arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2407.14560cs.LGcs.AIcs.AR
收藏

实现像素内智能的神经网络与ASIC的自动化整体协同设计

Automated and Holistic Co-design of Neural Networks and ASICs for Enabling In-Pixel Intelligence

Brookhaven National Laboratory(布鲁克海文国家实验室);Computational Science Initiative(计算科学倡议);Instrumentation Department(仪器部)

Shubha R. Kharel, Prashansa Mukim, Piotr Maj, Grzegorz W. Deptuch, Shinjae Yoo, Yihui Ren, Soumyajit Mandal

AI总结:

针对极端边缘AI系统严格的硬件约束与庞大设计空间,提出基于多目标贝叶斯优化的智能搜索方法,将神经网络搜索与ASIC综合集成于循环中,成功为读出ASIC像素内实时脉冲特征提取找到Pareto最优设计。

摘要:

极端边缘AI系统,例如用于辐射探测的读出ASIC中的系统,必须在严格的硬件约束下运行,如微米级尺寸、亚毫瓦功耗和纳秒级速度,同时需提供明显优于传统架构的准确率优势。寻找理想解决方案意味着要从这两个领域合并过程中爆炸式扩大的设计空间中,确定最优的AI和ASIC设计选择,这产生了非平凡的耦合,随着约束收紧,这些耦合共同作为约束作用于少量解集。在即使是小规模问题中也轻易超过数十亿种可能性中手动确定理想选择,即使不是不可能的,也是不切实际的。弥合这一差距的现有方法利用对硬件的理论理解来指导架构搜索。然而,计算此类理论指标时所做的假设过于理想化,无法在困难地搜索实际实现过程中提供充分指导。同时,许多其他关键指标(如延迟)的理论估计甚至不存在,且同样具有可变性,取决于工艺设计套件(PDK)的参数。为了应对这些挑战,我们提出了一项研究,采用基于多目标贝叶斯优化的智能搜索,将神经网络搜索和ASIC综合集成在循环中。该方法对跨领域设计选择的集体影响提供可靠反馈。我们通过为执行实时特征提取的有效且高效的神经网络找到几个Pareto最优设计选择,展示了我们方法的有效性,这些神经网络从读出ASIC单个像素内的输入脉冲中提取特征。

英文摘要:

Extreme edge-AI systems, such as those in readout ASICs for radiation detection, must operate under stringent hardware constraints such as micron-level dimensions, sub-milliwatt power, and nanosecond-scale speed while providing clear accuracy advantages over traditional architectures. Finding ideal solutions means identifying optimal AI and ASIC design choices from a design space that has explosively expanded during the merger of these domains, creating non-trivial couplings which together act upon a small set of solutions as constraints tighten. It is impractical, if not impossible, to manually determine ideal choices among possibilities that easily exceed billions even in small-size problems. Existing methods to bridge this gap have leveraged theoretical understanding of hardware to f architecture search. However, the assumptions made in computing such theoretical metrics are too idealized to provide sufficient guidance during the difficult search for a practical implementation. Meanwhile, theoretical estimates for many other crucial metrics (like delay) do not even exist and are similarly variable, dependent on parameters of the process design kit (PDK). To address these challenges, we present a study that employs intelligent search using multi-objective Bayesian optimization, integrating both neural network search and ASIC synthesis in the loop. This approach provides reliable feedback on the collective impact of all cross-domain design choices. We showcase the effectiveness of our approach by finding several Pareto-optimal design choices for effective and efficient neural networks that perform real-time feature extraction from input pulses within the individual pixels of a readout ASIC.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑