Copyright: informationweek.com – “Reducing the Environmental Impact of Artificial Intelligence (AI)”

By adopting energy-efficient architectures, optimizing AI models for performance, and pushing for cloud providers to embrace renewable energy, businesses can help reduce the carbon footprint of their AI solutions.

Artificial intelligence is reshaping our world. Its transformative power fuels innovation across industries — delivering new value to organizations and consumers alike. As the proliferation of AI accelerates, people are starting to ask important questions: How does AI impact the environment? And furthermore, how do we keep pushing for progress without leaving a heavy carbon footprint on the planet? 

AI’s Eco Impact

Artificial intelligence software runs in data centers that consume large amounts of energy and often cause significant carbon emissions. According to Bloomberg, there are more than 7,000 data centers worldwide. Collectively, they can consume as much power annually as the entire electricity production of Australia or Italy. The growing use of AI will further increase this already substantial energy consumption of data centers. 

The use of AI can be separated into two main tasks: training and inferencing. During training, AI models learn from vast amounts of data that can take months depending on data complexity and volume. Once an AI model has been trained, it consumes energy each time it generates a new response or “inference.” The International Energy Agency (IEA) has reported a ChatGPT inquiry requires up to 10 times the electricity of a Google search to respond to a typical request. This energy consumption adds up and can quickly surpass the energy used for training.

The WEF estimates training comprises about 20% of an AI model’s overall energy use across its lifespan, while inferencing makes up the remaining 80%.[…]

Read more: www.informationweek.com

Der Beitrag Reducing the Environmental Impact of Artificial Intelligence (AI) erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Copyright: webuildvalue.com – “The application of Artificial Intelligence (AI) in Webuild’s construction sites, a new frontier of innovation”

From Paris to Melbourne, here are the worksites where AI to serve workers is being tested.

The Only Limit is Imagination. The application of artificial intelligence to the construction of large-scale projects, and specifically to construction site activities, are still largely unexplored. However, ongoing innovations in experimentation promise to revolutionize the industry. The use of robotics, the connection and integration of machines, sensors, and devices on site that “communicate” with each other, predictive maintenance driven by data collection and analysis, and augmented reality for staff training are just some examples of what artificial intelligence can do and how it can help transform the construction world. Starting with safety—one of the strategic sectors where innovations can enhance productivity while also saving lives. In this regard, Webuild’s construction sites around the world represent a new frontier of artificial intelligence, where innovative solutions are tested to minimize workers’ exposure to risk.

«Globally, explains Dr. Giampiero Astuti, Head of Innovation Program Management at Webuild, there is a hierarchy of safety controls ranging from the cultural factors, so anything that is linked to human behaviors, to the technical-organizational controls, including the technological one, which aim to intervene to the source of hazard. This is where artificial intelligence comes into play.»

The application of artificial intelligence in Webuild’s construction sites is being tested precisely where the risks are highest: in the movements of large construction machinery, in workers’ behavior inside tunnels, and in the operations of TBMs, the massive tunnel boring machines that dig tunnels.

«The potential is immense, continues Astuti, and we have begun testing these new AI applications in some sites and on certain models, starting from Paris and extending to Australia».

The application of AI in the Grand Paris Express Construction Sites

One of the first applications of artificial intelligence was tested in the construction of the Grand Paris Express, the new metro network in Paris that will connect the towns of Île-de-France.[…]

Read more: www.webuildvalue.com

Der Beitrag The Application Of Artificial Intelligence (AI) In Construction Sites, A New Frontier Of Innovation erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

AI Unicorns and Strategic Shifts – SwissCognitive AI Investment Radar

Private equity firms and governments are increasingly seeking to control their own compute power, a critical asset as domestic AI startups surge in growth. PwC Australia is bolstering its commitment to AI with an additional $11.5 million investment into a Centre of Excellence, aiming to help businesses navigate the complexities of AI and stay competitive globally.

DevRev’s remarkable $100.8 million Series A funding round has propelled it into the AI unicorn club, now valued at $1.15 billion. Meanwhile, Amazon’s $4 billion investment in AI firm Anthropic has caught the attention of the UK’s competition watchdog, sparking concerns over potential anti-competitive practices.

As we look toward 2024, AI remains at the forefront of technological innovation, with cloud computing and cybersecurity also expected to dominate investment trends. Despite budget constraints, the UK government is moving forward with a £32 million investment in 98 AI projects, underscoring the strategic importance of AI even amid fiscal challenges.

Big Tech’s massive AI investments, particularly by Microsoft, are driving significant demand for Nvidia’s GPUs, while DevOps teams prioritize AI-driven testing to enhance software quality. On the global stage, AI is playing a transformative role in ESG (Environmental, Social, and Governance) integration in emerging markets, aligning investments with sustainability goals.

Previous SwissCognitive AI Radar: AI Investment Highlights and Strategic Developments.

Our article does not offer financial advice and should not be considered a recommendation to engage in any securities or products. Investments carry the risk of decreasing in value, and investors may potentially lose a portion or all of their investment. Past performance should not be relied upon as an indicator of future results.

Der Beitrag AI Unicorns and Strategic Shifts – SwissCognitive AI Investment Radar erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Copyright: cio.com – “AI is driving productivity and wage increases: Report”

Business sectors using artificial intelligence are seeing significant gains in productivity while AI skills are commanding higher wages, according to a new PwC report.

Industries such as financial services, information technology, and professional services are seeing labor productivity growth nearly five times greater than industries with less AI integration, the consulting firm said in a statement.

The report also highlights that jobs requiring AI expertise are associated with a considerable wage premium in major labor markets, including the US, the UK, Canada, Australia, and Singapore.

In the US, for example, these positions can offer an average of 25% higher salaries than non-AI jobs. The wage premium varies across professions, reaching 18% for accountants, 33% for financial analysts, 43% for sales and marketing managers, and 49% for lawyers.

This wage disparity is consistent across all analyzed markets, with AI skills consistently valued higher.

The report, analyzing over half a billion job advertisements across 15 countries, indicates that AI could enable many nations to overcome long-standing low productivity growth. This could lead to economic development, higher wages, and improved living standards, PwC added.

Upskilling imperative

The report also pointed out that job postings for AI-related positions are increasing 3.5 times faster than the overall job market. For every AI job listed in 2012, there are now seven.

However, the situation also demands more effort in skill development. Occupations significantly exposed to AI are experiencing a 25% faster change in skill requirements compared to those less exposed to it.[…]

Read more: www.cio.com

Der Beitrag AI is driving productivity and wage increases: Report erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Copyright: itbrief.com – “Five AI trends for 2024 – And How To Set Projects Up For Success”

We’ve seen a flurry of interest in not only generative AI (GenAI) based on large language models (LLMs), but our recent Annual Cloud Report revealed strong appetite for investment in AI more broadly, from computer vision systems to machine learning and data science for AI applications.

That’s great to see. AI has huge potential to lift productivity, improve customer service, and speed up product development. But let’s not forget, AI projects have traditionally had a high failure rate – 60 – 80% according to various reports by research groups.

There’s a growing sense of FOMO in the business community, which is leading to a headlong rush to develop and deploy AI platforms and services. Now is definitely the time to experiment. But the last thing you want to do is put time and money into projects that fizzle out or cause reputational damage because they create security or ethical issues.

Here are five trends we expect to see in AI in 2024 and some tips on how to make the most of the investment you put into your organisation’s AI efforts.

1. The Copilot productivity test

We’ve been told for years that intelligent assistants are coming that will cut through the admin and drudgery of office life, helping to manage our inboxes, draft documents and summarise information instantly. Well, the intelligent assistant era began in late 2023 with the arrival of Microsoft’s Copilot services for Microsoft 365 and rival services from the likes of Google.

In 2024, CIOs across Australia and New Zealand will be advising their senior leadership teams on whether to deploy these services to boost productivity. At a licence cost of around A$45, Copilot for Microsoft 365 it’s a hefty investment. We expect limited rollout to test the productivity promise before widespread deployment. The Australian Government is piloting Copilot across several government agencies.

Beyond productivity, there’s huge potential for these services to transform knowledge management by allowing an intelligent agent to analyse an organisation’s data in a secure environment to provide insights. Currently, the indexing costs of doing so can be prohibitive. But the price will come down in 2024 as adoption increases.

2. Rise of the model gardens

OpenAI and its free and premium ChatGPT services hogged the limelight this year. However, hundreds of LLMs have been developed and deployed across the tech ecosystem. The business model for providing LLMs is starting to take shape, with platforms offering a range of LLMs to suit your needs. AWS has its Bedrock service with foundational models from the likes of Stable Diffusion, Antropic, the open source Llama 2, and Amazon’s own Titan models. Google’s Model Garden features 100+ models allowing you to pick and choose what you need. The public cloud consumption model is now underpinning use of GenAI. In 2024, we will see the rise of ‘chaining’, where you use several models optimised for specific tasks to power a single product or service.[…]

Read more: www.itbrief.com

Der Beitrag Five AI trends for 2024 – And How To Set Projects Up For Success erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

SwissCognitive Guest Bloggers: Bidyut Sarkar, Senior Solution Manager, IBM USA and Rudrendu Kumar Paul – Boston University, Boston, USA – “Conversational AI on Manufacturing Floors With NLP-Enabled Assistants”

Takeaways:

• NLP-enabled assistants are transforming manufacturing by simplifying human-machine interactions.
• Industry giants like Toyota, Boeing, and Shell have witnessed enhanced efficiency and reduced errors through AI integration.
• The future of manufacturing envisions plant floors driven by data-rich, conversational interactions.

The advent of artificial intelligence in the manufacturing sector has brought a transformative era. As industries evolve, the integration of AI technologies becomes not just advantageous but essential. Natural language interfaces are a pivotal innovation among the myriad of advancements. These interfaces, rooted in human language and cognition principles, offer a seamless bridge between intricate machine operations and human understanding. In contemporary production settings, the ability to communicate with machines using everyday language can redefine operational efficiency. Such interfaces eliminate the barriers of complex coding languages, making data queries and command executions more intuitive. The shift towards these natural language interfaces underscores a broader movement in manufacturing: embracing AI not as a mere tool but as a collaborative partner. This partnership, built on the foundation of mutual understanding, promises to reshape the dynamics of production floors, making them more agile, responsive, and intelligent.

For Europe, the anticipated compound annual growth rate (CAGR) for the natural language processing industry from 2023 to 2030 is projected to be 15.19%, leading to an estimated market value of $17.41 billion by the end of the period. (Statista)

Source: Statista

At the same time, the value added by the manufacturing market in Europe is anticipated to reach $3.54 trillion in 2028, with an expected compound annual growth rate (CAGR) of 3.93% from 2023 to 2028. (Statista)

Source: Statista

The Role of Conversational Interfaces in Manufacturing

Conversational interfaces represent a paradigm shift in how humans interact with machines. At their core, these interfaces harness the nuances of human language, enabling a more intuitive communication pathway with technological systems. In the context of manufacturing, this innovation holds profound significance. Historically, interactions with machines required specialized knowledge, often demanding intricate command sequences or coding.

Conversational interfaces, on the other hand, simplify this interaction. They allow operators to engage with systems using natural language, making the process more accessible and less daunting. This shift democratizes access and accelerates response times as the need for translating thoughts into machine-specific commands diminishes.

Comparing conversational interfaces with their traditional counterparts reveals stark contrasts. Standard interfaces, often graphical or command-line-based, necessitate a learning curve and can limit responsiveness. Causal models break these barriers, offering a more fluid, adaptive, and user-centric approach. In essence, the evolution from traditional to conversational interfaces in manufacturing marks a transition from rigid, prescriptive systems to more flexible, understanding, and adaptive ones. This transition holds the potential to redefine the efficiency and adaptability of manufacturing processes. (Swiss Cognitive)

Technological Foundations

The underpinnings of conversational interfaces lie in two pivotal technological advancements: Natural Language Processing (NLP) and neural networks. NLP, a subfield of artificial intelligence, delves into the interaction between computers and human language. Its primary objective is to enable machines to understand, interpret, and generate human language meaningfully and contextually relevantly. This understanding forms the bedrock of any conversational interface, ensuring that interactions are syntactically correct and semantically coherent.

Neural networks, inspired by the structure and function of the human brain, play a complementary role. (IJAIM) These interconnected algorithms process information in layers, allowing for recognizing patterns and relationships in vast datasets. In NLP, neural networks facilitate the deep learning processes that drive language comprehension, sentiment analysis, and response generation.

When NLP and neural networks converge, the result is a conversational interface capable of understanding intricate language patterns, discerning context, and generating appropriate responses. Unlike traditional systems that rely on explicit programming for every possible interaction, these interfaces learn and adapt. They draw from vast linguistic datasets, refining their understanding with each interaction. This continuous learning, underpinned by the combined might of NLP and neural networks, empowers conversational interfaces to be dynamic, adaptive, and increasingly attuned to the nuances of human language. In the manufacturing sector, this translates to responsive and predictive interfaces, heralding a new age of intelligent interaction.

Leading Innovators in the Field

Several trailblazers have emerged in the dynamic landscape of conversational interfaces, each carving a distinct niche with innovative solutions. CoPilot.ai, Sigma, and Arria NLG have garnered significant attention for their pioneering contributions to manufacturing.

CoPilot.ai stands at the forefront of integrating artificial intelligence with human-centric design.

Their platform emphasizes intuitive interactions, ensuring operators can query and command production systems seamlessly. By prioritizing user experience, CoPilot.ai has managed to bridge the gap between sophisticated AI algorithms and the practical needs of manufacturing floors.

Sigma, on the other hand, has taken a data-driven approach. Their platform harnesses the power of big data analytics, combined with NLP, to offer insights and recommendations. This means real-time feedback, predictive maintenance alerts, and actionable insights that can significantly enhance operational efficiency in manufacturing. Sigma’s strength lies in transforming raw data into meaningful, actionable intelligence.

Arria NLG, focusing on the Natural Language Generation, brings a fresh perspective. Instead of merely understanding or interpreting human language, Arria NLG’s solutions excel in generating human-like text based on data. In manufacturing, this capability translates to detailed reports, summaries, and explanations generated on the fly, providing operators with a clear understanding of complex processes and data streams.

These innovators are redefining the boundaries of what’s possible in manufacturing. While varied in approach, their unique solutions share a common goal: to enhance the symbiotic relationship between humans and machines. By doing so, they are not only elevating the capabilities of individual operators but also setting the stage for a more collaborative and intelligent manufacturing future.

Source: Avnet

Real-world Applications and Case Studies

The theoretical promise of AI-powered conversational interfaces is compelling, but it’s in real-world applications where their transformative potential truly shines. Several industry giants have already begun harnessing these technologies, yielding tangible benefits.

Synonymous with automotive excellence, Toyota has integrated AI-powered assistants into its production lines. The primary objective was to combat the perennial challenge of downtime. By leveraging these advanced interfaces, Toyota’s operators can swiftly diagnose issues, receive instant feedback, and implement corrective measures. The result was a significant reduction in unproductive hours, ensuring that assembly lines run smoother and more efficiently.

Boeing, a behemoth in the aerospace sector, has turned to conversational interfaces to streamline its intricate manufacturing processes. Given the complexity of aircraft production, even minor inefficiencies can lead to substantial delays. Boeing’s adoption of these interfaces has enabled its engineers and technicians to access critical data, seek clarifications, and receive guidance without wading through cumbersome manuals or databases. The outcome has marked improved workflow efficiency and reduced production bottlenecks.

Shell, a global leader in the energy sector, faces the daunting task of managing vast and complex operations. The introduction of AI-guided processes has been a game-changer. These systems assist in monitoring equipment, predicting maintenance needs, and even guiding operators in crisis scenarios. The result is a more streamlined operation with a notable decrease in errors, leading to safer and more efficient energy production.

Beyond these industry leaders, several other enterprises have embraced the power of conversational AI. For instance, pharmaceutical companies use these interfaces for precision drug formulation, while textile manufacturers employ them for quality control. The common thread across these applications is straightforward: conversational interfaces, backed by robust AI, are ushering in a new era of enhanced productivity, reduced errors, and more intuitive human-machine collaboration.

Benefits of AI-Powered Production Assistants

Integrating AI-powered production assistants into manufacturing processes has ushered in a series of tangible benefits that are reshaping the industry landscape. One of the most pronounced advantages is the substantial reduction in downtime. By providing real-time diagnostics and predictive insights, these assistants enable swift identification and rectification of issues, ensuring that production lines remain operational and minimizing costly disruptions.

Furthermore, the precision and vigilance of AI assistants have led to a marked decrease in errors and mistakes. Unlike human operators, AI systems maintain consistent accuracy and may overlook anomalies or misinterpret data under pressure. Their ability to process vast amounts of data quickly and identify discrepancies means that potential issues are flagged and addressed before they escalate.

Lastly, the overarching impact of these advancements is the enhancement of overall efficiency and productivity. Production rates improve with streamlined workflows, instant access to data, and the elimination of common bottlenecks. Moreover, operators, freed from routine troubleshooting, can focus on more value-added tasks, driving innovation and quality.

In essence, adopting AI-powered assistants in manufacturing is not just about automating processes; it’s about elevating the entire production ecosystem to new heights of excellence.

The Future of Conversational Plant Floors

The journey through the intricacies of NLP-enabled assistants underscores their transformative potential in reshaping manufacturing dynamics. These advanced interfaces, bridging human intuition with machine precision, promise a future where communication barriers on production floors become relics of the past. As industries evolve, the vision is clear: plant floors will become hubs of data-driven conversations, where machines execute commands and offer insights, fostering a collaborative atmosphere. This synergy between human expertise and AI-driven insights is set to redefine manufacturing, heralding an era where conversational interactions drive innovation, efficiency, and unparalleled growth.

About the Authors:

Bidyut Sarkar, Fellow of the IET (UK) and author of books on AI is an expert in life sciences and industrial manufacturing industry solutions with applied AI/ML experience, having served as a keynote speaker and judge at startup competitions. His professional experience has taken him to various parts of the world, including the USA, Netherlands, Saudi Arabia, Brazil, Australia, and Switzerland.

Rudrendu Kumar Paul is an applied AI and machine learning expert and the author of multiple books on AI, with over a decade of experience in leading data science teams at Fortune 50 companies across industrial high-tech, automation, and e-commerce industries. Rudrendu holds an MBA, an MS in Data Science from Boston University (USA), and a bachelor’s degree in electrical engineering.

Der Beitrag Conversational AI on Manufacturing Floors With NLP-Enabled Assistants erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

SwissCognitive Guest Blogger: Artem Pochechuev, Head of Data Science at Sigli – “Advantages of AI in Building Solutions For People With Limited Mobility”

According to various studies around 10%-12% of people worldwide have mobility issues that result in significant difficulty walking or climbing stairs. Governments and NGOs all over the globe invest money and effort in the identification of the existing barriers and their gradual removal. However, the reality is that a huge number of people with limited mobility still have no other choice but to spend the majority of time at home because a lot of cities and towns do not have a suitable infrastructure for those who have to use mobility aids like wheelchairs, mobility scooters, crutches, walkers, and other tools.

When we are talking about people with limited mobility we should bear in mind that in many cases we will deal not only with those who were born with disabilities. In this case, we should also mention people who faced mobility issues as a result of an accident or illness. Moreover, such issues may also have a temporary character. Given all these factors, it is vital to focus not only on urban infrastructure but also on the home environment, as well as conditions for working and studying that people will need to deal with. It means that we need to cover a quite wide category of existing issues and possible solutions.

As well as in many other cases, modern technologies can be a good answer to address the existing difficulties. Artificial intelligence has already found its use cases in many projects aimed at increasing the quality of life for people with disabilities. It’s worth highlighting that the needs of people with limited mobility can be also catered to with the power of AI.

Let us share our thoughts on how is it possible to use the opportunities that AI offers us to make it easier for everyone to deal with numerous daily tasks without external help.

Examples of AI-powered solutions

To begin with, we’d like to start with the solutions that are already widely used in many families and facilitate a lot of processes for them, even if their members do not experience any health problems or issues with their physical conditions. However, for people with physical disabilities and mobility difficulties, they will have a much more significant role.

We are talking about voice-controlled home automation systems that are able to interpret voice commands and perform various tasks from simple ones like turning music on or off to telling a user who is ringing at the doorbell (in this case, the system should be also enriched with face recognition tools). Amazon’s Alexa or Google Home Assistant are among the most well-known examples provided by tech giants but today there are many more solutions of this kind being developed by ambitious startups. Already now we can speak about the growing adoption of such systems and it will be reasonable to presuppose that in the future even more families will leverage the benefits of such solutions.

AI-powered self-driving cars also have huge potential for enhancing the freedom of mobility and eliminating physical isolation. Thanks to driverless cars developed by Waymo, Tesla, General Motors and others will allow people to avoid a lot of issues related to the use of public transport and traditional cars as well as the lack of possibility to get from point A to point B due to their physical disabilities. Though there are still a lot of things to do to make the use of such means of transportation fully safe, today the expectations are very high and we can hope that in some years they will become more accessible and affordable for a wide audience of drivers.

Another example of using the capacities of AI in the context of reducing barriers for people with limited mobility is the introduction of robotic assistance for mobility and the creation of new solutions built upon the existing mobility aids, such as wheelchairs or walkers, for example. Quite often people with disabilities are not able to use traditional wheelchairs on their own (or have serious difficulties while doing it). Nevertheless, AI can address this problem by perceiving audio commands from the user which can provide users with much more freedom of movement. Moreover, not every person can use traditional joysticks. Such conditions as spinal cord injury, spinal muscular atrophy, motor neurone disease, and some others can affect the hand function of people. An Australia-based company Control Bionics that focuses on advanced assistive technology developed a wireless wearable device NeuroNode. It lets people rely on their brain signals to fulfill different tasks that are typically performed with the help of a touch screen, keyboard, joystick, or mouse. To control a cursor on a screen people can eye-track and to opt for some actions they need to send a neural signal. The adoption of such solutions can also become a real game-changer for people who have difficulties in controlling the muscles that are required for speaking as they will be able to communicate with others with the help of brain-operated text-to-speech tools.

In collaboration with Deakin University’s Applied AI Institute, Control Bionics created DROVE which is known as the first autonomous driving wheelchair module in the world. The module has been already tested by users at their homes. It is powered by the NeuroNode interface and a digital camera system mounted on a wheelchair. Moreover, to ensure centimeter accuracy, researchers installed sensors at the locations where it was planned to use the solution. The system has proved its efficiency by demonstrating its ability to navigate tight doorways, detect unexpected obstacles, and always leave the wheelchair in the required position.

The researchers at The University of Texas at Austin have also contributed to building an inclusive mobility future. They developed the technology that allows users to control the movement of a wheelchair with the power of their mind. It includes a skullcap with electrodes and they can detect those brain signals that regulate movement. A laptop that is mounted on a wheelchair has AI-powered software that can translate these signals into wheel movements. It means that to move a wheelchair a person needs to imagine how he or she is moving legs or arms.

The row of examples of smart robotic solutions for people with disabilities is rather wide.

LEA (the Lean Empowering Assistant) is a robotic walker introduced by Robot Care Systems with a view to increasing stability and safety for elderly people and people with reduced mobility. LEA is powered by sensor technology that can ensure autonomous navigation by scanning the environment and reacting to different conditions. Let’s suppose that LEA detects that there is an object on the floor that can cause a fall of a user. After the detection of an obstacle, the system will notify a person and the walker will slow down to ensure a high level of safety.

There are also different projects that address some particular needs of people with highly limited mobility. While some robots, like RIBA, can lift up a person from a bed and set down to a wheelchair (or vice versa), some others can become dining companies and feed them. Obi is a good example of the solutions from the latter category. It is designed to reduce assistance required for people with disabilities. Obi has several bowls for food and a robotic arm. It can learn where it should deliver a spoon with food after a caregiver once shows this to it.

Speaking about possible solutions for people with limited mobility, we also have to mention AI-powered prosthetics and exoskeletons. AI helps to overcome the existing limitations of traditional prosthetics. It is possible thanks to providing enhanced functionality, much more intuitive control, and better signal decoding. ML and AI help to adapt prosthetics to various conditions and environments in accordance with the feedback from the user’s body.

Exoskeletons are wearables that support movement and expand the physical capacities of people. Artificial Intelligence can make movements more neutral and reduce the required physical efforts which makes such devices more user-friendly.

In 2021, researchers at the University of Waterloo, Ontario, made headlines when it was revealed that they had started testing AI-powered semi-autonomous exoskeletons. These exoskeletons are intended to help people with limited mobility walk again with the help of robotic prostheses enriched with deep-learning technologies. What makes this project special is that human thought is not required for controlling exoskeletons. It is possible to compare them with autonomous vehicles that can drive themselves. These exoskeletons can walk themselves. The robotic limbs have sensors and cameras that provide images to computer-vision algorithms for analyzing the surroundings. For example, when stairs are detected, the exoskeleton’s control system will get a signal to start a relevant scenario. It will include such commands as lifting a knee and stepping down or up depending on the exact location of a user. It means that people wearing exoskeletons can move while their movements will be controlled by special software.

AI-powered exoskeletons are expected to play a significant role in helping individuals with disabilities and also elderly people. Such wearables of different types can be useful not only for moving around but also for lifting and carrying heavy objects and reducing the load on the human body. Though right now, a lot of products are available only as prototypes, we can presuppose that with the time flow, great progress will be made in this field.

Conclusion

It’s always inspiring to observe how technologies can be used in real life. And it is even more inspiring to see that they can be used to help people who face various difficulties caused by impairments and health states. While traditional walking aids have a lot of restrictions and drawbacks, AI can become a great booster in the development of more advanced, more powerful, and more convenient-to-use solutions. Of course, at the current moment, such solutions (even if already commercially available) have a rather limited target audience due to their high costs. Nevertheless, their appearance is already excellent proof that the work in this sphere is going on and a lot of researchers and businesses are ready to make their contributions.

In the next articles in this series, we will discuss other use cases of AI and possibilities to leverage the power of this technology. If it sounds interesting to you, just don’t miss our posts!

About the Author:

In his current position, Artem Pochechuev leads a team of talented engineers. Oversees the development and implementation of data-driven solutions for Sigli’s customers. He is passionate about using the latest technologies and techniques in data science to deliver innovative solutions that drive business value. Outside of work, Artem enjoys cooking, ice-skating, playing piano, and spending time with his family.

Artem will be speaking at the SwissCognitive World-Leading AI Network AI Conference focused on The AI Trajectory 2024 – Invest for Impact on 13th December.

Der Beitrag Advantages of AI in Building Solutions For People With Limited Mobility erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

SwissCognitive Guest Bloggers: Bidyut Sarkar, Senior Solution Manager, IBM USA and Rudrendu Kumar Paul – Boston University, Boston, USA – “AI-Powered Predictive Maintenance in Advanced Manufacturing”

Takeaways:

• Deep learning offers unparalleled precision in predictive maintenance by analyzing intricate patterns in sensor data.
• Collaboration between domain experts and data scientists is crucial for effective model implementation.
• Embracing deep learning in maintenance strategies can lead to significant operational efficiency and cost savings.

Once a novel concept, predictive maintenance has evolved significantly with technological advancements. Historically, industries relied on rudimentary methods to predict equipment failures. However, the landscape transformed with artificial intelligence and deep learning. These cutting-edge technologies have ushered in a new era, offering unparalleled insights into the health and longevity of machinery. Deep understanding, in particular, has demonstrated its prowess by analyzing intricate patterns from sensor data, thereby enhancing the precision of predictions. This shift not only underscores the potential of modern algorithms but also highlights the transformative impact of technology on industrial operations.

In Europe, the market for machine learning (which includes deep learning applications) is forecasted to expand from $43.40 billion in 2023 to $144.60 billion by 2030, with a compound annual growth rate (CAGR) of 18.76% during this period.

Source: Statista

At the same time, the global predictive maintenance market is projected to grow to $64.3 billion by 2030, with a compound annual growth rate (CAGR) of 31% from 2022 to 2030. (Statista)

Source: Statista

The Limitations of Traditional Predictive Maintenance

Historically, predictive maintenance relied on essential monitoring tools and heuristic techniques. These methods, while foundational, often fell short of accurately forecasting equipment malfunctions. Relying on rudimentary sensors and manual inspections, traditional approaches needed more granularity to detect subtle anomalies or predict failures with high confidence. Furthermore, these techniques were susceptible to human error and could not adapt to the evolving complexities of modern machinery. Such limitations underscored the pressing requirement for innovations that could offer more detailed insights and higher predictive accuracy. As industries grew and machinery became more intricate, the inadequacies of conventional predictive maintenance became increasingly evident, paving the way for the integration of advanced technological solutions.

Deep Learning: A Game Changer for Predictive Maintenance

Deep learning, a subset of artificial intelligence, harnesses neural networks with multiple layers to analyze vast amounts of data; unlike traditional algorithms that plateau after a certain data threshold, deep learning thrives on extensive datasets, extracting intricate patterns often invisible to other methods. In the context of predictive maintenance, this capability is invaluable. (Swiss Cognitive)

Machinery, especially in industrial settings, generates a plethora of sensor data. This data, rich in minute details, holds the key to understanding the health and potential vulnerabilities of equipment. With their advanced neural structures, deep learning models efficiently sift through this data, identifying patterns and anomalies that might indicate impending failures. By doing so, these models offer a nuanced understanding of equipment health, allowing industries to address issues before they escalate preemptively.

The true prowess of AI (which includes deep learning) lie in its ability to discern patterns from seemingly random data points. In predictive maintenance, this means recognizing the early signs of wear and tear or the subtle hints that a machine part might be on the brink of malfunction. Thus, deep learning is a beacon of innovation, revolutionizing how industries approach equipment maintenance. (IJAIM)

Source: KSB

Real-world Applications and Case Studies

In the evolving landscape of predictive maintenance, several trailblazing entities have emerged, leveraging deep learning to redefine industry standards. Their applications provide compelling evidence of the transformative potential of this technology.

Uptake

One notable entity in this domain is Uptake, which has made significant strides in forecasting outages. By harnessing the power of deep learning, Uptake’s models analyze vast datasets to predict potential disruptions. The implications of such precise forecasting are profound. By averting unplanned downtimes, industries can optimize operations, reduce costs, and enhance overall productivity. Moreover, the ripple effect of these advancements extends beyond mere operational efficiency, influencing supply chains, labor management, and even environmental sustainability.

Augury

Another pioneer, Augury, has carved a niche in detecting nuanced data indicators that hint at equipment health. Traditional methods often overlook these subtle signs, but with deep learning’s intricate pattern recognition, Augury’s models can pinpoint anomalies with remarkable precision. Such capabilities enable industries to undertake timely interventions, ensuring machinery longevity and reducing the risk of catastrophic failures.

C3 AI

C3 AI stands out with its commendable achievement of over 85% accuracy in predictive analytics. Such a high degree of precision is a testament to the prowess of deep learning models that can sift through complex data structures, identifying patterns that would otherwise remain obscured. This accuracy bolsters confidence in predictive maintenance strategies and underscores the potential for further refinements and innovations in the field.

Delving deeper into specific applications:

• ML Forecasting Bearing Faults: Bearings, critical components in many machines, can exhibit faults that, if undetected, can lead to significant operational challenges. Deep learning models have demonstrated their capability to forecast these faults by analyzing vibrational data, temperature fluctuations, and other sensor outputs, ensuring timely interventions.
• Pump Cavitation Detection: Cavitation in pumps, where vapor bubbles form in the liquid due to pressure changes, can harm equipment health. Through deep learning, subtle signs of cavitation, often missed by conventional methods, can be detected, allowing for preventive measures.
• Predicting Wind Turbine Failures: Wind turbines, monumental feats of engineering, are not immune to wear and tear. When processed through deep learning algorithms, their vast data outputs can predict potential failures, from blade issues to gearbox malfunctions, ensuring optimal energy production and equipment longevity.

These real-world applications underscore the transformative impact of deep learning on predictive maintenance, heralding a new era of efficiency and precision.

Solution Strategies in Implementing Deep Learning for Predictive Maintenance

Incorporating deep learning into predictive maintenance is a nuanced endeavor, necessitating adherence to certain best practices to ensure optimal outcomes.

Model Governance

At the heart of any deep learning initiative lies the model itself. Ensuring its reliability and consistency is paramount. This involves rigorous testing, validation, and monitoring of the model in real-world scenarios. A robust governance framework makes the model behave as expected, even when encountering diverse and evolving datasets. Furthermore, documentation of model parameters, training methodologies, and validation results aids in maintaining transparency and trust.

Iterative Improvement

The dynamic nature of machinery and operational environments means that a one-size-fits-all model is a myth. As such, continuous refinement of deep learning models is essential. Industries can enhance predictive accuracy over time by revisiting and updating models based on new data and feedback. This iterative approach ensures that models remain relevant and practical, even in changing industrial landscapes.

Practitioner Collaboration

The success of any predictive maintenance initiative hinges on the synergy between data scientists and maintenance experts. While data scientists bring expertise in model development and data analysis, maintenance experts possess invaluable domain knowledge. Collaborative efforts between these professionals can lead to models that are not only technically sound but also contextually relevant. Such collaboration ensures that the insights derived from deep learning are actionable and aligned with on-ground realities.

Adhering to these best practices can significantly augment the efficacy of deep learning in predictive maintenance, ensuring sustainable and impactful results.

The Future of Predictive Maintenance with Deep Learning

The trajectory of predictive maintenance, guided by deep learning, paints a promising picture. As computational capabilities expand and datasets grow more affluent, the potential for refining and enhancing predictive models becomes increasingly evident. These advancements could lead to even more nuanced detections, capturing the minutest of anomalies that might have previously gone unnoticed.

Industries stand at the cusp of this transformative era, and preparation is crucial. Embracing a culture of continuous learning and fostering an environment conducive to innovation will be pivotal. Investing in training programs that bridge the knowledge gap between traditional maintenance practices and modern data-driven approaches can also prove beneficial. Moreover, as technology evolves, so should the strategies, ensuring that industries remain agile and adaptive.

In essence, the fusion of deep learning with predictive maintenance heralds a future marked by unparalleled precision, proactive interventions, and enhanced operational efficiency.

Challenges and Considerations

While integrating deep learning into predictive maintenance offers immense promise, it has challenges. A primary consideration is the data itself. Both quality and quantity are paramount; models trained on insufficient or skewed data can produce misleading results, potentially leading to costly misjudgments.

Additionally, the intricacies of machinery and equipment demand domain expertise. Mere algorithmic prowess needs to be improved. Collaborative efforts between domain experts and data scientists are essential to ensure that models are grounded in practical realities.

Lastly, concerns surrounding transparency and trustworthiness arise, as with any AI-driven initiative. Black-box models, which offer little insight into their decision-making processes, can be a source of apprehension for industries. Addressing these concerns through explainable AI methodologies and rigorous validation can help build confidence and ensure the responsible adoption of deep learning in predictive maintenance.

Conclusion

The fusion of deep learning with predictive maintenance signifies a pivotal shift in how industries approach equipment health and longevity. This synergy offers an unparalleled opportunity to detect intricate patterns, forecast potential failures, and ensure timely interventions. As the technological landscape continues to evolve, industries stand to gain immensely from these advancements, reaping benefits in terms of operational efficiency, cost savings, and machinery lifespan. Forward-thinking entities must recognize this potential and actively integrate deep learning methodologies into their maintenance strategies. Doing so, they pave the way for a future marked by precision, proactivity, and enhanced productivity.

About the Authors:

Bidyut Sarkar, Fellow of the IET (UK) and author of books on AI is an expert in life sciences and industrial manufacturing industry solutions with applied AI/ML experience, having served as a keynote speaker and judge at startup competitions. His professional experience has taken him to various parts of the world, including the USA, Netherlands, Saudi Arabia, Brazil, Australia, and Switzerland.

Rudrendu Kumar Paul is an applied AI and machine learning expert and the author of multiple books on AI, with over a decade of experience in leading data science teams at Fortune 50 companies across industrial high-tech, automation, and e-commerce industries. Rudrendu holds an MBA, an MS in Data Science from Boston University (USA), and a bachelor’s degree in electrical engineering.

Der Beitrag AI-Powered Predictive Maintenance in Advanced Manufacturing erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Billion-Dollar AI Bets – SwissCogntive AI Investment Radar

Time has consistently showcased that the AI realm is akin to a meteor shower – each breakthrough a brilliant spark, every investment a testament to a tech future that’s brighter than we might have imagined.

In this edition, our gaze first rests on Microsoft’s colossal commitment to Australia, not just financially, but in reshaping an entire digital ecosystem. This strategic play promises to forge a new digital frontier, championing Australia’s rise as a formidable player in the AI epoch. Yet, the world of tech is vast and varied, and as Apple embarks on its own journey into the depths of generative AI, we see glimpses of a future filled with smarter, more intuitive devices, perhaps changing the way we engage with technology altogether.

But it’s not all smooth sailing. As we bask in the glow of AI’s potential, shadows of challenges loom. Regulatory hurdles, ethical dilemmas, and the sheer pace of advancement demand our attention. Amidst these movements, AI bets are becoming more strategic than ever. It’s about startups like Celes pushing boundaries in niche sectors, it’s about the very chips that power our AI dreams, and it’s about the confluence of AI with other revolutionary technologies, such as VR, reshaping industries we hold dear.

From investments in AI-driven music revolutions to advancements in AI finance tools and analytic platforms, our radar’s sweep is comprehensive. We’re here to ensure you not only stay informed but deeply connected to every significant pulse and pivot in the world of AI. So, as we unfurl the tapestry of this edition’s stories, join us on this enlightening voyage into the heart of AI’s transformative age.

Previous SwissCognitive AI Radar: AI Pioneers, Unicorns, and Market Movement Updates.

Our article does not offer financial advice and should not be considered a recommendation to engage in any securities or products. Investments carry the risk of decreasing in value, and investors may potentially lose a portion or all of their investment. Past performance should not be relied upon as an indicator of future results.

Der Beitrag Billion-Dollar AI Bets – SwissCognitive AI Investment Radar erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.

Copyright: newatlas.com – “AI Analysis will Predict Future Disease from Bone Density Scans”

Just as calcification, or the deposit of calcium, on the inner wall of blood vessels in the heart can be problematic, so can calcification of the aorta, the largest artery in the body. Exiting the heart, it branches upward to supply blood to the brain and arms and extends down to the abdomen, where it splits into smaller arteries that supply each leg.

Abdominal aortic calcification (AAC), calcification in the section of the aorta that runs through the abdomen, can predict the development of cardiovascular diseases such as heart attack and stroke and determine mortality risk. Previous studies have also found that it’s also a reliable marker for late-life dementia. AAC is visible on bone density scans typically used to detect osteoporosis in the lumbar vertebrae, but a highly trained professional is required to analyze these images, which takes time.

AAC is commonly quantified by trained imaging specialists using a 24-point scoring system, AAC-24. A score of zero represents no calcification, and a score of 24 represents the most severe degree of AAC. Now, researchers from Edith Cowan University in Australia have turned to machine learning to speed up the calcification assessment and scoring process.

The researchers input 5,012 spinal images, taken by four different models of bone density machines, into their machine learning model. Though other algorithms have been developed to assess AAC from these types of images, the researchers say this study is the biggest and the first to be tested in a real-world setting using images taken from routine bone density testing.

They then assessed the model’s performance in accurately classifying images into low, moderate and high categories of calcification based on their AAC-24 score. To check for accuracy, the machine-learning-based AAC scores were compared with scores given by human specialists. The specialist and the software arrived at the same determination 80% of the time. Three percent of people with high AAC scores were incorrectly diagnosed as having low scores by the software.[…]

Read more: www.newatlas.com

Der Beitrag AI Analysis will Predict Future Disease from Bone Density Scans erschien zuerst auf SwissCognitive | AI Ventures, Advisory & Research.