NSF Convergence Accelerator - Track C: SQAI: Scalable Quantum Artificial Intelligence for Discovery

NSF融合加速器支持以使用为灵感、以团队为基础的多学科努力，以应对国家重要性的挑战，并将在不久的将来产生对社会有价值的成果。这笔赠款将通过压缩药物发现时间表和降低成本而造福社会。这将对经济和人类健康产生社会影响。人工智能的一个新分支--生成性对抗网络(GAN)在探索大型化学空间和产生针对某种疾病的新候选药物方面表现出了希望。量子GaN(QGAN)已经成为加速经典GaN的一条途径。该项目将创建一个实验性的量子GaN框架，以在嘈杂的中间尺度量子计算机(NISQ)上探索化合物。这将产生一个新的基于发现的框架，可扩展量子人工智能(SQAI)，可以用于其他应用。这个融合加速器团队将在药物发现的背景下回答基本问题，例如：(I)如何最大限度地利用量子优势，使用NISQ时代的计算机进行药物发现？我们是应该将量子资源用于搜索、辨别和加强奖励，还是应该将它们完全分配给搜索并依赖于常规任务的经典范式？(Ii)特定的分子表示法是否有利于NISQ时代的计算机？(Iii)我们能否增强QGAN中的量子空间，以探索化学空间中与药理相关的领域？(4)在训练和生成过程中，噪声对QGAN有什么影响？(V)是否有其他材料系统可为现有易受噪音影响的量子位元提供抗噪性？(Vi)我们可以为用户提供什么样的灵活性来探索他们的发现应用的量子计算？主要任务是开发软件工具链，以弥合量子人工智能算法和硬件之间的差距，使用NISQ计算机探索药物发现，并准备一支量子智能和多样化的劳动力队伍。面向K-12教师的外展活动将包括专业发展研讨会和与量子计算入门内容相关的课程材料。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The NSF Convergence Accelerator supports use-inspired, team-based, multidisciplinary efforts that address challenges of national importance and will produce deliverables of value to society in the near future. This grant will benefit society by compressing the pharmaceutical discovery timeline and reducing cost. This would have societal impacts both economically and in human health. A new branch of artificial intelligence, Generative Adversarial Networks (GAN), shows promise for exploring a large chemical space and generating novel pharmaceutical candidates targeted for a certain disease. Quantum GAN (QGAN) has emerged as a path to accelerate classical GANs. This project will create an experimental quantum GAN framework to explore chemical compounds on Noisy Intermediate Scale Quantum (NISQ) computers. This will yield a new discovery-based framework, Scalable Quantum Artificial Intelligence (SQAI) which could be employed for other applications.This Convergence Accelerator team will answer fundamental questions in the context of drug discovery such as: (i) How to exploit quantum advantage to the fullest for drug discovery using NISQ-era computers? Should we use quantum resources for search, discrimination and reinforcement of reward or allocate them fully for search and rely on classical paradigm for regular tasks? (ii) Does a particular molecular representation benefit NISQ-era computers? (iii) Can we enhance the quantum ansatz in QGAN to explore pharmacologically relevant areas of chemical space? (iv) What is the implication of noise on QGAN during training and generation? (v) Are there other material systems that will provide noise immunity to the existing noise-prone qubits? (vi) What kind of flexibility can we offer to the users in exploring quantum computing for their discovery application? The main tasks are focused on developing a software toolchain to bridge the gap between quantum AI algorithms and hardware, exploring drug discovery using NISQ computers, and preparing a quantum smart and diverse workforce. Outreach to K-12 teachers will include a professional development workshop and curricular materials related to introductory quantum computing content.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.