#machinelearning Archives - Page 2 of 2

App development by Machine Disrupting Mobile app Industry

Posted on 18/01/202012/06/2021 by Tushar

When we talk about the present, we don’t realize that we are actually talking about yesterday’s future. And one such futuristic technologies to talk about is Machine learning app development or use of AI in mobile app development services. Your next seven minutes will be spent on learning how Machine Learning technology is disrupting today’s mobile app development industry.

“Signature-based malware detection is dead. Machine learning based Artificial Intelligence is the most potent defence the next-gen adversary and the mutating hash.” ― James Scott, Senior Fellow, Institute for Critical Infrastructure Technology

The time of generic services and simpler technologies is long gone and today we’re living in a highly machine-driven world. Machines which are capable of learning our behaviors and making our daily lives easier than we ever imagined possible.

If we go deeper into this thought, we’ll realize, how sophisticated a technology has to be for learning on its own any behavioral patterns that we subconsciously follow. These are not simple machines, these are more than advanced.

Technological realm today is fast-paced enough to quickly switch between Brands and Apps and technologies if one happens to not fulfill their needs in the first five minutes of them using it. This is also a reflection upon the competition this fast pace has led to. Mobile app development companies simply cannot afford to be left behind in the race of forever evolving technologies.

Today, if we see, there is machine learning incorporated in almost every mobile application we decide to use. For instance, our food delivery app will show us the restaurants which deliver the kind of food we like to order, our on-demand taxi applications show us the real-time location of our rides, time management applications tell us what is the most suitable time for to complete a task and how to prioritize our work. The need of worrying over simple, even complicated things is ceasing to exist because our mobile applications and our smartphone devices are doing that for us.

Looking at the stats, they will show us that

AI and Machine Learning-driven apps is a leading category among funded startups
Number of businesses investing in ML is expected to double over the next three years
40% of US companies use ML to improve sales and marketing
76% of US companies have exceeded their sales targets because of ML
European banks have increased product sales by 10% and lower churn rates by 20% with ML

The idea behind any kind of business is to make profits and that can only be done when they gain new users and retain their old users. It might be a bizarre thought for mobile app developers but it is as true as it can be that Machine learning app development has the potential of turning your simple mobile apps into gold mines. Let us see how:

How Machine Learning Can Be Advantageous For Mobile App Development?

Personalisation: Any machine learning algorithm attached to your simpleton mobile application can analyze various sources of information from social media activities to credit ratings and provide recommendations to every user device. Machine learning web app, as well as mobile app development, can be used to learn.

Who are your customers?
What do they like?
What can they afford?
What words they’re using to talk about different products?

Based on all of this information, you can classify your customer behaviors and use that classification for target marketing. To put simply, ML will allow you to provide your customers and potential customers with more relevant and enticing content and put up an impression that your mobile app technologies with AI are customized especially for them.

To look at a few examples of big brands using machine learning app development to their benefits,

Taco Bell as a TacBot that takes orders, answers questions and recommends menu items based on your preferences.
Uber uses ML to provide an estimated time of arrival and cost to its users.
ImprompDo is a Time management app that employs ML to find a suitable time for you to complete your tasks and to prioritise your to-do list
Migraine Buddy is a great healthcare app which adopts ML to forecast the possibility of a headache and recommends ways to prevent it.
Optimise fitness is a sports app which incorporates an available sensor and genetic data to customise a highly individual workout program.

Advanced Search: Machine learning app ideas let you optimize search options in your mobile applications. ML makes the search results more intuitive and contextual for its users. ML algorithms learn from the different queries put by customers and prioritize the results based on those queries. In fact, not only search algorithms, modern mobile applications allow you to gather all the user data including search histories and typical actions. This data can be used along with the behavioural data and search requests to rank your products and services and show the best applicable outcomes.

Upgrades, such as voice search or gestural search can be incorporated for a better performing application.

Predicting User Behavior: The biggest advantage of machine learning app development for marketers is that they get an understanding of users’ preferences and behavior pattern by inspection of different kind of data concerning the age, gender, location, search histories, app usage frequency, etc. This data is the key to improving the effectiveness of your application and marketing efforts.

Amazon’s suggestion mechanism and Netflix’s recommendation works on the same principle that ML aids in creating customized recommendations for each individual.

And not only Amazon and Netflix but mobile apps such as Youbox, JJ foodservice and Qloo entertainment adopt ML to predict the user preferences and build the user profile according to that.

More Relevant Ads: Many industry experts have exerted on this point that the only way to move forward in this never-ending consumer market can be achieved by personalizing every experience for every customer.

“Most analog marketing hits the wrong people or the right people at the wrong time. Digital is more efficient and more impactful because it can hit only the right people, and only at the right time.” – Simon Silvester, Executive Vice President Head of Planning at Y&R EMEA

According to a report by The Relevancy Group, 38% of executives are already using machine learning for mobile apps as a part of their Data Management Platform (DMP) for advertising.

With the help of integrating machine learning in mobile apps, you can avoid debilitating your customers by approaching them with products and services that they have no interest in. Rather you can concentrate all your energy on generating ads that cater to each user’s unique fancies and whims.

Mobile app development companies today can easily consolidate data from ML that will in return save the time and money went into inappropriate advertising and improve the brand reputation of any company.

For instance, Coca-Cola is known for customizing its ads as per the demographic. It does so by having information about what situations prompt customers to talk about the brand and has, hence, defined the best way to serve advertisements.

Improved Security Level: Besides making a very effective marketing tool, machine learning for mobile apps can also streamline and secure app authentication. Features such as Image recognition or Audio recognition makes it possible for users to set up their biometric data as a security authentication step in their mobile devices. ML also aids you in establishing access rights for your customers as well.

Apps such as ZoOm Login and BioID use machine learning for mobile apps to allow users to use their fingerprints and Face IDs to set up security locks to various websites and apps. In fact, BioID even offers a periocular eye recognition for partially visible faces.

ML even prevents malicious traffic and data from reaching your mobile device. Machine Learning application algorithms detect and ban suspicious activities.

How are developers using the Power of Artificial intelligence In Mobile Application development?

After learning that what is machine learning app, let us take a look at the advantages of AI-powered mobile apps which are never-ending for Users as well as for mobile app developers. One of the most sustainable uses for developers is that they can create hyper-realistic apps using Artificial Intelligence.

The best usages can be:

Machine learning can be incorporated as a part of Artificial Intelligence in mobile technology.
It can be used for predictive analysis which is basically the processing of large volumes of data for predictions of human behaviour.
Machine learning for mobile apps can also be used for assimilating security and filtering out harmful data.

Machine learning empowers an optical character recognition (OCR) application to identify and remember the characters which might have been skipped from the developer’s end.

The concept of machine learning also stands true for Natural Language Processing (NLP) apps. So besides reducing the development time and efforts, the combination of AI and Quality Assurance also reduces the update and testing time phases.

What Are The Challenges with Machine Learning and their solutions?

Like any other technology, there is always a series of challenges attached to machine learning as well. The basic working principle behind machine learning is the availability of enough resource data as a training sample. And as a benchmark of learning, the size of training sample data should be large enough so as to ensure a fundamental perfection in machine learning algorithms.

In order to avoid the risks of misinterpretation of visual cues or any other digital information by the machine or mobile application, following are the various methods that can be used:

Hard Sample mining – When a subject consists of several objects similar to the main object, the machine is ought to confuse between those objects if the sample size provided for analysis as the example if not big enough. Differentiating between different objects with the help of multiple examples is how the machine learns to analyse which object is the central object.
Data Augmentation – When there is an image in question in which the machine or mobile application is required to identify a central image, there should be modifications made to the entire image keeping the subject unchanged, thereby enabling the app to register the main object in a variety of environments.
Data addition imitation – In this method, some of the data is nullified keeping only the information about the central object. This is done so that the machine memory only contains the data regarding the main subject image and not about the surrounding objects.

Which are the Best Platforms for the development of a mobile application with Machine Learning?

Azure – Azure is a Microsoft cloud solution. Azure has a very large support community, and high-quality multilingual documents, and a high number of accessible tutorials. The programming languages of this platform are R and Python. Because of an advanced analytical mechanism, the developers can create mobile applications with accurate forecasting capabilities.
IBM Watson – The main characteristic of using IBM Watson, is that it allows the developers to process user requests comprehensively regardless of the format. Any kind of data. Including voice notes, images or printed formats is analyzed quickly with the help of multiple approaches. This search method is not provided by any other platform than IBM Watson. Other platforms involve complex logical chains of ANN for search properties. The multitasking in IBM Watson places an upper hand in the majority of the cases since it determines the factor of minimum risk.

Tensorflow – Google’s open-source library, Tensor, allows developers to create multiple solutions depending upon deep machine learning which is deemed necessary to solve nonlinear problems. Tensorflow applications work by using the communication experience with users in their environment and gradually finding correct answers as per the requests by users. Although, this open library is not the best choice for beginners.

Api.ai – It is a platform that is created by the Google development team which is known to use contextual dependencies. This platform can be very successfully used to create AI based virtual assistants for Android and iOS. The two fundamental concepts that Api.ai depends on are – Entities and Roles. Entities are are the central objects (discussed before) and Roles are accompanying objects that determine the central object’s activity. Furthermore, the creators of Api.ai have created a highly powerful database that strengthened their algorithms.

Wit.ai – Api.ai and Wit.ai have largely similar platforms. Another prominent characteristic of Wit.ai is that it converts speech files into printed texts. Wit.ai also enables a “history” features which can analyze context-sensitive data and therefore, can generate highly accurate answers to user requests and this is especially the case of chatbots for commercial websites. This is a good platform for the creation of Windows, iOS or Android mobile applications with machine learning.

Some of the most popular apps such as Netflix, Tinder, Snapchat, Google maps and Dango are using AI technology in mobile apps and machine learning business applications for giving their users a highly customised and personalised experience.

Machine learning to benefit mobile apps is the way to go today because it loads your mobile app with enough personalization options to make it more usable, efficient and effective. Having a great concept and UI is one pole of the magnet but incorporating machine learning is going a step forward to provide your users with the best experiences.

How is MLOps Becoming a CPG Requisite?

Posted on 03/02/201919/06/2021 by Tushar

Before singing praises on the wonders that MLOps can do, let me shine some lights on a few new learnings, thanks to the post-pandemic crisis, that the companies across the globe have learned, especially the CPG.

Digital channels, or at least, digitization is a requisite. It is like Yoda said – do or don’t, there is no try! CPG companies who have toiled for years to see their brand sprout across the market witnessed a sharp decline in sales in a matter of months! Logistics became a big problem, yes, but their poorly implemented strategies were the actual Gordon Knot.
Today, consumers have a plethora of options. CPG firms cannot rely on their standard go-to-market strategies. How to connect with end-consumers? Now, there is an addendum to the question – how to connect with end-consumers and win them?
Companies across the world, irrespective of the size and market presence, have started moving from offline to online, in one or another way – Who does not think and act ‘online’ is up for a loss.
Health and wellness have become essential factors for the customers.
Millennials shop online; nothing drives them more except the cost to value. They want convenience, a sense of belonging, and too at lower prices.

Well, these are just the picture’s skeleton, the actual painting factors in multiple new developments, such as:

The emergence of small and medium-sized companies, focusing on target customers.
Manufacturers and distributors share data to streamline the logistics.
A surge in the usage of automated systems.
Shift towards local consumption.
E-Logistics companies collaborating with the retail stores.

The list is long.

A quick glimpse of how a product reaches the end consumer.

If you start eagle eyeing each step, you will find tremendous opportunities hidden in them.
Here are a few.

Opportunity 1 – Introduce a forecasting functionality based on new data.
Opportunity 2 – Bring in an integrated system that synchronizes the data across the process.
Opportunity 3 – Factor in self-learning feature that would comprise the market changes, customers’ buying behavior, etc.

You can cash on the above opportunities by implementing automation systems with various machine learning (ML) algorithms. You can introduce ML algorithms, such as:

Route optimization to make the best of the sales reps’ time.
Product optimization to solve the product mix problems.
NLP to analyze the consumers’ behavior.
Trade promotion optimization to plan and execute your trade spends.

Again, this list is endless.

So, you have the solution – build ML models and deploy them.
What are the critical roadblocks in adopting Machine Learning?

Problem 1 – Continuous delivery of value

Your team who works on the use case and writes the ML codes do not deploy them. Or at least, they do not have expertise on the delivery. So, relying your success entirely on the data science team can frustrate them and derail your ML journey.

Problem 2 – Composite and complex ML builds

Unlike traditional development builds, ML models make predictions by (indirectly) capturing data patterns without following the explicit rules. The ML build runs a pipeline that extracts patterns from the data to create model artifacts, making it far too complex and experimental.

Problem 3 – Productionizing ML models

Gartner figures 80% of the data science projects fail or never make it to production. To run the project successfully in a real-time environment, you need to find the problem situation and solve the problem when it occurs. You need to continuously monitor the process to find the difference between correct and incorrect predictions (bias) and know in advance how your training data will represent real-time data.

Areas to Focus: Identify Where Things Might Go Wrong for You

Beyond ML deployment difficulties and risks in the CPG, there are several other key areas where things can go wrong, so instead:

Find out the exact use case; if you try solving the wrong problems, things will go wrong.
Do not build models that do not map well to your business processes.
Check if you have any flawed assumptions about the data.
Convert the results of your experimentation into a production-ready model.

There are opportunities, there are problems, and there are ML models. However, the only requirement that delays the models’ deployments or often triggers performance issues is simply the lack of means to deploy it successfully. Anteelo can reduce your effort in solving the ML deployment challenges through its state-of-the-art ML Works platform that provides you the means to run thousands of ML models at scale and at once.

Is programming required for a Data Science career?

Posted on 18/01/201917/06/2021 by Tushar

This is a common dilemma faced by folks who are beginning their careers. What should young data scientists focus on — understanding the nuances of algorithms or faster application of them using the tools? Some of the veterans see this as an “analytics vs technology” question. However, this article agrees to disagree with this concept. We will soon discover the truth as we progress through the article. How should you build a career in data science?

Analytics evolved from a shy goose, a decade back, to an assertive elephant. The tools of the past are irrelevant now. Some of the tools lost market share, their demise worthy of case studies in B-schools. However, if we are to predict its future or build a career in this field, there are some significant lessons it offers.

The Journey of Analytics

A decade back, analytics primarily was relegated to generating risk scorecards and designing campaigns. Analytical companies were built around these services.

Their teams would typically work on SAS, use statistical models, and the output will be some sort of score -risk, propensity, churn etc. Its primary role was to support business functions. Banks used various models to understand customer risk, churn etc. Retailers were active in their campaigns in the early days of adoption patterns.

And then “Business Intelligence” happened. What we saw was a plethora of BI tools addressing various needs of the business. The focus was primarily in various ways of efficient visualizations. Cognos, Business Objects, etc. were the rulers of the day.

But the real change to the nature of Analytics happened with the advent of Big Data. So, what changed with Big data? Was the data not collected at this scale, earlier? What is so “big” about big data? The answer lies more in the underlying hardware and software that allows us to make sense of big data. While data (structured and unstructured) existed for some time before this, the tools to comb through the big data weren’t ready.

Now, in its new role, analytics is no more just about algorithmic complexity. It needs the ability to address the scale. Businesses wanted to understand the “marketed value” of this newfound big data. This is where analytics started courting programming. One might have the best models, but they are of no use unless you trim and extract clean data out of zillions of GBs of data.

This also coincided with the advent of SaaS (Software as a service) and PaaS (Platform as a service). This made computing power more and more affordable.

By now, there is an abundance of data clubbed with economical and viable computing resources to process that data. The natural question was – What can be done with this huge data? Can we perform real-time analytics? Can the algorithmic learning be automated? Can we build models to imitate human logic? That’s where Machine Learning and Artificial Intelligence started becoming more relevant.

What then is machine learning? Well, to each his own. In its more restrictive definition, it limits itself to situations where there is some level of feedback-based learning. But again, the consensus here is to include most forms of analytical techniques into it.

While the traditional analytics need a basic level of expertise in statistics, you can perform most of your advanced NLP, Computer vision etc. without any knowledge of their details. This is made possible by the ML APIs of Amazon/Google. For example, a 10th grader can run facial recognition on a few images, with little or no knowledge of Analytics. Some of the veteran’s question if this is real analytics. Whether you agree with them or not, it is here to stay.

The Need for Programming

Imagine a scenario where your statistical model output needs to be integrated with ERP systems, to enable the line manager to consume the output, or even better, to interact with it. Or a scenario where the inputs given to your optimization model change in real-time, and model reruns. As we see more and more business scenarios, it is becoming increasingly evident that embedded analytical solutions are the way forward. the way analytical solutions interact with the larger ecosystem is getting the spotlight. This is where the programming comes into the picture.

Part 1 of the Machine Learning Operations (MLOP) series

Posted on 13/01/201917/06/2021 by Tushar

Introduction to Machine Learning Operations

Machine learning – a tech buzz phrase that has been at the forefront of the tech industry for years. It is almost everywhere, from weather forecasts to the news feed on your social media platform of choice. It focuses on developing computer programs that can acquire data and “learn” by recognizing patterns and making decisions with them.

Although data scientists build these models to simplify and make business processes more efficient, their time is, unfortunately, split and rarely dedicated to modeling. In fact, on average, data scientists spend only 20% of their time on modeling; the other 80% is spent on the machine learning lifecycle.

Building

This exciting step is unquestionably the highlight of the job for most data scientists. This is the step where they can stretch their creative muscles and design models that best suits the application’s needs. This is where Anteelo believes that data scientists ought to spend most of their time to maximize their value to the firm.

Data Preparation

Though information is easily accessible in this day and age, there is no universally accepted format. Data can come from various sources, from hospitals to IoT devices; to feed the data into models, sometimes, transformations are required. For example, machine learning algorithms generally need data to be numbers, so textual data may need to be adjusted. Statistical noise or errors in data may also need to be corrected.

Model Training

Training a model means determining good values for all the weights and bias in a model. Essentially, the data scientists are trying to find an optimal model that can minimize loss – an indication of how badly the prediction is performed on a single example.

Parameter Selection

During training, it is necessary to select some parameters that will impact the prediction of the model. Although most are selected automatically, some subsets cannot learn and require expert configuration. These are known as hyper parameters. Experts trying to configure hyper parameters have to implement various optimization strategies to tune the hyper parameters.

Transfer Learning

It is quite common to reuse machine learning models across various domains. Although models may not be directly transferrable, some can serve as excellent foundations or building blocks for developing other models.

Model Verification

At this stage, the trained model will be tested to see if the validated model can provide sufficient information to achieve its intended purpose. For example, when the trained model is presented with new data, can it still maintain its accuracy?

Deployment

At this point, the model has been thoroughly trained & tested and has passed all requirements. The step aims to use this model for the firm and ensure that it can continue to perform with a live stream of data.

Monitoring

Now that the model is deployed and live, many businesses generally consider the process to be final. Unfortunately, this is far from reality. Like any tool, the model will wear out after use. If not tested regularly, it will provide irrelevant information. To make matters worse, since most machine learning models work in a “black box,” they lack the clarity to explain the model’s predictions, making the predictions challenging to defend.

Without this entire process, models would never see the light of day. That said, the process often weighs heavily on data scientists, simply because many steps require direct actions on their end. Enter Machine Learning Operations (MLOps).

MLOps (Machine Learning Operations) is a set of practices, frameworks, and tools that combines Machine Learning, DevOps, and Data Engineering to deploy and maintain ML models in production reliably and efficiently. MLOps solutions provide Data engineers, scientists, and engineers with the necessary tools to make the entire process a breeze. Next time, find out how Anteelo Engineers have developed a tool that targets one of these steps to make the lives of data scientists’ easier.

From machine intelligence to security and storage, AWS re:Invent opens up new options.

Posted on 10/01/201918/06/2021 by Tushar

Technology as an enabler for innovation and process improvement has become the catchword for most companies. Whether it’s artificial intelligence and machine learning, gaining insights from data through better analytics capabilities, or the ability to transfer data and knowledge to the cloud, life sciences companies are looking to achieve greater efficiencies and business effectiveness.

Indeed, that was the theme of my presentation at the AWS re:Invent conference: the ability to innovate faster to bring new therapies to market, and how this is enabled by an as-a-service digital platform. For example, one company that had an increase in global activity needed help to accommodate the growth without compromising its operating standards. Rapid migration to an as-a-service digital platform led to a 23 percent reduction in its on-premises system.

This was my first re:Invent, and it was a real eye opener to attend such a large conference. The week-long AWS re:Invent conference, which took place in November 2018, brought together nearly 55,000 people in several venues in Las Vegas to share the latest developments, trends, and experiences of Amazon Web Services (AWS), its partners and clients.

The conference is intended to be educational, giving attendees insights into technology breakthroughs and developments, and how these are being put into use. Many different industries take part, including life sciences and healthcare, which is where my expertise lies.

This slickly organized, high-energy conference offered a massive amount of information shared across numerous sessions, but with a number of overarching themes. These included artificial intelligence, machine learning and analytics; serverless environments; and security, to mention just a few. The main objective of the meeting was to help companies get the right tool for the job and to highlight several new features.

During the week, AWS also rolled out new functionalities designed to help organizations manage their technology, information and businesses more seamlessly in an increasingly data-rich world. For the life sciences and healthcare industry — providers, payers and life sciences companies — a priority is being able to gain insights based on actual data so as to make decisions quickly.

That has been difficult to do in the past because data has existed in silos across the organization. But when you start to connect all the data, it’s clear that a massive amount of knowledge can be leveraged. And that’s critical in an age where precision medicine and specialist drugs have replaced blockbusters.

A growing number of life sciences companies recognize that to connect all this data — across the organization, with partner, and with clients — they need to move to the cloud. As such, cloud, and in particular major services such as AWS, are becoming more mainstream. There’s a growing need for platforms that allow companies to move to cloud services efficiently and effectively without disrupting the business, but at the same time make use of the deeper functionality a cloud service can provide.

Putting tools in the hands of users

One such functionality that AWS launched this year is Amazon Textract, which automatically extracts text and data from documents and forms. Companies can use that information in a variety of ways, such as doing smart searches or maintaining compliance in document archives. Because many documents have data in them that can’t easily be extracted without manual intervention, many companies don’t bother, given the massive amount of work that would involve. Amazon Textract goes beyond simple optical character recognition (OCR) to also identify the contents of fields in forms and information stored in tables.

Another key capability with advanced cloud platforms is the ability to carry out advanced analytics using machine learning. While many large pharma companies have probably been doing this for a while, the resources needed to invest in that level of analytics has been beyond the scope of most smaller companies. However, leveraging an observational platform and using AWS to provide that as a service puts these capabilities within the reach of life sciences companies of all sizes.

Having access to large amounts of data and advanced analytics enabled by machine learning allows companies to gain better insights across a wide network. For example, sponsors working with multiple contract research organizations want a single view of the performance at the various sites and by the different contract research organizations (CRO). At the moment, that can be disjointed, but by leveraging a portal through an observational platform, it’s possible to see how sites and CROs are performing: Are they hitting the cohort requirements set? Are they on track to meet objectives? Or, is there an issue that needs to be managed?

Security was another important theme at the conference and one that raised many questions. Most companies know theoretically that cloud is secure, but they’re less certain whether what they have in place gives them the right level of security for their business. That can differ depending on what you put in the cloud. In life sciences, if you are putting research and development systems into the cloud, it’s vital that your IT is secure. But with the right combination of cloud capabilities and security functionality, companies can get a more secure site there than they would on-premises.

The conference highlighted multiple new functions and services that help enterprises gain better value from moving to the cloud. These include AWS Control Tower, which allows you to automate the setup of a well-architected, multi-account AWS environment across an organization. Storage was also on the agenda, with discussions about getting the right options for the business. Historically, companies bought storage and kept it on-site. But these storage solutions are expensive to replace, and it’s questionable whether they are the best way forward for companies. During the re:Invent conference, AWS launched its new Glacier Deep Dive storage facility, which allows companies to store seldom-used data much more cost effectively than legacy tape systems, at just $1.01/TB per month. Consider the large amount of historical data that a legacy product will have. In all likelihood, that data won’t be needed very often, but for companies selling or acquiring a product or company, it may be important to have access to that data.

One of the interesting things I took from the week away, apart from a Fitbit that nearly exploded with the number of steps I took in a day, was how the focus on cloud has shifted. Now the discussion has turned to: “How do I get more from the cloud, and who can help me get there faster?” rather than: “Is the cloud the right thing for my business?” Conversations held when standing in queues waiting to get into events or onto shuttle buses were largely about what each organization is doing and what the next step in its digital journey would be. This was echoed in the Anteelo booth, where many people wanted more information on how to accelerate their journey. One of the greatest concerns was the lack of internal expertise many companies have, which is why having a partner allows them to get real value and innovation into the business faster.

Order Cancellation Prediction: How a Machine Learning Solution Saved Thousands of Driver Hours

Posted on 09/01/201917/06/2021 by Tushar

‘Efficiency’ roots from processes, solutions, and people. It is one of the main driving forces leading to significant changes in the way companies work in the first decade of the 21st century. The following decennary further accelerated this dynamic. Now, post-COVID, it is vital for us to become efficient, productive, and environmentally friendly.

One of our clients manufactures and sells precast concrete solutions that improve their customers’ building efficiency, reduce costs, increase productivity on construction sites, and reduce carbon footprints. They provide higher quality, consistency, and reliability while maintaining excellent mechanical properties to meet customers’ most stringent requirements. The customers rely on their quality service and punctual delivery to receive products. This is possible because their supply chain model is simple. They prepare the order by date, call the driver the day before, and load the concrete the next morning. The driver delivers the exact specific product to the specified address.

However, a large percentage of customers cancel orders. One of the main reasons for the cancellation is the weather.

The client turned to Anteelo to provide an analytical solution for flagging such orders so that their employees do not have to prepare for such deliveries.

I’ll abridge the journey so far that it led to the creation of a promising solution.

How it all started?

One of the business units of the client suffered huge operational losses due to the cancellation of orders. Although the causes were(are) beyond their control, they always had(have) to compensate truck driver and concrete workers. To improve the demand and supply planning process’s efficiency, they had to encounter order cancellation risks. Though they might have increased their resource capacity by adding more people or working in shifts, this option may not have paved well in the long run. Apart from this, the risks may not have mitigated as anticipated, which might have further reduced the RoI.

Although they put forward various innovative ideas, the results did not reflect the expectations, resulting in the loss of thousands of drivers’ hours. Before deciding to use an analytical solution, they discovered that their existing system has two main shortcomings.

Extensive reliance on conventional methods for dispatch
Absence of a data-driven approach

Thus, they wanted to leverage a powerful ML-enabled solution to empower ‘order dispatching’ to effectively get ahead of order cancellation and minimize high labor costs.

Roadmap that led to the solution’s development

The analytics team from Anteelo pitched the idea of developing a pilot solution and executing it in the decided test market and then creating a full-blown working solution.

We used retrospective data in the sterile concept (the idea was to solve as many challenges as possible for POC (Proof of Concept)). Later, when the field team gave positive feedback, we planned to deploy a cloud-based working model with a real-time front-end. Next, measure its benefits in terms of hours saved in the next 12 to 24 months.

Proof of Concept (POC)

To reap the maximum benefits and minimize risks on the analytical initiative, we opted to start with the proof of concept (POC) and execute a lightweight version of the ML tool. We developed a predictive model to flag orders at risk of cancellation and simulated operational savings based on the weather and previous years’ data. We found that:

50% of orders were canceled each year
A staggering percentage of orders were canceled after a specific time the day before the scheduled delivery – ‘Last-minute cancellations.’
Because of these last-minute cancellations, hundreds of thousands of driving hours were lost

Creating the Most Viable Product (MVP)

Before we could go any further or zero down to the solution deployment, we had to understand the cancellation’s levers. And once the POC was ready, we decided to evaluate the results based on the baselines and expectations and compare them with the original goals. Next, we decided to proceed with the pilot test and modify the solution based on its result. Therefore, we selected a location and deployed some field representatives to provide real-time feedback and relied on our research for this purpose. The results (savings potential) were as follows:

Fewer large orders canceled
More orders canceled on Monday
When the temperature dropped to certain degrees, the number of cancellations increased
More than half of the last-minute cancellations were from the same customers
If a certain proportion of the orders were canceled at least one day in advance, the remaining orders were canceled at the last minute
On days with rain, the number of cancellations increased

Overall, order quantity, project, and customer behavior were the essential variables.

The MVP stage provided a staggering number, representing the associated monetary loss (in millions) due to the last-minute cancellations. The reasons behind such a grim figure were the lack of a data-oriented approach and prioritization method.

The deployed MVP helped reduce the idle hours. It helped flag the cancellations that we usually would have missed with our heuristic model. It also provided the market-wise potential, which we ultimately decided to roll out.

Significant findings (and refinements) in the ML model based on pilot test

Labor planning is a holistic process

An effective labor plan must deliberate factors other than the quantity (orders), such as the distribution of orders throughout the day, the value of the relationship with customers, and so on.

Therefore, the model output was modified to predict the quantity based on the hourly forecast.

Order quantity may vary with resource plan

‘Order quantity’ shows a considerable variation between the forward order book and the tickets, making it impossible to use it as a predictor variable.

Resources are reasonably fixed during the day

This contradicts one of the POC’s assumptions that resources will be concentrated in the market on a given day. This has led to corresponding changes in forecast reports, accuracy calculations, etc.

Building and Deploying a Full-blown ML-model at Scale

At this stage, we had the cancelation metrics, levers that worked, and exact variables to use in the solution. Now, the team has enough data to build an end-to-end solution comprising intuitive UI screens & functions, automated data flows, and model runs. And finally, measure the impact in monetary equivalent.

Benefits’ (Impact) Measurement

To turn the wheel and get it on track, we have to extract the model’s maximum value and evaluate it over time. We decided on two evaluation time metrics for measuring the impact.

Year-on-Year
Month-on-Month

The following is a summary table of improvements to key operational KPIs. Based on TPD change, the estimated savings are calculated based on the annual business volume.

	TPD	Location-specific	US
Metric value (YoY)	30% (up)	>$350k	>$3M
Metric value (MoM)	12% (up)	>$150k	>$3M

*data is speculative and based on the pilot run.

Predictive Model’s Key Features

Visual Insights
Weekly Model Refresh
Modular Architecture for seamless maintenance

Results

Reduced Deadheading
Streamlined dispatch planning
Higher Labor Utilization
Greater Revenue Capture

Why should you consider Anteelo’s ML/AI solutions?

We have successfully tested the pilot solution, and the model has shown annual savings of more than $3 million. Now, we will build and deploy the full version of the model.

Anteelo is one of the top analytics and data engineering companies in the US and APAC regions. If you need to make multi-faceted changes in your business operations, let us understand your top-of-mind concerns and help you with our unique analytics services. Reach out to us at https://anteelo.com/contact/.

MLOps: Is This the Only Way to Eat an Elephant?

Posted on 08/01/201917/06/2021 by Tushar

Managing ML production requires a combination of data scientists (algorithm procrastinators) and operations (data architects, product owners? Yes, why not?).

Operationalizing ML solutions in on-prem or cloud environments is a challenge for the entire industry. Enterprise customers usually have a long and random software update cycle, usually once or twice a year. Therefore, it is impractical to couple the deployment of the ML model with irregular update cycles. Besides, data scientists have to deal with:

Data governance
Model serving & deployment
System performance drifts
Picking model features
ML model training pipeline
Setting the performance threshold
Explainability

And data architects have enough databases and systems to develop, install, configure, analyze, test, maintain… the verb would keep on accumulating, depending on the ratio of the company’s size to the number of data architects.

This is where MLOps come in to rescue the team, solution, and the enterprise!

What is MLOps?

MLOps is a new coinage, and the ML community keeps on adding/ perfecting its definition (as the ML life cycle continues to evolve, its understanding is also evolving). In layman terminology, it is a set of practices/disciplines to standardize & streamline ML models in production.

It all started when a data scientist shared his plight with a DevOps engineer. Even the engineer was unhappy with the incumbent (inclusion of) data and models in the development life cycle. In cahoots, they decided to amalgamate the practices and philosophies of DevOps and ML. Lo and behold! MLOps came into existence. This may not be entirely true, but you have to give credits to the growing community of ML & DevOps personnel.

Five years ago, in 2015, a research paper highlighted the shortcomings of traditional ML systems (third reference on this Wikipedia page). Even then, the ML implementation grew exponentially. After three years of the research’s publication, MLOps became mainstream – 11 years after DevOps! Yes, it took this long to combine the two. The reason is simple – AI became mainstream only a few years back, 2016, 2018, or 2019 (the year is debatable).

MLOps Lifecycle

MLOps brings the DevOps principles to your ML workflow. It allows continuous integration into data science workflows, automates code creation and testing, helps create repeatable training pipelines, and then provides continuous deployment workflow to automate the package, model validation, and deployment to the target server. It then monitors the pipeline, infrastructure, model performance, and new data and creates a data feedback flow to restart the pipeline.

These practice involving data engineers, data scientists, and ML engineers enables the retraining of models.

All seems hunky-dory at this stage; however, in my numerous encounters with the enterprise customers, and after going through several use cases, I have seen MLOps, although evolutionary & state-of-the-art, failing several times in delivering the expected result or RoI. The foremost reason, often discovered, because of –

The singular, unmotivated performance monitoring approach
Unavailability of KPIs to set/measure the performance
And lack of threshold to raising model degradation alerts

In contrast, these are the technical hindsight that is often vindictive because of the lack of MLOps standardization; However, a few business factors, such as lack of discipline, understanding, resources, can slog or disrupt your entire ML operations.

Critical MLOps Roadblocks that Will Delay Your AI Journey in the Enterprise

Posted on 07/01/201917/06/2021 by Tushar

How do you retract the steps that led to the model’s creation, say, your data scientists are away for some reason?

How will you reproduce predictions to validate its outcome, say, someone shoots the question?

It is not just about resourcing data scientists, software developers, or data engineers to work in isolation to achieve the operationalization and automation of the ML lifecycle. It is about how the three can work in tandem as a unit. For this, the data’s quality or availability must remain identical across the process & environment to ensure the model performs on par with the set metrics. Again, the core problem boils down to operation and automation, which we diligently tried/try to address via MLOps.

To solve the problem’s crux, you first need to answer a few questions:

How do the three personas, i.e., data engineers, data scientists, and ML engineers, use different tools and techniques?
How do you collaborate on the ML workflow within and between teams?

As you cannot share models like other software packages, you need to share the ML pipeline that can reproduce and tune the model based on new data specific to the new environment/scenario. A ubiquitous work culture or norm in large enterprises is to have independent data science teams, and most of whom are engaged, day in and day out, on similar workflows.

Now, how do you collaborate and share the results?
When it comes to enterprise readiness, how do you plan data/ ML model governance while dealing with data & ML?
When you deal with specialized hardware, cost management comes into play as you have to compute with large amounts of GPU, memory, jobs that take a long time to run. Some of these jobs can take days or even weeks to run to get a good model. So how do you establish the trust?

Having insights, even dismal, will help you identify the real-time use cases and factor in an Enterprise AI plan.

What Next? Martian Version for Earthling Solution?

ML Works Will Just Do!

Most MLOps toolkit often focus on the technical aspect of the MLOps, while ignoring its real-life impact. Other factors that can weigh in its contribution are having a 360-degree view and control on the micro/macro aspects of the data science process.

At Anteelo, we have tried reordering the ML alphabets with our proprietary suite of toolkits, in which we take immense pride. We call it ML Works. The solution, which is cloud-agnostic and scalable, automates the model’s build, deploy, and monitoring processes, thereby reducing the need for larger teams.

Tag: #machinelearning