Progress Inc11月13日推出神经形态技术PANC：以关联比较库完成识别与分类、公司2013年创立于密歇根州
超级智能研究与知识产权公司 Progress Inc. 于 2026 年 11 月 13 日在密歇根州西布卢姆菲尔德镇宣布，推出专有神经形态技术 Progress Associative Neuromorphic Comparator，简称 PANC。该公司是一家 2013 年在密歇根州创立的私营公司。
PANC 通过结构化的关联比较库，对图像、图表、表格、文本及其他形式的数据进行比较、识别、分类与分析。它不依赖深度神经网络与计算密集的训练过程，而是把知识组织进关联比较库，生成新数据与已存样本之间的相似系数。
公司称，这一架构可以完成识别、分类、聚类、泛化、近似、内插与外推等智能操作。建立识别库不需要深度学习，训练被缩减为两步：先把几乎任何文档或图形格式的信息转成其专有 BCF 平台格式，再把选定文件组合成有组织的比较库，已有库可以更新或扩展而不必重训整个系统。
公司强调的另一点是透明。比较过程、识别逻辑与分类路径都可供检查与分析，技术人员可以复查促成某一结果的信息与操作。由于其比较过程是确定性的而非概率性的，同一输入在重复运行与不同计算环境下给出可复现的结果，公司称这使该框架支持审计、人工监督与决策可追溯。
Progress 首席执行官 Vladimir Proseanic 表示，PANC 建立在比较是智能核心这一原则之上，其基础是一种新的人工神经元设计与比较器架构。公司称已在人与物体识别分类、医疗诊断、商业分析与决策支持、焊接缺陷检测与质量检验等场景评估过 PANC，该平台可部署在从嵌入式设备与边缘系统到企业服务器的标准计算硬件上。
（来源：Progress Inc）
Progress Inc. Launches Its Neuromorphic Technology PANC on November 13: It Recognizes and Classifies Through Associative Comparison Libraries, From a Company Founded in Michigan in 2013
Progress Inc., a private artificial intelligence research and intellectual property company founded in Michigan in 2013, announced in West Bloomfield Township, Michigan on November 13, 2026 the launch of the Progress Associative Neuromorphic Comparator, known as PANC, a proprietary neuromorphic technology designed to compare, recognize, classify and analyze images, charts, tables, text and other forms of data through structured associative comparison libraries. Unlike conventional artificial intelligence systems that depend on deep neural networks and computationally intensive training, PANC organizes knowledge in associative comparison libraries and generates similarity coefficients between new data and previously stored examples. The company says the architecture supports recognition, classification, clustering, generalization, approximation, interpolation and extrapolation. Building a recognition library does not require deep learning. The training process is reduced to two steps, converting information from virtually any document or graphical format into the proprietary BCF platform format, and then combining selected files into organized comparison libraries. Existing libraries can be updated or expanded without retraining the entire system. The company also stresses transparency. The comparison process, recognition logic and classification pathways are accessible for inspection and analysis, and because PANC relies on deterministic comparison processes rather than probabilistic model behavior, it produces reproducible results across repeated runs and different computing environments. The company says the framework supports auditing, human oversight and decision traceability. Vladimir Proseanic, chief executive of Progress Inc., said PANC is built on the principle that comparison lies at the heart of intelligence, and is founded on a new artificial neuron design and comparator architecture. The company says it has evaluated PANC in human and object recognition and classification, medical diagnostics, business analytics and decision support, weld defect detection and quality inspection, and that the platform is optimized for standard computing hardware from embedded devices and edge systems to enterprise servers.
(Source: Progress Inc)
来源：Progress Inc
发布：GSM·美国（发布链接）
图片来源：图：主板上的中央处理器芯片特写。Capitalistcakes，CC0，经 Wikimedia Commons。 / Photo: Capitalistcakes, CC0, via Wikimedia Commons.
GSM Global · 本篇英文版 English edition of this story
