Distribution and democratization represent two complementary paradigms that are gaining increasing attention in manufacturing. Distributed manufacturing (DM) allows for geographically dispersed production, often at small scales and near the end user. Democratization enables large populations to engage in manufacturing. Massively distributed manufacturing (MDM), which combines these paradigms, is performed on demand by a large network of people located anywhere. Rather than rely on mass production in centralized factories, MDM promises to improve the responsiveness and resilience of manufacturing to urgent production demands (such as emergencies like pandemics); promote mass customization and cost-effective, low-volume production; gainfully employ many informally trained citizens in manufacturing (such as through the gig economy); and reduce the environmental footprint of manufacturing by producing items near their points of use. The Fourth Industrial Revolution will play an important role in enabling MDM by way of cyber-physical operating systems (CPOSs). From the First Industrial Revolution in the 18th century onward, manufacturing has been carried out predominantly through mass production in centralized factories, often far from the end user. Mass production enables large quantities of products to be produced with standardized quality, high productivity, and low cost. However, in the face of urgent demands or disruptions, it lacks flexibility, agility, and resilience and cannot readily provide consumers with personalized products in small quantities (mass customization) ([ 1 ][1]). Moreover, its environmental footprint is large, mainly because it often requires raw materials and finished goods to be transported over long distances. During the past decade, there has been growing interest and activity in distributed and democratized manufacturing as alternative or complementary paradigms to mass production. DM has been emphasized by the United Nations International Development Organization ([ 2 ][2]), the World Economic Forum ([ 3 ][3]), and other major agencies ([ 4 ][4], [ 5 ][5]) as critical to the future of manufacturing. Several companies engaged in DM, such as 3D Hubs, 3Diligent, Fast Radius, and Xometry, have sprouted. Xometry, for example, enables its customers to access the manufacturing capacity of a network of >5000 carefully curated partners—typically small- and medium-sized enterprises—distributed across the world. In terms of democratization ([ 6 ][6]), perhaps the most compelling example is the proliferation of desktop three-dimensional (3D) printers, which currently retail on average for ∼$1000 ([ 7 ][7]), which is within the purchasing power of large portions of the population. In 2019, >700,000 desktop 3D printers were sold globally ([ 7 ][7]). These printers can now be found in homes, offices, schools, maker spaces, public libraries, and other facilities, and people can use them for prototyping and small-scale or micromanufacturing without extensive technical training. ![Figure][8] Networked systems linking producers to consumers Massively distributed manufacturing uses a cyber-physical operating system and artificial intelligence tools to connect and coordinate consumers with producers. Producers in micromanufacturing units can use three-dimensional (3D) printing to fabricate customized products. Smart logistics such as drones and rideshare services enable the physical product delivery. GRAPHIC: C. BICKEL/ SCIENCE However, distributed and democratized manufacturing are still far from the goal of MDM ([ 8 ][9]–[ 10 ][10]), in which products are manufactured by a large, diverse, and geographically dispersed but coordinated network of individuals and organizations with agility and flexibility, but with near–mass-production quality, productivity, and cost effectiveness. For example, a company like Xometry would need to engage millions of users in micromanufacturing across the globe, similar to what companies like Uber and Lyft have achieved with transportation. The latent potential of MDM was evident during the early days of the COVID-19 pandemic, when personal protective equipment (PPE) were in short supply. Mass production was too slow to react to the sudden demands for PPE, including demands for simple but vital plastic products like face shields. Worldwide, thousands, if not millions, of people, many of whom did not have experience in making these products, organized themselves into small networks to produce millions of face shields and other PPE using desktop 3D printers and other small-scale manufacturing equipment ([ 11 ][11]). This effort exposed key challenges of MDM in terms of standardizing production requirements, guaranteeing quality and reliability, and attaining high production efficiencies that can rival those of mass production. This example illustrates the important role of technology in enabling MDM. The First, Second, and Third Industrial Revolutions, driven by mechanization, electrification plus assembly lines, and digital computing, respectively, paved the way for the Fourth Industrial Revolution (or Industry 4.0), undergirded by networked cyber-physical systems and artificial intelligence. For example, Xometry leverages cloud computing and machine learning to power its instant quoting engine that enables customers to receive pricing, expected lead times, and manufacturability feedback within seconds. Similarly, 3Diligent uses cloud computing to enable manufacturers in its network to route jobs across their shop floors and track quality. With advances in Industry 4.0, manufacturing machines (including low-cost 3D printers) are increasingly equipped with sensors and cloud connectivity ([ 12 ][12]). The large amounts of data generated by these sensors are being used in machine-learning algorithms to provide predictive and corrective actions ([ 13 ][13]). Advanced cloud-based controllers are being developed to improve the quality and productivity of the machines ([ 14 ][14]). These advances in technology and automation can converge into a cloud-based CPOS for MDM. An inspiration for CPOSs is the central coordinator used in distributed computing to automate the allocation and execution of large-scale computing tasks on distributed networks of computers. The central coordinator has enabled Folding at Home ([ 15 ][15]), a distributed computing cluster that leverages the idle capacity of >100,000 personal computers to run simulations that help scientists to understand how proteins fold. Similarly, a CPOS will intelligently, efficiently, and securely coordinate large networks of cloud-connected, autonomous, and geographically-dispersed manufacturing resources. It will optimally allocate manufacturing jobs to the resources connected to it and leverage distributed and democratized delivery systems, such as shared vehicles and drones, for logistics (see the figure). It will apply machine learning to the data gathered from sensors to help assure and improve quality and to optimize operations. Furthermore, CPOSs will leverage the ingenuity of humans through the crowdsourcing of ideas to improve manufacturing operations across networks of manufacturers as well as cybersecurity measures to protect intellectual property and the privacy of participants. CPOSs will thus allow the collaboration of large, autonomous, heterogeneous, and geographically dispersed networks of manufacturers to rapidly respond to production demands and disruptions with agility and flexibility, while ensuring the high quality, productivity, and cost effectiveness of MDM. 1. [↵][16]1. B. J. Pine II , Mass Customization: The New Frontier in Business Competition (Harvard Business School Press, 1993). 2. [↵][17]1. C. López-Gómez, 2. E. O'Sullivan, 3. M. Gregory, 4. A. C. C. Fleury, 5. L. Gomes , Emerging Trends in Global Manufacturing Industries (United Nations Industrial Development Organization, 2013). 3. [↵][18]1. B. Meyerson , Top 10 Emerging Technologies of 2015 (World Economic Forum, 2015). 4. [↵][19]Foresight, The Future of Manufacturing: A New Era of Opportunity and Challenge for the UK Summary Report (The Government Office for Science, London, UK, 2013). 5. [↵][20]European Factories of the Future Research Association, Factories of the Future: Multi-annual Roadmap for the Contractual PPP under Horizon 2020 (European Commission, 2013). 6. [↵][21]MForesight, Democratizing Manufacturing: How to Realize the Promise of the Maker Movement (2017). 7. [↵][22]Wohlers Associates, Wohlers Report 2020: 3D Printing and Additive Manufacturing: Global State of the Industry (Wohlers Associates, 2020); . 8. [↵][23]1. J. S. Srai et al ., Int. J. Prod. Res. 54, 6917 (2016). [OpenUrl][24] 9. 1. H. Stewart, 2. J. Tooze , Making Futures 4, 1 (2015). [OpenUrl][25] 10. [↵][26]1. P. Jiang, 2. J. Leng, 3. K. Ding, 4. P. Gu, 5. Y. Koren , Proc. Inst. Mech. Eng. B 230, 1961 (2016). [OpenUrl][27] 11. [↵][28]1. J. M. Pearce , J. Manuf. Mater. Process. 4, 49 (2020). [OpenUrl][29] 12. [↵][30]1. C. E. Okwudire, 2. S. Huggi, 3. S. Supe, 4. C. Huang, 5. B. Zeng , Inventions 3, 56 (2018). [OpenUrl][31] 13. [↵][32]1. T. Wuest, 2. D. Weimer, 3. C. Irgens, 4. K. D. Thoben , Prod. Manuf. Res. 4, 23 (2016). [OpenUrl][33] 14. [↵][34]1. C. E. Okwudire, 2. X. Lu, 3. G. Kumaravelu, 4. H. Madhyastha , Robot. Comput.-Integr. Manuf. 62, 101880 (2020). [OpenUrl][35] 15. [↵][36]Folding at Home, . Acknowledgments: C.E.O. is a founder of Ulendo, which has licensed research in advanced cloud-based 3D printer control algorithms. [1]: #ref-1 [2]: #ref-2 [3]: #ref-3 [4]: #ref-4 [5]: #ref-5 [6]: #ref-6 [7]: #ref-7 [8]: pending:yes [9]: #ref-8 [10]: #ref-10 [11]: #ref-11 [12]: #ref-12 [13]: #ref-13 [14]: #ref-14 [15]: #ref-15 [16]: #xref-ref-1-1 "View reference 1 in text" [17]: #xref-ref-2-1 "View reference 2 in text" [18]: #xref-ref-3-1 "View reference 3 in text" [19]: #xref-ref-4-1 "View reference 4 in text" [20]: #xref-ref-5-1 "View reference 5 in text" [21]: #xref-ref-6-1 "View reference 6 in text" [22]: #xref-ref-7-1 "View reference 7 in text" [23]: #xref-ref-8-1 "View reference 8 in text" [24]: {openurl}?query=rft.jtitle%253DInt.%2BJ.%2BProd.%2BRes.%26rft.volume%253D54%26rft.spage%253D6917%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [25]: {openurl}?query=rft.jtitle%253DMaking%2BFutures%26rft.volume%253D4%26rft.spage%253D1%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [26]: #xref-ref-10-1 "View reference 10 in text" [27]: {openurl}?query=rft.jtitle%253DProc.%2BInst.%2BMech.%2BEng.%2BB%26rft.volume%253D230%26rft.spage%253D1961%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [28]: #xref-ref-11-1 "View reference 11 in text" [29]: {openurl}?query=rft.jtitle%253DJ.%2BManuf.%2BMater.%2BProcess.%26rft.volume%253D4%26rft.spage%253D49%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [30]: #xref-ref-12-1 "View reference 12 in text" [31]: {openurl}?query=rft.jtitle%253DInventions%26rft.volume%253D3%26rft.spage%253D56%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [32]: #xref-ref-13-1 "View reference 13 in text" [33]: {openurl}?query=rft.jtitle%253DProd.%2BManuf.%2BRes.%26rft.volume%253D4%26rft.spage%253D23%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [34]: #xref-ref-14-1 "View reference 14 in text" [35]: {openurl}?query=rft.jtitle%253DRobot.%2BComput.-Integr.%2BManuf.%26rft.volume%253D62%26rft.spage%253D101880%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [36]: #xref-ref-15-1 "View reference 15 in text"

In the paper, we show how scalable, low-cost trunk-like robotic arms can be constructed using only basic 3D-printing equipment and simple electronics. The design is based on uniform, stackable joint modules with three degrees of freedom each. Moreover, we present an approach for controlling these robots with recurrent spiking neural networks. At first, a spiking forward model learns motor-pose correlations from movement observations. After training, intentions can be projected back through unrolled spike trains of the forward model essentially routing the intention-driven motor gradients towards the respective joints, which unfolds goal-direction navigation. We demonstrate that spiking neural networks can thus effectively control trunk-like robotic arms with up to 75 articulated degrees of freedom with near millimeter accuracy.

Agtech capabilities are bringing traditional farming into the 21st century. These solutions range from sprawling LED-equipped indoor farming facilities to robotically plucking ripe produce off the vine using computer vision and artificial intelligence (AI). On Thursday, John Deere and Intel announced a pilot program that relies on AI and computer vision to detect defects in manufacturing related to the welding process. "Welding is a complicated process. This AI solution has the potential to help us produce our high-quality machines more efficiently than before," said Andy Benko, quality director at John Deere Construction and Forestry Division.

When it comes to embracing new technology and digitising entire sectors of business, look no further than the Industry 4.0 revolution. All over the world, companies from the likes of manufacturing, warehousing and logistics are embracing Industry 4.0's key technologies to open up new values and benefits. However, Industry 4.0 trends shift and evolve as time goes on. As such, we want to take a look at Industry 4.0 technologies and projects in more detail. We've previously discussed the topic of Industry 4.0 before, but here's a quick recap.

Wire-feed laser additive manufacturing (WLAM) is gaining wide interest due to its high level of automation, high deposition rates, and good quality of printed parts. In-process monitoring and feedback controls that would reduce the uncertainty in the quality of the material are in the early stages of development. Machine learning promises the ability to accelerate the adoption of new processes and property design in additive manufacturing by making process-structure-property connections between process setting inputs and material quality outcomes. The molten pool dimensional information and temperature are the indicators for achieving the high quality of the build, which can be directly controlled by processing parameters. For the purpose of in situ quality control, the process parameters should be controlled in real-time based on sensed information from the process, in particular the molten pool. Thus, the molten pool-process relations are of preliminary importance. This paper analyzes experimentally collected in situ sensing data from the molten pool under a set of controlled process parameters in a WLAM system. The variations in the steady-state and transient state of the molten pool are presented with respect to the change of independent process parameters. A multi-modality convolutional neural network (CNN) architecture is proposed for predicting the control parameter directly from the measurable molten pool sensor data for achieving desired geometric and microstructural properties. Dropout and regularization are applied to the CNN architecture to avoid the problem of overfitting. The results highlighted that the multi-modal CNN, which receives temperature profile as an external feature to the features extracted from the image data, has improved prediction performance compared to the image-based uni-modality CNN approach.

Founded in 2013, Markforged is the creator of an integrated metal and carbon fiber additive manufacturing platform, The Digital Forge - a cloud and ML-based 3D printing platform designed to interconnect all of the company's systems currently being used around the world. It is claimed to be the first such platform to use machine learning, a feature that enables the company's Eiger print preparation software to constantly learn from the 12,000 systems in its 73-country-wide global fleet. As such, says the company, every print on a connected Markforged system should theoretically be more accurate than the last. A* was founded and is led by technology industry veteran and investor Kevin Hartz. The combined company will have an estimated post-transaction equity value of approximately $2.1 billion at closing.

In recent years, the'maker movement' has emerged as a social phenomenon driven by novel technological possibilities.1 With the help of inexpensive, yet highly versatile means of production (for example, CNC milling machines, 3D printers) and easy-to-use software tools, makers free themselves from their traditional role as passive consumers and evolve into innovators and producers. Although the act of physical production seems to be at the center of the movement, a large part of the creative work takes place in the online sphere. These digital activities and their outcomes provide a rich source of information that can be used to gain a more nuanced understanding of how the digitization affects the creative process itself. Of all the production methods available to makers, 3D printing is probably the most versatile and requires only a limited understanding of the production process. Several 3D design software packages allow even lay people to turn their ideas into printable designs.

From Star Trek's replicators to Richie Rich's wishing machine, popular culture has a long history of parading flashy machines that can instantly output any item to a user's delight. While 3D printers have now made it possible to produce a range of objects that include product models, jewelry, and novelty toys, we still lack the ability to fabricate more complex devices that are essentially ready-to-go right out of the printer. A group from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) recently developed a new system to print functional, custom-made devices and robots, without human intervention. Their single system uses a three-ingredient recipe that lets users create structural geometry, print traces, and assemble electronic components like sensors and actuators. "LaserFactory" has two parts that work in harmony: a software toolkit that allows users to design custom devices, and a hardware platform that fabricates them.

Deep neural networks have revolutionized many machine learning tasks in power systems, ranging from pattern recognition to signal processing. The data in these tasks is typically represented in Euclidean domains. Nevertheless, there is an increasing number of applications in power systems, where data are collected from non-Euclidean domains and represented as the graph-structured data with high dimensional features and interdependency among nodes. The complexity of graph-structured data has brought significant challenges to the existing deep neural networks defined in Euclidean domains. Recently, many studies on extending deep neural networks for graph-structured data in power systems have emerged. In this paper, a comprehensive overview of graph neural networks (GNNs) in power systems is proposed. Specifically, several classical paradigms of GNNs structures (e.g., graph convolutional networks, graph recurrent neural networks, graph attention networks, graph generative networks, spatial-temporal graph convolutional networks, and hybrid forms of GNNs) are summarized, and key applications in power systems such as fault diagnosis, power prediction, power flow calculation, and data generation are reviewed in detail. Furthermore, main issues and some research trends about the applications of GNNs in power systems are discussed.

We propose a new adversarial training approach for injecting robustness when designing controllers for upcoming cyber-physical power systems. Previous approaches relying deeply on simulations are not able to cope with the rising complexity and are too costly when used online in terms of computation budget. In comparison, our method proves to be computationally efficient online while displaying useful robustness properties. To do so we model an adversarial framework, propose the implementation of a fixed opponent policy and test it on a L2RPN (Learning to Run a Power Network) environment. That environment is a synthetic but realistic modeling of a cyber-physical system accounting for one third of the IEEE 118 grid. Using adversarial testing, we analyze the results of submitted trained agents from the robustness track of the L2RPN competition. We then further assess the performance of those agents in regards to the continuous N-1 problem through tailored evaluation metrics. We discover that some agents trained in an adversarial way demonstrate interesting preventive behaviors in that regard, which we discuss.