Artisan robots with AI smarts will juggle tasks, choose tools, mix and match recipes and even order materials – all without human help
Failure of a machine in a factory can shut it down. Lost production can cost millions of dollars per day. Component failures can devastate factories, power plants and battlefield equipment
To return to operation, skilled technicians use all the tools in their kit - machining, bending, welding and surface treating, making just the right part as quickly and as accurately as possible. But there’s a declining number of technicians with the right skills, and the quality of things made by hand is subject to the skills and mood of the artisan on the day the part is made.
Both problems could soon be solved by artificially intelligent robotic technicians. These systems can take measurements; shape, cut or weld parts using varied tools; pass parts to specialized equipment; and even purchase needed materials – all without human intervention. Known as hybrid autonomous manufacturing, this process involves automated systems that seamlessly use multiple tools and techniques to build high-quality components where and when they are needed.
I am a professor of metallurgical engineering. My colleagues and I design the recipes to make materials and components with just the right internal structure to create properties like strength and fracture resistance. With a network of colleagues at Ohio State and other universities, I have been developing a plan to give birth to these autonomous artisans.
How things are made
Components are either mass-produced or custom-made.
Most things people touch daily have been mass-produced. Quality is assured by using well-honed processes based on testing and monitoring large numbers of parts and assuring the process is done the same way every time.
Custom fabrication – making components on demand – is often essential, sometimes to conform to a patient’s specific anatomy or to replace aircraft landing gear that was forged and is no longer being made. Processes for making metallic parts – material removal, deposition, deformation, transformation, inspection – can all be done with small tools, with incremental actions rather than the kind of bulk processes, usually with big tools and dies, used in mass production.
Automation has long been a part of mass production, which includes sophisticated robots that handle parts and weld on automobile assembly lines. Additive manufacturing often referred to as 3D printing, is increasingly being used with a variety of materials to make components.
Now in development are robotic blacksmiths – robots that can hammer metallic parts into shape instead of cutting, building up or moulding them.
Automated customization – not an oxymoron
To automate custom fabrication, my colleagues and I are developing an automated suite of tools that can carry out all the steps for making a wide range of components, using multiple processes without human intervention. Sensors will also be central to hybrid autonomous manufacturing to control the processes and maintain and assure quality.
Such autonomous manufacturing systems will make the myriad decisions needed to create a component of the right strength, size and surface finish. Artificial intelligence will be required to handle the enormous number of choices of materials, machine settings and process sequences. Rather than finding a mass production recipe and never deviating, these autonomous manufacturing systems will choose from a very large set of possible recipes to create parts and will have the intelligence to assure that the chosen path produces components with the appropriate material properties.
Robots could either position small tools on manufactured components or transfer the component from one piece of equipment to another. A fully autonomous system could manufacture a wide range of products with a versatile set of tools. The systems could source materials and possibly even send work out to specialized cutting and deformation tools, just like a human artisan.
The production rate of such systems would not rival those of mass production, but because robots can work continuously they can be more productive than human technicians are. Data from sensors provide a digital record of all the steps and processes with critical temperatures, machine settings and even images. This record can assure quality by, for example, making sure the material was deformed the right amount and cracks were not produced during the process and covered up.
Manufacturing at or near the operating room is one example of a process that can be enabled with hybrid autonomous manufacturing. Often when patients with bone fractures undergo trauma surgery, metallic plates of varied shapes are required to hold bones together for healing. These are often created in the operating room, where the surgeon bends plates to fit the patient, sometimes using a 3D-printed model created from medical images of the patient as a form to bend the metal against.
Bending by hand is slow and imprecise, and stressing the plate in the wrong place can cause it to fracture. A robotic technician could cut and bend and finish a plate before surgery. Patients do better and save money if they spend less time in the hospital.
The road to robotic artisans
Numerous companies are now showing the way forward in autonomous manufacturing, including three venture-funded startups. FormLogic is developing automated high-quality machine shops. Path Robotics is putting the skills of a welder into a robot. And Machina Labs is out to create robotic blacksmiths. Other companies are developing systems to automate design and logistics.
Hybridization – the ability to carry out different tasks in different ways with multiple tools – is the next step. The key pieces of hybrid autonomous manufacturing exist now, and fully autonomous systems could be common in a decade. Companies adopting this approach to custom fabrication will need to draw on a new generation of students with the skills to combine these technologies.
The investments proposed in the United States Innovation and Competition Act passed by the Senate on June 8, 2021, and those in the Biden administration’s proposed American Jobs Plan could support the development of these kinds of advanced manufacturing technologies. Funds for the development of advanced manufacturing technologies and the associated skills base could make U.S. manufacturing more competitive.
Glenn S. Daehn, Professor of Materials Science and Engineering, The Ohio State University
This article is republished from The Conversation under a Creative Commons license. Read the original article.
AI Marketplace: the digital platform for tomorrow’s innovations
Challenges in product creation can be solved with the help of artificial intelligence
Posting challenges, finding solutions, and gathering know-how: this is what the AI Marketplace promises. For the past year, experts from science and industry have been working on a platform for AI in product development. The platform is now live in a beta version and brings together AI experts, suppliers, and users.
Whether it's automating technology scouting or optimising design data, the potential of AI in product creation is manifold. CLAAS, an internationally active agricultural machinery manufacturer, has recognised these possibilities.
The company is testing a special use case for the integration of AI in computer-aided design in the AI Marketplace, an initiative of the leading-edge cluster It’s OWL. The project is one of six pilot projects in which companies are working together with research institutions in the AI Marketplace on AI solutions for specific use cases.
The topics range from intelligent product monitoring and smart vehicle diagnosis to AI-supported manufacturability analysis. In addition to CLAAS, companies including Diebold Nixdorf, Düspohl Maschinenbau, Hella Gutmann Solutions, Westaflex and Ubermetrics Technologies are also involved in the pilot projects.
However, countless other partners will benefit from the AI Marketplace. After all, providers, users, and experts can jointly develop and exchange AI solutions on the platform of the same name.
“The AI Marketplace offers companies a central place to solve their product development challenges with the help of AI,” says Leon Özcan, Project Coordinator and Research Associate at Heinz Nixdorf Institute (Germany), part of Paderborn University. “AI providers gain direct access to their customers in this way.”
Sneak peek at the AI Marketplace platform: Suppliers, users and experts can develop and exchange AI solutions. Image provided courtesy of It’s OWL.
What sets the AI Marketplace apart
The AI Marketplace platform will be successively expanded by functionalities in the course of the project. Initially, users will be offered a matching function: for example, companies can post challenges while AI providers create competence profiles.
A subsequent matching of challenges with competence profiles brings together suitable platform actors, who can now jointly develop solutions.
“Our competitive analysis has shown that our concept of a marketplace for AI in product creation addresses a blank spot on the competitive map,” says Ruslan Bernijazov, Technical Project Coordinator and Research Associate at Fraunhofer Institute for Mechatronic Systems Design IEM. “Although there are numerous solutions that address concrete problems from product creation, there is currently no central point of contact for solution providers, AI experts, and users.”
In addition to matching, the next step is to create a protected data room for development and test data in order to continuously improve AI applications and adapt them to customer needs.
Here, the AI Marketplace relies on certified standards that guarantee data sovereignty while forming the basis for fair cooperation. An app store for AI solutions is also to be added, as well as a range of standardised AI building blocks that can be combined as required and used to develop new AI solutions.
More than a platform
The AI Marketplace already offers free services with which small and medium-sized enterprises in particular can tap into the potential of artificial intelligence in product creation.
By means of a Potential Analysis, structured application possibilities for AI in companies can be identified. In addition to concrete AI use cases, companies also receive recommendations for action for implementation. Using these, manufacturing companies can optimise their development capacities and shorten development cycles.
Companies can find out whether they are ready for the application of artificial intelligence in the AI Readiness Check on the AI Marketplace. In addition to determining the maturity level, companies receive helpful tips here on how to best prepare for the use of AI in product development.
In addition, companies can have their current corporate data governance assessed by the AI Marketplace with the help of a Data Check-Up. As a result, companies immediately receive a classification of their maturity level as well as concrete tips and information on how they can improve their corporate data governance so that nothing stands in the way of the use of AI applications.
The AI Marketplace creates a unique ecosystem that brings together AI experts, vendors, and users to realise the potential of AI. Image Source: Adobe Stock
A large network is behind the AI Marketplace
The success of the platform and its services is guaranteed by a consortium of 19 research institutions, networks, and companies, whose nucleus is It’s OWL. In this technology network, 200 companies and research institutions are jointly developing new technologies for tomorrow’s production.
Other networks also ensure a broader impact. prostep ivip Association, for example, bundles know-how in product development, while the International Data Spaces Association (IDSA) – both members of FIWARE Foundation – ensures secure data spaces.
Furthermore, the open-source platform FIWARE and the platform operator Inno-Focus are leaders in their fields. Artificial Intelligence and Machine Learning (ML) techniques are increasingly being incorporated in architectures powered by FIWARE to support the adoption of smarter decisions, going a step further in the automation of processes.
The AI Marketplace project is funded until December 2022 by the innovation competition “Artificial Intelligence as a Driver for Economically Relevant Ecosystems” of the German Federal Ministry for Economic Affairs and Energy (BMWi).
To learn more about our work and how you can get involved, visit the AI Marketplace webpage.
Hendrik Fahrenwald is a Marketing Manager at It’s OWL.
Header image courtesy of It’s OWL.
INDUSTRY VIEW FROM THE AI MARKETPLACE
‘Can I see your parts list?’ What AI’s attempted chat-up lines tell us about computer-generated language
Have you ever wondered what flirting with artificial intelligence would look like? Research scientist and engineer Janelle Shane have given us an idea by training a neural network – an algorithm loosely inspired by biological brain structures – to produce chat-up lines.
Some of the results are hilarious and completely nonsensical, such as the inelegant: “2017 Rugboat 2-tone Neck Tie Shirt”. But some of them turned out pretty well. At least, if you’re a robot:
I can tell by your red power light that you’re into me. You look like a thing and I love you. Can I see your parts list?
But how were these lines generated, and why do the results vary so much in terms of quality and cohesiveness? That’s down to the types of neural networks Shane worked with: all based on GPT-3, the world’s largest language model to date.
Language modelling
GPT stands for generative pre-trained transformer. Its current version, developed by OpenAI, is the third in a line of ever-improving natural language processing systems trained to produce human-like text or speech.
Natural language processing, or NLP, refers to the application of computers to process and generate large amounts of coherent spoken or written text. Whether you ask Siri for a weather update, request for Alexa to turn on the lights, or you use Google to translate a message from French into English, you’re able to do so because of developments in NLP.
It takes a variety of NLP tasks – from speech recognition to picking apart sentence structures – for applications such as Siri to successfully requests. The virtual assistant, much like any other language-based tool, is trained using many thousands of sentences, ideally as varied and diverse as possible.
Because human language is extremely complex, the best NLP applications rely increasingly on pre-trained models that allow “contextual bidirectional learning”. This means considering a word’s wider context in a sentence, scanning both left and right of any given word to identify the word’s intended meaning. More recent models can even pay attention to more nuanced features of human language, such as irony and sarcasm.
Computer compliments
GPT-3 is such a successful language-generating AI because it doesn’t need retraining over and over again to complete a new task. Instead, it uses what the model has already learned about language and applies it to something new – such as writing articles and computer code, generating novel dialogue in video games, or formulating chat-up lines.
Compared to its predecessor GPT-2, the third-generation model is 116 times bigger and has been trained on billions of words of data. To generate its chat-up lines, GPT-3 was simply asked to automate the text for an article headlined: “These are the top pickup lines of 2021! Amaze your crush and get results!”
Because GPT-3’s training updates have been added gradually over time, this same prompt could also be used on smaller, more basic variants – generating weirder and less coherent chat-up lines:
Hey, my name is John Smith. Will you sit on my breadbox while I cook or is there some kind of speed limit on that thing? It is urgent that you become a professional athlete. CAPE FASHION
But GPT-3’s “DaVinci” variant – its largest and most competent iteration to date – delivered some more convincing attempts which might actually pass for effective flirting – with a little fine-tuning:
You have the most beautiful fangs I’ve ever seen. I love you. I don’t care if you’re a doggo in a trenchcoat. I have exactly 4 stickers. I need you to be the 5th.
The latest variant of GPT-3 is currently the largest contextual language model in the world and is able to complete a number of highly impressive tasks. But is it smart enough to pass as a human?
Almost human
As one of the pioneers of modern computing and a firm believer in true artificial intelligence, Alan Turing developed the “Imitation Game” in 1950 – today is known as the “Turing Test”. If a computer’s performance is indistinguishable from that of a human, it passes the Turing Test. In language generation alone, GPT-3 could soon pass Alan Turing’s test.
But it doesn’t really matter if GPT-3 passes the Turing Test or not. Its performance is likely to depend on the specific task the model is used for – which, judging by the technology’s flirting, should probably be something other than the delicate art of the chat-up line.
And, even if it were to pass the Turing Test, in no way would this make the model truly intelligent. At best, it would be extremely well trained on specific semantic tasks. Maybe the more important question to ask is: do we even want to make GPT-3 more human?
Learning from humans
Shortly after its reveal in summer 2020, GPT-3 made headlines for spewing out shockingly sexist and racist content. But this was hardly surprising. The language generator was trained on vast amounts of text on the internet, and without remodelling and retraining, it was doomed to replicate the biases, harmful language and misinformation that we know to exist online.
Clearly, language models such as GPT-3 do not come without potential risks. If we want these systems to be the basis of our digital assistants or conversational agents, we need to be more rigorous and selective when giving them reading material to learn from.
Still, recent research has shown that GPT-3’s knowledge of the internet’s dark side could actually be used to automatically detect online hate speech, with up to 78% accuracy. So even though its chat-up lines look unlikely to kindle more love in the world, GPT-3 could be set, at least, to reduce the hate.
Stefanie Ullmann, Postdoctoral Research Associate, Centre for the Humanities and Social Change, University of Cambridge
This article is republished from The Conversation under a Creative Commons license. Read the original article.
AI is increasingly being used to identify emotions – here's what's at stake
Imagine you are in a job interview. As you answer the recruiter’s questions, an artificial intelligence (AI) system scans your face, scoring you for nervousness, empathy and dependability. It may sound like science fiction, but these systems are increasingly used, often without people’s knowledge or consent.
Emotion recognition technology (ERT) is in fact a burgeoning multi-billion-dollar industry that aims to use AI to detect emotions from facial expressions. Yet the science behind emotion recognition systems is controversial: there are biases built into the systems.
Many companies use ERT to test customer reactions to their products, from cereal to video games. But it can also be used in situations with much higher stakes, such as in hiring, by airport security to flag faces as revealing deception or fear, in border control, in policing to identify “dangerous people” or in education to monitor students’ engagement with their homework.
Shaky scientific ground
Fortunately, facial recognition technology is receiving public attention. The award-winning film Coded Bias, recently released on Netflix, documents the discovery that many facial recognition technologies do not accurately detect darker-skinned faces. And the research team managing ImageNet, one of the largest and most important datasets used to train facial recognition, was recently forced to blur 1.5 million images in response to privacy concerns.
Revelations about algorithmic bias and discriminatory datasets in facial recognition technology have led large technology companies, including Microsoft, Amazon and IBM, to halt sales. And the technology faces legal challenges regarding its use in policing in the UK. In the EU, a coalition of more than 40 civil society organisations have called for a ban on facial recognition technology entirely.
Like other forms of facial recognition, ERT raises questions about bias, privacy and mass surveillance. But ERT raises another concern: the science of emotion behind it is controversial. Most ERT is based on the theory of “basic emotions” which holds that emotions are biologically hard-wired and expressed in the same way by people everywhere.
This is increasingly being challenged, however. Research in anthropology shows that emotions are expressed differently across cultures and societies. In 2019, the Association for Psychological Science conducted a review of the evidence, concluding that there is no scientific support for the common assumption that a person’s emotional state can be readily inferred from their facial movements. In short, ERT is built on shaky scientific ground.
Also, like other forms of facial recognition technology, ERT is encoded with racial bias. A study has shown that systems consistently read black people’s faces as angrier than white people’s faces, regardless of the person’s expression. Although the study of racial bias in ERT is small, racial bias in other forms of facial recognition is well-documented.
There are two ways that this technology can hurt people, says AI researcher Deborah Raji in an interview with MIT Technology Review: “One way is by not working: by virtue of having higher error rates for people of color, it puts them at greater risk. The second situation is when it does work — where you have the perfect facial recognition system, but it’s easily weaponized against communities to harass them.”
So even if facial recognition technology can be de-biased and accurate for all people, it still may not be fair or just. We see these disparate effects when facial recognition technology is used in policing and judicial systems that are already discriminatory and harmful to people of colour. Technologies can be dangerous when they don’t work as they should. And they can also be dangerous when they work perfectly in an imperfect world.
The challenges raised by facial recognition technologies – including ERT – do not have easy or clear answers. Solving the problems presented by ERT requires moving from AI ethics centred on abstract principles to AI ethics centred on practice and effects on people’s lives.
When it comes to ERT, we need to collectively examine the controversial science of emotion built into these systems and analyse their potential for racial bias. And we need to ask ourselves: even if ERT could be engineered to accurately read everyone’s inner feelings, do we want such intimate surveillance in our lives? These are questions that require everyone’s deliberation, input and action.
Citizen science project
ERT has the potential to affect the lives of millions of people, yet there has been little public deliberation about how – and if – it should be used. This is why we have developed a citizen science project.
On our interactive website (which works best on a laptop, not a phone) you can try out a private and secure ERT for yourself, to see how it scans your face and interprets your emotions. You can also play games comparing human versus AI skills in emotion recognition and learn about the controversial science of emotion behind ERT.
Most importantly, you can contribute your perspectives and ideas to generate new knowledge about the potential impacts of ERT. As the computer scientist and digital activist Joy Buolamwini says: “If you have a face, you have a place in the conversation.”
Alexa Hagerty, Research Associate of Anthropology, University of Cambridge and Alexandra Albert, Research Fellow in Citizen Social Science, UCL
This article is republished from The Conversation under a Creative Commons license. Read the original article.