How AMAZON Leveraging Power of Artificial Intelligence Machine Learning

What is Artificial Intelligence?

Artificial Intelligence (AI) is the branch of computer sciences that emphasizes the development of intelligent machines, thinking and working like humans. For example, speech recognition, problem-solving, learning, and planning.

Artificial Intelligence (AI) refers to the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions. The term may also be applied to any machine that exhibits traits associated with a human mind such as learning and problem-solving. The ideal characteristic of artificial intelligence is its ability to rationalize and take actions that have the best chance of achieving a specific goal.

AI programming focuses on three cognitive skills: learning, reasoning, and self-correction.

Learning processes. This aspect of AI programming focuses on acquiring data and creating rules for how to turn the data into actionable information. The rules, which are called algorithms, provide computing devices with step-by-step instructions for how to complete a specific task.

Reasoning processes. This aspect of AI programming focuses on choosing the right algorithm to reach the desired outcome.

Self-correction processes. This aspect of AI programming is designed to continually fine-tune algorithms and ensure they provide the most accurate results possible.

What is Machine Learning ?

Machine learning is a concept or Application under Artificial intelligence, that enables systems to automatically learn, adjust actions without explicit programming, and improve from past experience.

machine-learning works like a human mind where the human mind can learn from experiences and predict some new, machine learning do the same. it can learn the pattern of the given input and output as experiences to the system and after analyzing the whole pattern it provides the ability to system to predict something future value using these experiences and patterns. Machine learning is nothing but human provide some ability to the system. in machine learning there is no requirement to program explicitly and Behind machine learning, there is a lot of algorithm processing that analyzes the pattern.

 ML is one of the most exciting technologies that one would have ever come across. As it is evident from the name, it gives the computer that makes it more similar to humans: The ability to learn. Machine learning is actively being used today, perhaps in many more places than one would expect.

Artificial Intelligence (AI) refers to human Intelligence exhibited by machines. AI can be classified into Strong and Weak AI. Strong AI (or artificial general intelligence) is a machine with consciousness, sentience, and mind, and this machine has intelligence in more than one specific area. Weak AI (or artificial narrow intelligence) focuses on specific narrow tasks (e.g., self-driving cars). Machine learning is an approach to achieve artificial intelligence, and deep learning is a branch of machine learning and a technique for realizing machine learning. Deep learning focuses on algorithms inspired by the structure and function of the human brain. Robotics deals with the design, development, operation, and application of robots.

Case Study`

A M A Z O N :

Amazon’s approach to Artificial Intelligence is called a flywheel. It’s a simple tool that efficiently stores rotational energy. It stores energy when a machine isn’t working constantly. It does not waste energy in turning the machine on or off, it keeps the energy constant and expands it to other areas of the machine.

At Amazon, artificial intelligence and machine-learning technology are not confined to one business segment. The technology is everywhere, utilized across the teams that back the Alexa suite of voice-activated devices, the Amazon Go stores, and the recommendation engine that cause “Frequently Bought Together” or “Customers Who Bought This Item Also Bought” purchase recommendations to populate. 

However, AI-powered technology and deep learning power one of Amazon’s most critical elements of its business — delivery, which is fully dependent on a fluid warehouse operation.

As the U.S. pioneer of one-day shipping, the intricacies of the company’s fulfillment center processes will continue to adapt and evolve to make the end-to-end fulfillment model more streamlined, automated, and sophisticated. In this article, we highlight how Amazon leverages artificial intelligence and machine learning to optimize delivery at scale.

For the record, AI and machine learning go hand in hand, so the way Amazon uses AI works alongside machine learning (ML). Amazon is building many businesses that run on machine learning-based systems.

Amazon fulfillment center :

According to Amazon, the integration of robots at fulfillment centers make it possible to store 40% more inventory, which in turn enable it to more rapidly fulfill Prime one- or two-day shipping orders, and more robotics and AI-related innovations will continue to develop and become integrated into Amazon’s global operations model.

With anywhere from 1-4 million product bins per fulfillment center, the Amazon Worldwide Operations team is constantly refining its processes and leveraging technology to determine, in real-time, which orders should be picked at the same time to get the products that belong together in the same box.

The team leverages computer vision systems that analyze images to securely track where each item is located throughout the warehouses, which are typically set up with a Manhattan-style grid and have specific, structured paths for the product pods to follow. Each pod has about nine rows of shelves to hold products on all four sides and each warehouse typically ranges between 600,000 to 1 million square feet in size.

Although Amazon says that fully automated shipping warehouses are at least a decade away, robots are currently operating these product pods — in harmony with humans — at 26 of Amazon’s more than 175 fulfillment centers. Spearheaded by Amazon Robotics, which was founded in 2003 as a subsidiary of Amazon.com, fulfillment center automation is backed by technologies including autonomous mobile robots, sophisticated control software, language perception, computer vision, depth sensing, machine learning, object recognition, and semantic understanding of comments.

The company also utilizes artificial intelligence to determine how many units of a specific product it anticipates customers to buy, which then factors into where the product is stocked so it is as close as possible to the people who will buy it. 

At Amazon’s re:MARS conference in June 2019, Brad Porter, head of Amazon Robotics, announced two platforms that Amazon has begun integrating into its warehouse operations globally. The first is Pegasus, an item-categorizing system that is predicted to cut down wrongly sorted products by 50%. The other, Xanthus, is a modular drive system that quickly adapts to new applications and attachments that allow it to carry different types of cargo.

With decision engines and decision logic, the technology uses data and information to minimize the distance the pods have to travel at a given time, making decisions constantly as the data on the back end changes. Additionally, the technology builds predictions on how likely a pod may need to be accessed in the next one, two, or three hours. Eventually, a worker will take a specific item from a predetermined station that the robot will travel to and put it on a conveyor belt to prepare it to be shipped. 

Once a shipping label is put on the box, the transportation execution processes take over and use machine learning to determine the most effective way the package should travel from point A to point B. The box is sent to a waiting trailer based on its shipping method, speed of delivery, and location. On a daily basis, the team at Amazon uses machine learning and optimization algorithms to improve each warehouse process for one-day shipping, which consumers are already taking advantage of with more than 10 million products.

Amazon Recommendation Engine :

From using AI to predict the number of customers willing to buy a new product to running a cashier-less grocery store, Amazon’s AI capabilities are designed to provide customised recommendations to its customers. According to a report, Amazon’s recommendation engine is driving 35% of its total sales.

Amazon Recommendations: Amazon practically invented the concept of giving personalized product recommendations after online purchases, using an algorithm they call “item-based collaborative filtering.” This algorithm makes the homepage of each of its many millions of customers unique, based on their interests and previous purchasing history. The Semantics Scholar article, “Two Decades of Recommender Systems at Amazon.com” puts it best:

Amazon.com has been building a store for every customer. Each person who comes to Amazon.com sees it differently because it’s individually personalized based on their interests. It’s as if you walked into a store and the shelves started rearranging themselves, with what you might want moving to the front, and what you’re unlikely to be interested in shuffling further away.

Amazon doesn’t only use the purchase data of each of their customers, they also utilize the purchase histories of other people that purchased the same product, giving “frequently bought together” information on their product listings. Furthermore, they factor in customer feedback and ratings. How? By offering recommendations that match a customer’s interests as well as reported customer satisfaction, price, and quality level.

Amazon continues to improve their collaborative filtering by connecting purchase history with browsing data. If, for instance, a customer purchased socks, Amazon may not suggest just socks in the future. Instead, their algorithm may look at an individual’s browsing history, see they watch a superhero movies on Prime, and in turn, recommend Marvel brand shirts.

To Amazon, interpreting massive amounts of divergent data in real-time is key, and this recommendation engine is responsible for a whopping 35% of their total revenue.

Amazon Go :

To align with their expanding artificial intelligence footprint and foray into various leading businesses (cloud computing, advertising, and voice technology, for example), Amazon has created two brick-and-mortar grocery stores in their home base of Seattle that are the first of its kind — no lines, no checkout stands, and fully tailored to grab-and-go convenience.

Amazon describes Amazon Go as “a new kind of store with no checkout required. We created the world’s most advanced shopping technology so you never have to wait in line.” Through what the e-commerce leader has coined as “Just Walk Out Technology,” customers solely need to download the Amazon Go app before arriving and scan a QR code on the app for entry before they are free to explore the store. Interestingly enough, having an Amazon Prime account is not a requirement for store entry.

The company compares the technology used in the Amazon Go stores to the same kind used in self-driving cars — computer vision, deep learning algorithms, and sensor fusion. There are cameras mounted throughout the store that track customers’ movements from all angles.

The combination of these elements allows for instant detection of when products are taken off or returned to the shelves.

The AI technology monitors the items as if they were sitting in a virtual cart. When you have completed your shopping, you can just leave the store.

After the fact, Amazon emails you the receipt from that transaction and charges your Amazon account.

In the future, Amazon plans to expand the cashier-less stores to Chicago and San Francisco. The two current Seattle stores feature a variety of ready-to-eat breakfast, lunch, dinner, and snack options, as well as Amazon’s meal kits that can be prepared at home and take about 30 minutes to prepare. The app also lists prices for all of the items sold in the store, so you can decide ahead of time whether the price point of a specific item makes sense for you before making the trip to the store. To find out what one of the specific stores sells, customers can go to the Discover tab in the Amazon Go app

Amazon Alexa :

Amazon Alexa, also known simply as Alexa, is a virtual assistant AI technology developed by Amazon, first used in the Amazon Echo smart speakers developed by Amazon Lab126. It is capable of voice interaction, music playback, making to-do lists, setting alarms, streaming podcasts, playing audiobooks, and providing weather, traffic, sports, and other real-time information, such as news. Alexa can also control several smart devices using itself as a home automation system. Users are able to extend the Alexa capabilities by installing "skills" (additional functionality developed by third-party vendors, in other settings more commonly called apps such as weather programs and audio features).

Most devices with Alexa allow users to activate the device using a wake-word (such as Alexa or Amazon); other devices (such as the Amazon mobile app on iOS or Android and Amazon Dash Wand) require the user to push a button to activate Alexa's listening mode, although, some phones also allow a user to say a command, such as "Alexa" or "Alexa wake". Currently, interaction and communication with Alexa are available only in English, German, French, Italian, Spanish, Portuguese, Japanese, and Hindi. In Canada, Alexa is available in English and French (with the Quebec accent).

As of November 2018, Amazon had more than 10,000 employees working on Alexa and related products. In January 2019, Amazon's devices team announced that they had sold over 100 million Alexa-enabled devices.

In September 2019, Amazon launched many new devices achieving many records while competing with the world's smart home industry. The new Echo Studio became the first smart speaker with 360 sound and Dolby sound. Other new devices included an Echo dot with a clock behind the fabric, a new third-generation Amazon Echo, Echo Show 8, a plug-in Echo device, Echo Flex, Alexa built-in wireless earphones, Echo buds, Alexa built-in spectacles, Echo frames, an Alexa built-in Ring, and Echo Loop.

How it Works :

According to Adi Agashe, Program Manager at Microsoft, Alexa is built based on natural language processing (NLP), a procedure of converting speech into words, sounds, and ideas.

• Amazon records your words. Indeed, interpreting sounds takes up a lot of computational power, the recording of your speech is sent to Amazon’s servers to be analyzed more efficiently.

Computational power: refers to the speed that instructions are carried out and is normally expressed in terms of kiloflops, megaflops, etc.

• Amazon breaks down your “orders” into individual sounds. It then consults a database containing various words’ pronunciations to find which words most closely correspond to the combination of individual sounds.
• It then identifies important words to make sense of the tasks and carry out corresponding functions. For instance, if Alexa notices words like “sport” or “basketball”, it would open the sports app.
• Amazon’s servers send the information back to your device and Alexa may speak. If Alexa needs to say anything back, it would go through the same process described above, but in reverse order(source)

Analysis of an “order”

The above command has 3 main parts: Wake word, Invocation name, Utterance. (this part is extracted from Kiran Krishnan’s article)

• Wake word :
• When users say ‘Alexa’ which wakes up the device. The wake word put the Alexa into the listening mode and ready to take instructions from users.
• Invocation name :
• Invocation name is the keyword used to trigger a specific “skill”. Users can combine the invocation name with an action, command or question. All the custom skills must have an invocation name to start it.

Alexa “skills”: voice-driven Alexa capabilities.

• Utterance :
• ‘Taurus’ is an utterance. Utterances are phrases the users will use when making a request to Alexa. Alexa identifies the user’s intent from the given utterance and responds accordingly. So basically the utterance decides what user want Alexa to perform.

After, Alexa enabled devices sends the user’s instruction to a cloud-based service called Alexa Voice Service (AVS).

Think the Alexa Voice Service as the brain of Alexa enabled devices and perform all the complex operations such as Automatic Speech Recognition (ASR) and Natural Language Understanding (NLU).

Alexa Voice Service process the response and identify the user’s intent, then it makes the web service request to third party server if needed.

Amazon Alexa and Go :

Amazon’s predictive algorithm goes far beyond their website. They now possess a number of alternate purchasing options—such as through Amazon’s voice assistant, Alexa, or their new physical stores, Amazon Go—that make it easier to create a seamless shopping experience for their customers. Through products like Amazon Echo, Amazon makes it easier for customers to purchase products. but also gives itself other data points, such as what music they are listening to. These data points make it easier to give more holistic suggestions.

Then there is Amazon Go, the company’s foray into physical stores, which we wrote about here. These stores automatically keep track of every customer purchase better than any store currently in existence. As Forbes states, “Data from customers’ smartphone cameras tracks shopping activities and not only helps Amazon Go, but can also be shared with the machine learning team for continued development.”

The takeaway? No matter where they collect their data, Amazon creates as many touchpoints as possible to better understand customers and create fully holistic views of their behavior. Take this example from Forbes: “A customer can visit the Amazon Go store to get a few items for dinner, ask Alexa to look up a recipe and the product recommendation engine can determine that the customer likely needs to purchase a certain type of saucepan.” This integration allows Amazon to create an extremely detailed profile of every one of their customers, and provide multiple ways of delivering exactly what they need through whatever method the person desires.