Making Manufacturing More Intelligent: AI and the World of Printed Circuit Boards

A tremendous opportunity is emerging in the PCB manufacturing industry whereby artificial intelligence (AI) can be utilized to streamline production processes and improve production outcomes in ways never before achievable. Arguably a market disruptor, the introduction of AI into PCB manufacturing is especially important as the market moves toward Industry 4.0 – or the “Smart Factory” – in which automation systems communicate and cooperate with each other and with humans in real-time, decentralizing the decision-making processes and offering numerous benefits. For AI implementations in the PCB domain to be successful, however, a rich combination of deep expertise in both PCB manufacturing and AI fundamentals is required – and this is where Orbotech comes in.

What is AI?

Over the last decade, AI has emerged from a futuristic concept to a genuinely applicable, relevant and effective technology. This breakthrough is largely due to increased and inexpensive computing power, more sophisticated learning algorithms, and the advent of big data, which itself provides the foundational information from which AI systems “learn.”

AI describes a machine or software that mimics human cognitive functions, such as problem solving and learning. Many important innovations are becoming subsets of AI, including machine learning and deep learning. Machine learning is a type of AI that uses various algorithmic techniques to enable computers to improve the performance of a task using data, but without being explicitly programmed to do so. In PCB manufacturing, machine learning enables a host of benefits including improved operations, reduced scrap rates, optimized fab operations, reduction of “non-sophisticated labor,” and the more efficient management of assets, inventory and the supply chain. All of these relevant features are the very definition of Industry 4.0.

Deep learning is an even more complex form of AI in which computers very effectively learn sophisticated representations of insights, patterns and context in data, all of which can be applied to process improvements in manufacturing. Deep learning utilizes multi-faceted, multi-layer artificial neural networks, which mimic the human brain’s ability to learn, understand and extrapolate.

What Orbotech Can Bring to AI for PCB Manufacturers

AI is exceedingly complex, requiring deep expertise in the technology itself, and in the industry domain in which the AI is to be used. As experts in both PCB manufacturing and AI, Orbotech is currently developing AI solutions for Industry 4.0 in PCB manufacturing.

To successfully implement an AI solution, process expertise must be matched by tight operational integration. An extensive and lengthy training phase is needed to build an AI system within the complex manufacturing environment. These training processes are very operationally demanding and require massive computational and human orchestration for optimal performance. Combined with Orbotech’s AI expertise, this ensures that Orbotech is able to optimize this critical training phase for AI’s neural network framework development.

Data is Key

In each variant of AI described above, access to data is a key component of success and efficacy. AI cannot be successful without very high-quality, labeled data. In PCB manufacturing, for example, defect classification is a crucial aspect of an automated optical inspection (AOI) solution. An Orbotech AOI machine sends images of potentially defective PCBs to a remote multi-image verification (RMIV) station where a human operator views the images and then classifies them as “True Defects” or “False Alarms.” A human operator is expensive, however, and can easily make mistakes in classification. An AI solution for defect classification would make these decisions autonomously and with consistent accuracy once it has learned from many thousands of decisions made by human operators. Such an AI system is dependent on accurate data patterns to learn correct algorithmic behavior over time. The input data – the examples of manually classified defects – must be verified and classified carefully to realize the benefits of AI. Even the smallest of changes in the data may lead to game-changing results from the AI system.

Success with AI through Collaboration

While quality data is essential for the success of AI, strong collaboration between each customer and its artificial intelligence solution provider, will also be crucial for long-term success. The complexities of AI demand that the solution provider apply its knowledge of PCB manufacturing processes and AI to guide customers in the implementation of AI solutions. For example, when classifying defects, Orbotech and its customers will collaborate to codify insights behind each human operator's classification evaluation results, and brainstorm together to create a meaningful system that prioritizes which PCB characteristics are the most important to detect. It is this level of customer collaboration and engagement that will yield the best results for an AI engine.

The Future

Given the significant benefits AI will inevitably provide to PCB manufacturers, Orbotech continues to engage heavily in R&D to perfect the integration of AI into Orbotech Smart Factory – the company’s Industry 4.0 solution – as well as its other solutions. In doing so, Orbotech is evaluating the many ways in which AI can be applied to manufacturing in general, learning through testing, and is helping to drive the industry towards smarter, more robust manufacturing processes that will help boost yields and reduce operational costs.

Author: Dima Pundik, Algorithm Group Manager, Orbotech Ltd.

Cautionary Statement Regarding Forward-Looking Statements

Except for historical information, the matters discussed in this website (including in press releases, webcasts, presentations, posts and other places) are forward-looking statements within the meaning of the U.S. Private Securities Litigation Reform Act of 1995. These statements relate to, among other things, future prospects, developments and business strategies and involve certain risks and uncertainties. The words “anticipate,” “believe,” “could,” “will,” “plan,” “expect” and “would” and similar terms and phrases, including references to assumptions, have been used in this website to identify forward-looking statements. These forward-looking statements are made based on management’s expectations and beliefs concerning future events as of the date of the applicable information (press releases, webcasts, presentations, or posts) and are subject to uncertainties and factors relating to Orbotech’s operations and business environment, the previously announced acquisition of Orbotech by KLA, the manner in which the parties plan to effect the transaction, including the share repurchase program, the ability to raise additional capital necessary to complete the repurchase program within the time frame expected, the expected benefits, synergies and costs of the transaction, management plans relating to the transaction, the expected timing of the completion of the transaction, the parties’ ability to complete the transaction considering the various closing conditions, including conditions related to regulatory and Orbotech shareholder approvals, the plans, strategies and objectives of management for future operations, product development, product extensions, product integration, complementary product offerings and growth opportunities in certain business areas, the potential future financial impact of the transaction, and any assumptions underlying any of the foregoing. Actual results may differ materially from those referred to in the forward-looking statements due to a number of important factors, including but not limited to the foregoing matters and the possibility that expected benefits of the transaction may not materialize as expected, that the transaction may not be timely completed, if at all, that KLA-Tencor may not be able to successfully integrate the solutions and employees of the two companies or ensure the continued performance or growth of Orbotech’s products or solutions, the risk that the Company may not achieve its revenue and margin expectations within and for 2018 (including, without limitation, due to shifting move-in dates); cyclicality in the industries in which the Company operates, the Company’s supply chain management and production capacity, order cancelation (often without penalty), timing and occurrence of product acceptance (the Company defines ‘bookings’ and ‘backlog’ as purchase arrangements with customers that are based on mutually agreed terms, which, in some cases for bookings and backlog, may still be subject to completion of written documentation and may be changed or cancelled by the customer, often without penalty), fluctuations in product mix within and among divisions, worldwide economic conditions generally, especially in the industries in which the Company operates, the timing and strength of product and service offerings by the Company and its competitors, changes in business or pricing strategies, changes in the prevailing political and regulatory framework in which the relevant parties operate, including as a result of the United Kingdom’s prospective withdrawal from the European Union (known as “Brexit”) and political uncertainty in the United States, or in economic or technological trends or conditions, including currency fluctuations, inflation and consumer confidence, on a global, regional or national basis, the level of consumer demand for sophisticated devices such as smart mobile devices, automotive electronics, flexible applications and devices, augmented reality/virtual reality and wearable devices, high-performance computing, liquid crystal display and organic light emitting diode screens and other sophisticated devices, the Company’s global operations and its ability to comply with varying legal, regulatory, exchange, tax and customs regimes, the timing and outcome of tax audits, including the best judgment tax assessment issued by the Israel Tax Authority with respect to the audit of tax years 2012-2014 in Israel and the related criminal investigation, the Company’s ability to achieve strategic initiatives, including related to its acquisition strategy, the Company’s debt and corporate financing activities; the timing, final outcome and impact of the criminal matter and ongoing investigation in Korea, including any impact on existing or future business opportunities in Korea and elsewhere, any civil actions related to the Korean matter brought by third parties, including the Company’s customers, which may result in monetary judgments or settlements, expenses associated with the Korean matter, and ongoing or increased hostilities in Israel and the surrounding areas.

The foregoing information should be read in connection with the Company’s Annual Report on Form 20-F for the year ended December 31, 2017, and subsequent SEC filings. This information is also supplemented by the information in the applicable document on this website (e.g., press release, webcast, presentation, posts and other document). The Company is subject to the foregoing and other risks detailed in those reports. The Company assumes no obligation to update the information in this website (including press releases, webcasts, presentations, posts and other places) to reflect new information, future events or otherwise, except as required by law.