While exploring various ideas for the next issue of the blog the author came across a unique terminology that seems to have got the supply chain industry by storm. The trend is somewhat in the embryonic stage in terms of operational implementation, but the way the logistics sector is being catapulted towards a digital overhaul, it would not be a prophetic thing to say that a few years down the line, this phenomenon would gradually take the centre stage in supply chain management and optimization. The term that is being referred to over here is 'digital twin'.

For those who have not yet been initiated into the definition of it, digital twin, in the simplest words, alludes to the virtual replication of every possible aspect that is associated with the supply chain and that can include assets, warehousing, inventory positions et al. With the industrial world is undergoing the 4th revolution in the digital space and the coinage ‘Industry 4.0’ gaining momentum, digital twinning has become a key player in which many of the big players have invested judiciously and are reaping the benefits of it due to the technical and operational competence that it has brought in the ambit.

It must be noted that digital twinning is an umbrella term that contains a number of digital and computational upping such as cloud computing, IoT, AI, ML, simulation and so on. The biggest advantage that digital twin offers is that by digitizing the entire process of manufacturing and proving real-time data in operation, it upholds the entire lifecycle of a product, ranging from commissioning to maintenance. And in tandem, it generates reliable metadata that renders help in crunching numbers in the planning and optimization process.

While simulating the vantage points of the business, digital twinning also assists in pointing out the volatility in the businesses which in turn is fruitful in resolving these stumbling blocks. The process of digital twinning has been brought to prominence especially in the post-covid outbreak. The pandemic exposed the drawbacks that lay latent in the framework of the industry on the macro scale. And to withstand further breakdowns and disruptions, companies have scaled their business efficiencies to get armed with resilience and that is exactly where the efficacy of digital twinning comes to the fore.

There is no dearth in the veracity of the statement that supply chain business is an extremely complex multi-tier affair that involves multiple stakeholders and copious internal, as well as external factors impacting the business performances and balancing the difficult trade-offs, can be a cumbersome task. Since the digital twinning process encloses facets such as manufacturers, suppliers, distribution facilities, transportation routes and customer locations, it is efficacious in assessing the inventory, throughput, services by providing end-to-end visibility in AR/VR mode. With the real-time representation bringing transparency in the entire process, companies can strike a balance between the expectations and the reality in every possible sphere and can come up with corrective measures to mitigate the raptures and resolving bottleneck issues.

Since digital twinning can be an expensive enterprise, companies may start the procedure by short-term planning. Steadily they may veer into simulating sales and operational planning with the objective of optimizing the plan, minimizing losses and evoking scalable data to load the business further. Identifying the latent bottlenecks can empower the business, provide durability and help gain insights even in issues like market, customer retention, replenishment and maintenance.

The companies that have already adapted to the digital twinning and have created a digital space for themselves have a better edge over the others in gaining perspectives from a vantage point. Hence, it is indeed advisable that business leaders and executives ponder upon the viability of digital twinning and brace themselves for a massive swap in the system.

Artificial Intelligence(AI) is an umbrella term that embeds multiple aspects in it! To keep the nomenclature simple, we need to understand its torn down version i.e. Machine Learning(ML). Before delving into the topic, we need to know in brief what human learning is and how it is related with computers. In cognitive science, learning is typically referred to as the process of gaining information through observation.  In this context, the most obvious question that emerges is: 'Why do we need to learn at all?'.

Human body and mind are customised to carry certain activities which might be as simple and routine as walking down the street or doing the homework or it may be something very complex such as calculating the angle in which a rocket would be launched. To accomplish all these activities, we need to have some prior information on one or more things related to these task. As we keep learning, we tend to garner more experience and information related to the task. To elucidate the whole thing, we can cite the example of home assignment. With more knowledge we arm ourselves with the ability to do homework with less mistakes.  Thus with more learning our efficacy in performing the tasks increases.

There are mainly three types of human learning : learning under expert guidance, learning guided by knowledge gained from experts and learning by self. The readers might be wondering how does all these even correspond with our topic of discussion and the author probably tends to beat around the bush a little bit too much. But all the possible branches of knowledge in this world is linked with questions that we tend to overlook. It is only when we seek the answer, we learn in the process.

Hence, keeping strains with the three segments of learning process, we have to alight to this discussion: do inanimate objects like machines really learn? If so, then how do they learn?

If we go by a standalone, superficial answer then it’s a big No! Machines do not learn on their own but they do learn in the presence of a supervisor i.e. humans. If we be a bit more precise, learning is fed into machines under the direction of the human beings!

This essentially means that a machine can be considered to be gleaning learnings if it is able to gather experience by doing a certain task and improve its performance by doing the similar tasks in future.  When we talk about experience, it simply alludes to the past data related to the task. This data is basically instilled as an input to the machine by some source.

Let us now understand this with a simple example.  A machine has been given the task of image classification. In this context, let’s say that E is the collection of all the past data with images having  names or assigned classes(for eg. Whether the image is of class cat or class dog or a class elephant etc.). We now consider a  task of assigning a moniker  to new unlabelled images(by unlabelled we mean that the random image should fall under some particular category which the computer will decide) and P is the collection of performance measure i.e. the percentage of images correctly classified. It means how aptly the computer finds out whether the image of a cat is the image of a cat and not of a dog. The more we feed the machine with inputs the more its performance in identifying random images increases.

Thus for every problem out there that needs to be solved there are three fundamental questions that should be clear: 'What’s the problem? ', 'Why does the problem need to be solved?', 'How would I solve the problem?'

You might be musing why we are even discussing ML? We were supposed to talk about AI!

Well let me tell you that ML forms the basis of AI. In other words, AI is a bigger concept to create intelligent machines that can simulate human thinking capability and behavior, whereas machine learning is an application or subset of AI that allows machines to learn from data without being programmed explicitly.

Let’s see some of the basic applications of Machine Learning. Apart from the example that we have discussed before here are some of the basic examples based on our day to day life that uses the ML algorithms.

●  Speech Recognition: In some search engines, we get a feature of assistance through voice. It comes under the subset of speech recognition and it is a popular application of machine learning. Speech recognition is a process of converting voice instructions into text, and it is also known as 'speech to text', or 'computer speech recognition'. At present, machine learning algorithms are widely used by various applications. Many robotics developers are using speech recognition technology to follow the voice instructions.

●  Traffic Prediction: If we want to visit a new place, we take help of the navigation systems which show us the correct path with the shortest route and predicts the traffic conditions.

It predicts the traffic conditions too on the basis of the examination of data and can opine whether the traffic is smooth or is slow-moving or if the road is heavily congested. It is achieved with the help of two ways:

1.  Realtime location of the vehicle amassed through sensors

2.  Data on the average time taken on the previous days at the same time.

Everyone who is using these navigation systems is helping this app in making it better. It takes information from the user and sends back to its database to improve the performance.

●  Product Recommendations:  Machine learning is widely used by various e-commerce and entertainment companies  for the sake of product  recommendation to the user. Whenever we search for some product on the apps or websites of these online retailers, we start getting an advertisement for the same product while surfing internet on the same browser and this is because of machine learning.

Information pertaining to the user interest is collected using various machine learning algorithms and helps suggest the product as per customer interest. Similarly, when we use digital streaming platforms, we get to see some recommendations for entertainment series, movies et al and this is also done with the help of machine learning.

●  Automatic Language translation:  Nowadays, if we visit a new place and we are not aware of the language then we would not feel lost and helpless. Machine learning helps us in transcending this human barrier by converting the text into languages that we know. The technology behind the automatic translation is a sequence-to-sequence learning algorithm, which is used with image recognition and translates the text from one language to another.

●  Self Driving Cars: One of the most exciting applications of machine learning is self-driving cars. Machine learning plays a very significant role in bringing this marvel of science to being. One of the most popular car manufacturing companies in the world is working on self-driving car. It is using unsupervised learning method to train the car models to detect people and objects while driving.

●  ML in Supply Chain:   Supply chain, being a heavily data reliant industry, has many applications of machine learning. Some of the uses and benefits  of how ML can impact the Supply Chain Industry are as follows:

1.  AI/ML is implemented in the warehouses to optimise pick-path routes through voice recognition devices and robotic channels such as automatic guided vehicles and autonomous mobile robots. These mechanized models are fed with navigation systems and sensors that help in onboarding and seamless functioning within the warehouses. Other mechanized vehicles like GTP, carousels, conveyors, vertical lifts, forklifts help reduce errors in warehouse movement and amp up productivity and profitability.

2.  AI works towards optimisation of product flow in the supply chain through inventory control. Stockout or overstocking or staffing are the challenges that logistics industries meet on a daily basis. These sophisticated and innovative technologies render accuracy in inventory management and thus business performances are also achieved well.

3.  Machine learning helps derive actionable insights, allowing for quick problem solving and continual improvement.

As the conclusion to this article that tried to capture the essence of this much-debated, much-discussed grain of learning that what we see is just the beginning of a colossal digital revolution. In the  coming years ML/AI is going to automate nearly each and every human task.  Machine Learning is going to open up unparalleled opportunities for organizations enabling automation, efficiency, and innovation thereby making our lives much easier than ever before.