As I reflect on the Twelve Days of Teamcenter, which include some of the most important technology trends that would impact enterprise software in general and PLM solutions in particular, I can’t help but muse and pen down a few thoughts on the impact that enterprise Artificial Intelligence (AI) and Machine Learning (ML) may have across all industries that our customers are in.
Truly transformational innovations influence human lives and businesses as remarkably as they dissolve amorphously into every fabric of society. Electricity and internet have been two such landmark innovations of the past. Looking at the present and the future, looking at the pace of democratization in Artificial Intelligence / Machine Learning, it has the potential to surpass the impact of any other.
Focusing purely on the manufacturing sector, the demands of industry 4.0, edge computing (including autonomous vehicles), Smart Everything and overall market demands have been fueling a plethora of technologies based on AI/ML that help companies (small and large) create new business models and transform at scale.
The business-critical systems could broadly be classified into:
Data Systems – That provide simple digitalization mechanism of storing the data
Information Systems – That provide the ability to deduce insights from data, and provide automated events, process flows, structure relations and means to collaborate
Intelligent systems that provide insights, make decisions and perform actions
Improving on the present day data and information systems requires a move to leveraging AI/ML to make these systems “Intelligent” and provide more value for our customers. This is evident from openly available empirical and qualitative research available on this topic.
Let us analyze how cognitive science could be applied to transform the manufacturing sector and lead your business through intelligent systems:
Fundamental challenges for enterprise AI is the customers’ demand for automation and improved productivity.
• Unintuitive, labor intense and repetitive
• Requires a learning curve
• Complex constructs and paradigms, tough to discern
• Non-cognitive automation
• Data entry focused
• Information retrieval is linear to request
• Minimal analytics
• No Insights or Predictive Analytics
• Rule-based decision making – require complex syntax and tribal knowledge
Delivering enterprise AI/ML in Teamcenter
Siemens has been at the forefront of understanding our customers’ demands and delivering prescribed solutions. We will incorporate the power of enterprise AI/ML into our solutions and products for improved business value for our customers. In the next year, you will see these technologies emerging in our new Teamcenter releases.
As we tread the path of providing end-to-end solutions to our customers and enable you to achieve maximum efficiency in managing the digital thread and digital twin, we will introduce AI/ML services that will help you. For example, improve the “reusability” of parts and design by providing “visual search.” Augmenting it with “speech to text” and NLG (natural language generation) based interaction with the applications will open new avenues of user productivity. A “plant manager” could interact hands-free with a business-critical application using voice and retrieve knowledge-on-demand for an actionable outcome.
The magnitude of open source communities, technologies and the contribution by industry behemoths and academic institutions alike is a clear indicator of the rapid pace of innovation. In the areas of enterprise AI/ML, not only are their contributions of significant proportions, but there have been joint efforts among competitors. The underlying reason is simple; the more the technologies are democratized and made available for “citizen developers” the faster will be the realization of AI/ML into every walk of life. For example, Google has developed TensorFlow (an open source machine learning framework) and created a huge community of contributors and projects. Amazon Web Services' SageMaker has open sourced TensorFlow and MXNet enabled docker containers, in addition to tens of active projects in its open source initiative. Microsoft and other major industry leaders are on the same path
Looking at industry 4.0 and edge-computing needs, there are three major areas of focus: autonomous driving, robotics, and maintenance/monitoring. AI/ML has revolutionized all these areas, by making the edge devices more responsive, flexible, cognitive, cost-effective, secure and simple.
In the field of autonomous driving vehicles where reinforced learning models hold the key, AWS has announced DeepRacer League, as an example of reinforced learning and its democratization. Reinforced learning could understand the data patterns received from the sensors and perform fairly autonomous tasks on a “near real-time basis”. The self-learning and self-healing reinforced learning models can help improve decision making in real time and at scale.
Another ground-breaking trend with ML is AutoML, where the data scientists and business analysts need not spend time and effort preparing the data, selecting and training the model. AutoML will identify the data identification, and delivery outcome (predictions), with automating the intermediary steps. AutoML addresses some of the challenges of lack of expected accuracy with pre-trained models, and laborious process in creating a training pipeline on your own. The promise of AutoML has motivated Google Cloud, Microsoft, and AWS to invest heavily.
At Siemens, we are continuously making efforts to research, improve, and adopt these technologies into our products for helping the customer realize increased business value. We will continue to innovate and announce our solutions and products powered by AI/ML and those that bring the best business value for our customers, in a responsible way.