<p>Robin Yang</p>
<p>电话：+1 (646) 975-9495</p>
<p>电子邮件：wimi@icrinc.com</p>
<p><img src="https://imageproxy.pbkrs.com/https://www.globenewswire.com/newsroom/ti/query-bmY9T1RZeE9UQXhPQ00zTXpReU1EYzFJekl5TlRjek1EQT0?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" width="1" height="1" original-src="https://imageproxy.pbkrs.com/https://www.globenewswire.com/newsroom/ti/query-bmY9T1RZeE9UQXhPQ00zTXpReU1EYzFJekl5TlRjek1EQT0"/></p>
<p><img src="https://imageproxy.pbkrs.com/https://ml.globenewswire.com/media/ODQwZDU1NjctOTc5Yi00MjAzLWE2YjItYmU5ZjBkNGQ2YjE5LTEyNjg4NDktMjAyNi0wMS0wMi1lbg==/tiny/WiMi-Hologram-Cloud-Inc-.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" width="1" height="1" original-src="https://imageproxy.pbkrs.com/https://ml.globenewswire.com/media/ODQwZDU1NjctOTc5Yi00MjAzLWE2YjItYmU5ZjBkNGQ2YjE5LTEyNjg4NDktMjAyNi0wMS0wMi1lbg==/tiny/WiMi-Hologram-Cloud-Inc-.png"/></p>

微美全息

<p>WiMi Hologram Cloud Inc. 宣布在混合量子 - 经典深度学习技术方面取得突破，推出了 QB-Net，该技术将轻量级量子模块集成到 U-Net 架构中。这一创新将瓶颈层中的参数数量减少了多达 30 倍，同时保持了性能。QB-Net 利用量子态进行高维特征表达，为深度学习提供了一种新的优化范式。这一发展标志着量子人工智能技术的重大进步，使混合架构成为未来人工智能的主流</p>

<p>BEIJING, Jan. 02, 2026 (GLOBE NEWSWIRE) -- WiMi Hologram Cloud Inc. (NASDAQ: WiMi) (&#34;WiMi&#34; or the &#34;Company&#34;), a leading global Hologram Augmented Reality (&#34;AR&#34;) Technology provider, released a breakthrough achievement—a hybrid quantum-classical deep learning technology based on parameter-efficient quantum modules, QB-Net (Quantum Bottleneck Network). This technology achieves a major breakthrough by embedding lightweight quantum computing modules into the classical U-Net deep learning architecture, reducing the number of parameters in the bottleneck layer by up to 30 times while maintaining performance comparable to that of the classical U-Net. This research and development outcome not only demonstrates the cutting-edge potential of hybrid quantum-classical artificial intelligence but also provides a brand-new optimization paradigm for traditional deep learning architectures.</p>
<div class="lb-trans"><p>北京，2026 年 1 月 2 日（全球新闻通讯社）——WiMi Hologram Cloud Inc.（纳斯达克：WiMi）（&#34;WiMi&#34;或&#34;公司&#34;），一家全球领先的全息增强现实（&#34;AR&#34;）技术提供商，发布了一项突破性成就——基于参数高效量子模块的混合量子 - 经典深度学习技术 QB-Net（量子瓶颈网络）。该技术通过将轻量级量子计算模块嵌入经典 U-Net 深度学习架构，实现了重大突破，瓶颈层的参数数量减少了多达 30 倍，同时保持了与经典 U-Net 相当的性能。这一研发成果不仅展示了混合量子 - 经典人工智能的前沿潜力，还为传统深度学习架构提供了一种全新的优化范式。</p>
</div><p>The core advantage of quantum computing lies in its ability to express high-dimensional information through the superposition states of qubits and perform linear operations in exponentially dimensional spaces, endowing it with expressive and transformative capabilities that surpass classical architectures. However, at the current stage, quantum hardware is still unable to support large-scale quantum neural networks or construct complete quantum U-Net or quantum Transformer.</p>
<div class="lb-trans"><p>量子计算的核心优势在于其能够通过量子比特的叠加态表达高维信息，并在指数维空间中执行线性操作，使其具备超越经典架构的表达和变革能力。然而，在当前阶段，量子硬件仍无法支持大规模量子神经网络或构建完整的量子 U-Net 或量子 Transformer。</p>
</div><p>Therefore, WiMi has taken a completely different path: instead of building fully quantized AI models, it constructs quantum enhancement modules.</p>
<div class="lb-trans"><p>因此，WiMi 采取了一条完全不同的路径：不是构建完全量子化的人工智能模型，而是构建量子增强模块。</p>
</div><p>This concept stems from a key observation: the bottleneck layer of deep networks is essentially a problem of high-density expression of high-dimensional features, while quantum states are naturally suited to express extremely high-dimensional vector spaces.</p>
<div class="lb-trans"><p>这一概念源于一个关键观察：深度网络的瓶颈层本质上是高维特征的高密度表达问题，而量子态自然适合表达极高维的向量空间。</p>
</div><p>When a classical network requires tens of thousands of parameters to accomplish a mapping task, a single quantum state can theoretically achieve the same or even higher expressive power with only a few dozen qubits. This means that as long as classical features can be mapped into quantum states and transformed through quantum circuits, it is possible to achieve equivalent capabilities with extremely low parameter counts.</p>
<div class="lb-trans"><p>当一个经典网络需要数万个参数来完成映射任务时，理论上一个量子态只需几十个量子比特就可以实现相同甚至更高的表达能力。这意味着，只要经典特征能够映射到量子态并通过量子电路进行转换，就有可能以极低的参数数量实现等效能力。</p>
</div><p>Based on this idea, WiMi designed a pluggable Quantum Bottleneck Module. This module takes minimal parameter count, structural stability, trainability, and the ability to be integrated into classical networks as its core objectives and has been embedded into the classical U-Net, forming QB-Net.</p>
<div class="lb-trans"><p>基于这一理念，WiMi 设计了一个可插拔的量子瓶颈模块。该模块以最小的参数数量、结构稳定性、可训练性以及能够集成到经典网络中的能力为核心目标，并已嵌入经典 U-Net，形成 QB-Net。</p>
</div><p>QB-Net retains the overall structure of U-Net, including the encoder, upsampling path, and skip connections. However, at the bottleneck layer position, the traditional multiple convolutional layers are replaced with a quantum feature compression-transformation-reconstruction module. This module consists of three key steps:</p>
<div class="lb-trans"><p>QB-Net 保留了 U-Net 的整体结构，包括编码器、上采样路径和跳跃连接。然而，在瓶颈层位置，传统的多个卷积层被替换为量子特征压缩 - 转换 - 重建模块。该模块由三个关键步骤组成：</p>
</div><p>The first step is the encoding of classical features into quantum states. The encoding module uses techniques such as linear projection or amplitude encoding to map the classical feature tensor into a compact vector form suitable for entering quantum circuits. The design of the encoding strategy follows two major principles: minimizing the number of qubits as much as possible while preserving the key information of the features without loss.</p>
<div class="lb-trans"><p>第一步是将经典特征编码为量子态。编码模块使用线性投影或幅度编码等技术，将经典特征张量映射为适合进入量子电路的紧凑向量形式。编码策略的设计遵循两个主要原则：尽可能最小化量子比特的数量，同时保留特征的关键信息而不丢失。</p>
</div><p>The second step is feature transformation through quantum circuits, which is the core link of the entire system and the key to parameter efficiency. A traditional convolutional bottleneck layer may contain hundreds of thousands or even millions of parameters, whereas a quantum circuit requires only tens to hundreds of adjustable rotation parameters to achieve equivalent expressive transformation.</p>
<div class="lb-trans"><p>第二步是通过量子电路进行特征转换，这是整个系统的核心环节，也是参数效率的关键。传统的卷积瓶颈层可能包含数十万甚至数百万个参数，而量子电路只需几十到几百个可调旋转参数即可实现等效的表达转换。</p>
</div><p>WiMi uses parameterized quantum circuits (PQC) and builds a deeply controllable quantum state transformer through layer stacking. The quantum circuit includes entanglement structures to ensure sufficient information flow between qubits, forming higher-dimensional representation capabilities than classical linear transformations.</p>
<div class="lb-trans"><p>WiMi 使用参数化量子电路（PQC），通过层叠构建一个深度可控的量子态变换器。量子电路包括纠缠结构，以确保量子比特之间的信息流动充足，形成比经典线性变换更高维的表示能力。</p>
</div><p>The third step is decoding the quantum state back into a classical tensor. The results obtained from quantum measurement are reconstructed through a classical integration and correction module and finally returned to the decoding path of the classical U-Net. The features compressed through the quantum bottleneck retain expressive power yet complete the filtering and abstraction of high-dimensional information with an extremely low number of parameters. The entire process can be directly embedded into existing models without modifying the U-Net architecture or changing the training paradigm, achieving true “plug-and-play quantum enhancement”.</p>
<div class="lb-trans"><p>第三步是将量子态解码回经典张量。通过经典集成和校正模块重建从量子测量中获得的结果，最终返回到经典 U-Net 的解码路径。通过量子瓶颈压缩的特征保留了表达能力，同时以极低的参数数量完成了高维信息的过滤和抽象。整个过程可以直接嵌入现有模型中，而无需修改 U-Net 架构或改变训练范式，实现真正的 “即插即用量子增强”。</p>
</div><p>The release of WiMi's QB-Net marks a key step forward for our company on the path of quantum AI technology. It not only proves that quantum computing can deliver real value right now but also demonstrates the enormous potential of deep integration between quantum technology and deep learning. In the future, hybrid quantum-classical architectures will no longer be regarded as transitional technologies but will become one of the mainstream forms of AI for a long time to come.</p>
<div class="lb-trans"><p>WiMi 的 QB-Net 的发布标志着我们公司在量子人工智能技术道路上的关键一步。它不仅证明了量子计算现在可以提供真正的价值，还展示了量子技术与深度学习之间深度融合的巨大潜力。在未来，混合量子 - 经典架构将不再被视为过渡技术，而将成为长期以来人工智能的主流形式之一。</p>
</div><p>QB-Net represents a brand-new way of thinking: letting quantum computing become the most valuable part of artificial intelligence rather than the entirety. The hybrid deep learning framework based on parameter-efficient quantum modules will bring a new structural optimization paradigm to the global AI industry and provide a completely new performance improvement path for enterprise-level intelligent systems.</p>
<div class="lb-trans"><p>QB-Net 代表了一种全新的思维方式：让量子计算成为人工智能中最有价值的部分，而不是全部。基于参数高效量子模块的混合深度学习框架将为全球人工智能产业带来全新的结构优化范式，并为企业级智能系统提供全新的性能提升路径。</p>
</div><p><strong>About WiMi Hologram Cloud</strong></p>
<div class="lb-trans"><p><strong>关于 WiMi Hologram Cloud</strong></p>
</div><p>WiMi Hologram Cloud Inc. (NASDAQ: WiMi) focuses on holographic cloud services, primarily concentrating on professional fields such as in-vehicle AR holographic HUD, 3D holographic pulse LiDAR, head-mounted light field holographic devices, holographic semiconductors, holographic cloud software, holographic car navigation, metaverse holographic AR/VR devices, and metaverse holographic cloud software. It covers multiple aspects of holographic AR technologies, including in-vehicle holographic AR technology, 3D holographic pulse LiDAR technology, holographic vision semiconductor technology, holographic software development, holographic AR virtual advertising technology, holographic AR virtual entertainment technology, holographic ARSDK payment, interactive holographic virtual communication, metaverse holographic AR technology, and metaverse virtual cloud services. WiMi is a comprehensive holographic cloud technology solution provider. For more information, please visit http://ir.wimiar.com.</p>
<div class="lb-trans"><p>WiMi Hologram Cloud Inc.（纳斯达克：WiMi）专注于全息云服务，主要集中在专业领域，如车载增强现实全息抬头显示器、3D 全息脉冲激光雷达、头戴式光场全息设备、全息半导体、全息云软件、全息汽车导航、元宇宙全息增强现实/虚拟现实设备以及元宇宙全息云软件。它涵盖了全息增强现实技术的多个方面，包括车载全息增强现实技术、3D 全息脉冲激光雷达技术、全息视觉半导体技术、全息软件开发、全息增强现实虚拟广告技术、全息增强现实虚拟娱乐技术、全息增强现实 SDK 支付、互动全息虚拟通信、元宇宙全息增强现实技术和元宇宙虚拟云服务。WiMi 是一个综合性的全息云技术解决方案提供商。有关更多信息，请访问 http://ir.wimiar.com。</p>
</div><p><strong>Translation Disclaimer</strong></p>
<div class="lb-trans"><p><strong>翻译免责声明</strong></p>
</div><p>The original version of this announcement is the officially authorized and only legally binding version. If there are any inconsistencies or differences in meaning between the Chinese translation and the original version, the original version shall prevail. WiMi Hologram Cloud Inc. and related institutions and individuals make no guarantees regarding the translated version and assume no responsibility for any direct or indirect losses caused by translation inaccuracies.</p>
<div class="lb-trans"><p>本公告的原始版本是官方授权的唯一具有法律约束力的版本。如果中文翻译与原始版本之间存在任何不一致或意义上的差异，以原始版本为准。WiMi Hologram Cloud Inc. 及相关机构和个人对翻译版本不作任何保证，并对因翻译不准确而导致的任何直接或间接损失不承担责任。</p>
</div><p><strong>Investor Inquiries, please contact:</strong></p>
<div class="lb-trans"><p><strong>投资者咨询，请联系：</strong></p>
</div><p>WIMI Hologram Cloud Inc.</p>
<div class="lb-trans"><p>WIMI Hologram Cloud Inc.</p>
</div><p>Email: pr@wimiar.com</p>
<div class="lb-trans"><p>电子邮件：pr@wimiar.com</p>
</div><p>ICR, LLC</p>
<div class="lb-trans"><p>ICR, LLC</p>
</div><p>Robin Yang</p>
<div class="lb-trans"><p>Robin Yang</p>
</div><p>Tel: +1 (646) 975-9495</p>
<div class="lb-trans"><p>电话：+1 (646) 975-9495</p>
</div><p>Email: wimi@icrinc.com</p>
<div class="lb-trans"><p>电子邮件：wimi@icrinc.com</p>
</div><p><img src="https://imageproxy.pbkrs.com/https://www.globenewswire.com/newsroom/ti/query-bmY9T1RZeE9UQXhPQ00zTXpReU1EYzFJekl5TlRjek1EQT0" alt="" original-src="https://www.globenewswire.com/newsroom/ti?nf=OTYxOTAxOCM3MzQyMDc1IzIyNTczMDA="/></p>
<p><img src="https://imageproxy.pbkrs.com/https://ml.globenewswire.com/media/ODQwZDU1NjctOTc5Yi00MjAzLWE2YjItYmU5ZjBkNGQ2YjE5LTEyNjg4NDktMjAyNi0wMS0wMi1lbg==/tiny/WiMi-Hologram-Cloud-Inc-.png" alt="" original-src="https://ml.globenewswire.com/media/ODQwZDU1NjctOTc5Yi00MjAzLWE2YjItYmU5ZjBkNGQ2YjE5LTEyNjg4NDktMjAyNi0wMS0wMi1lbg==/tiny/WiMi-Hologram-Cloud-Inc-.png"/></p>

微美全息通过将量子模块高效嵌入到 U-Net 中，实现了轻量化设计与高性能的共存 | WiMi Hologram Cloud - Class B 股票新闻