In this paper, we present a data-driven model for estimating optimal rework policies in manufacturing systems. We consider a single production stage within a multistage, lot-based system that allows for optional rework steps. While the rework decision depends on an intermediate state of the lot and system, the final product inspection, and thus the assessment of the actual yield, is delayed until production is complete. Repair steps are applied uniformly to the lot, potentially improving some of the individual items while degrading others. The challenge is thus to balance potential yield improvement with the rework costs incurred. Given the inherently causal nature of this decision problem, we propose a causal model to estimate yield improvement. We apply methods from causal machine learning, in particular double/debiased machine learning (DML) techniques, to estimate conditional treatment effects from data and derive policies for rework decisions. We validate our decision model using real-world data from opto-electronic semiconductor manufacturing, achieving a yield improvement of 2 - 3% during the color-conversion process of white light-emitting diodes (LEDs).

In manufacturing, rework refers to an optional step of a production process which aims to eliminate errors or remedy products that do not meet the desired quality standards. Reworking a production lot involves repeating a previous production stage with adjustments to ensure that the final product meets the required specifications. While offering the chance to improve the yield and thus increase the revenue of a production lot, a rework step also incurs additional costs. Additionally, the rework of parts that already meet the target specifications may damage them and decrease the yield. In this paper, we apply double/debiased machine learning (DML) to estimate the conditional treatment effect of a rework step during the color conversion process in opto-electronic semiconductor manufacturing on the final product yield. We utilize the implementation DoubleML to develop policies for the rework of components and estimate their value empirically. From our causal machine learning analysis we derive implications for the coating of monochromatic LEDs with conversion layers.

SiteAware, the pioneers of Digital Construction Verification (DCV) technology, announced today a $15 million Series B financing round led by Vertex Ventures, with participation from existing investors Bosch, Axon, Orizon, and lool Ventures. The company plans to use the funds to further scale its industry presence and to make Digital Construction Verification the new standard throughout the US construction industry. SiteAware was created to help contractors and developers overcome these challenges. The company's unique AI-powered DCV platform turns every construction element from the project plans into tagged data to create a digital twin of the building under construction. Each stage of new construction, be it the core, shell, or interior, is documented using drones, on-sitecameras, or people on the ground.

A group of more than 30 democratic lawmakers led by Representatives Tom Malinowski (D-NJ) and Anna G. Eshoo (D-CA) are calling on Facebook, Twitter and YouTube to make substantive changes to their recommendation algorithms. In three separate letters addressed to the CEOs of those companies, the group makes a direct link to the January 6th US Capitol attack and the part those platforms played in radicalizing the individuals who took part in the uprising. "On Wednesday, January 6th the United States Capitol was attacked by a violent, insurrectionist mob radicalized in part in a digital echo chamber that your company designed, built and maintained," the letter addressed to Google and YouTube CEOs Sundar Pichai and Susan Wojcicki says. A letter from some Congress members to Google CEO Sundar Pichai and YouTube CEO Susan Wojcicki flexes research on how YouTube's algorithms have promoted conspiracy theories and political extremism. Citing the Capitol attacks, they request changes to its recommendations systems.

From food to cars to complex manufacturing machinery, quality is a top concern of manufacturers. Factors such as safety, efficiency, and reliability affect product quality and ultimately influence customer satisfaction. Sourcing, design, testing, and inspection all play a crucial role in ensuring products meet the bar when it comes to quality. Product inspections at early stages in the production cycle help reduce risks and cost. While inspections can be conducted at any point throughout the production process, the goal is to identify, contain and resolve issues as quickly as possible.

Today's automobile development process is highly complicated. In the past, vehicles were just a combination of mechanical pieces. Now, automobiles are a multi-faceted combination of mechanical parts, electronics and in-vehicle software. As automotive designers, Honda R&D must continuously adapt our development processes to handle these complexities. Many design problems actually occur early in the development process, but don't become apparent until the latter stages.

In the Age of Big Data, humans generate 2.5 quintillion bytes of data each day. Looking at data related to land surveying for construction, the lack of automated information processing makes the activity expensive and tedious, costing clients an average range of $10,000 to $20,000 for a two-month project. David Morczinek and Adam Kernowski established Airworks, a Cambridge, M.A.-based data analytics startup that automates the processing of 20 million points of data from drone images to help decision-makers at land development and construction make sense of their aerial data. They are playing a more significant role in a variety of commercial applications such as aerial filming and package delivery. The global management consulting firm, McKinsey & Company, estimates that the commercial drone industry will have an economic impact ranging from $31 to $46 billion on U.S. GDP.

Oracle this week announced it will extend the reach of machine learning algorithms and related artificial intelligence (AI) applications into the realm of manufacturing. Ramchand Raman, vice president of product development for Oracle, says Oracle Adaptive Intelligent Applications for Manufacturing will employ machine learning and AI to process vast amounts of data to identify issues before they occur while also serving to improve overall operational efficiency. Oracle has already developed 150 key performance indicators (KPIs) for manufacturing firms that will serve as the baseline for which it will analyze manufacturing processes using machine learning algorithms and AI models. Over time, customers will be able to add their own KPIs, machine learning algorithms and AI models, says Raman. Specifically, Oracle Adaptive Intelligent Analytics for Manufacturing identifies patterns and correlations across variables such as manpower, machine, method, material and other management-related information.

Aerial robotics research has brought us flapping hummingbirds, seagulls, bumblebees, and dragonflies. But if these robots are to do anything more than bear a passing resemblance to their animal models, there is one thing they'll definitely need: better vision. In February, at the International Solid-State Circuits Conference (ISSCC) in San Francisco, two teams presented new work (PDF) aimed at building better-performing and lower-power vision systems that would help aerial robots navigate and aid them in identifying objects. Dongsuk Jeon, a graduate student working with Zhengya Zhang and IEEE Fellows David Blaauw and Dennis Sylvester at the University of Michigan, in Ann Arbor, outlined an approach to drastically lower the power of the very first stage of any vision system--the feature extractor. That system uses an algorithm to draw out potentially important features like circles and squares from an overall image.

In the former weblog publish, we learnt how to coach a convolutional neural network (CNN). 1 of the most preferred use instances for a CNN is to classify illustrations or photos. Once the CNN is properly trained, we need to have to know how to use it to classify an not known impression. The properly trained model files will be saved as "caffemodel" files, so we need to have to load those files, preprocess the input illustrations or photos, and then extract the output tags for those illustrations or photos. In this publish, we will see how to load those trained model files and use it to classify an impression. Let's go in advance see how to do it, shall we? Instruction a full network normally takes time, so we will use an existing properly trained model to classify an impression for now.