What is TensorFlow And How Is It Used For Machine Learning

What is TensorFlow And How Is It Used For Machine Learning

Python is traditionally very popular among developers for web development, scripting, process automation, and general applications. Yet, in recent times, Python programming language has become the go-to language for Machine Learning.

Python language has several libraries for Machine learning and Deep learning development. Many of which have become very popular. But in recent times, none have become as famous or widely as Tensorflow.

What is TensorFlow? Why is TensorFlow the most preferred library for Machine learning? How popular is Tensorflow? What are some of the use cases? These are a few of the many questions that one has and let’s try to understand them one at a time.

What is a Software Library?

A software library is a set of helper blocks or objects that is called upon by your application or program whenever there is a need for specific functionality. These blocks are called modules. They typically include configuration data, documentation, message templates, subroutines, classes, values or type specifications etc.

If you have read my previous article about the python frameworks, you may have a bit of confusion now. How is a library different from a framework?

In simple terms, a framework is the skeleton of an application and a library is a certain function in an application.

The key difference is “Inversion Of Control (IOC)”. Or rather who controls who. If your application calls for a library, you are in control. But with a framework, the control is inverted. It calls your application.

If you need a functionality for your code you can call it in a library. With a framework, however, you can insert your code only at certain places.

Now that we understand libraries and frameworks a bit, let’s move on to the main subject of this article. Python Library- TensorFlow.

There are several great libraries in Python for machine learning and deep learning programming. Libraries like Keras, Theanos, TensorFlow, and Scikit-Learn have made machine learning development relatively easy.

With these libraries, it is easy to gather data, train models, serve predictions and refine future results.

What is TensorFlow?

TensorFlow is an open source machine learning library developed by Google. With this library, it is possible for high-performance numerical computation.

Google uses machine learning in all of its products to improve the search engine, translation, image captioning, recommendations etc. A few years ago Google found out that deep learning outperforms all other machine learning algorithms when given huge data sets.

Google wanted to use these deep neural networks to improve its services. Together with a team of researches and developers, they built an Artificial Intelligence Model called Tensorflow.

Why is TensorFlow Popular?

If you a large set of data or if you want to build the most advanced AI system for your brand, there is no better option than TensorFlow. It has a very great feature in it called, Tensorboard. This feature gives developers a visual representation of complex computations – graphs. With Tensorboard it is easier to debug the program and scale and deploy. TensorFlow runs both on CPU and GPU.

You can easily visualize each & every part of the graph which is not an option while using Numpy or SciKit. In the pre-Tensorflow age, there were some good ML libraries with good functionalities and methods. However, they were not as comprehensive as TensorFlow.

Great potential and continuous support and easy readability meant TensorFlow was the one-stop solution for all ML requirements.

Tensorflow attracts the largest popularity on GitHub compare to the other deep learning framework.

Use Cases of TensorFlow

TensorFlow is mainly used for the following applications in machine learning,

  • Classification
  • Perception
  • Understanding
  • Discovering
  • Prediction
  • Creation

From discovering new exoplanets to screen patients for diabetic retinopathy, Tensorflow has seen several useful applications.


PlantMD is an app that lets you detect diseases in plants using TensorFlow. The app aims to assist farmers to grow better cassava, a crop in Africa that provides food for over half a billion people daily. After annotating thousands of cassava plant images, identifying and classifying diseases they trained a machine learning model using TensorFlow. Once the model was trained to identify diseases, it was deployed in the app.

Farmers can wave their phone in front of a cassava leaf and if a plant had a disease, the app could identify it and give options on the best ways to manage it.

Food Industry

Safety and quality have always plagued the food industry. Artificial Intelligence (AI) enabled technology has a place in many restaurants today.

Nowadays in many municipal health agencies, various kind of facial and object recognition is used. In the kitchen of many restaurants, AI enabled camera is used. With the help of this, the owner of the restaurant can make sure that the chef wears masks for protecting hair fall on the food.

One interesting development is a food detection app that is currently in development.

A renowned Japanese food processing company, Kewpie Corporation, uses AI enabled TensorFlow machine learning. It has the capacity to detect various anomalies in food.

Banking & Finance

Companies like Mastercard are using ML for their ‘Decision Intelligence’ projects. It helps them discover patterns from historical shopping and spending habits of cardholders to detecting fraudulent activities. Many finance companies like Zest Finance and Destacame use ML to quickly complete their underwriting tasks or identify new credit-worthy borrowers.

In many cases, Tensorflow plays an important role in developing key machine learning models.

Pharma & Healthcare

Machine learning has the capability to predict healthcare outcomes. For example, the likely time a patient will be discharged, the propensity for a heart attack and a wide range of other capabilities.

It is possible to find patients exhibiting similar characteristics to other patients who have been diagnosed for a particular disease in the past. For instance, using classification and regression algorithms on data of patients diagnosed for AIDS or Cancer, an ML model can be implemented to confirm if a potential patient is infected with the same.

Machine learning models build with TensorFlow also enables personalization of patient treatment and better management of hospital resources.


The Iowa Department of Transportation teamed up with researchers at Iowa State University to use machine learning, including TensorFlow, to provide insights into traffic behavior in Iowa.

Iowa State’s technology helps analyze the visual data gathered from stationary cameras and cameras mounted on snow plows. They also capture traffic information using radar detectors. Machine learning transforms that data into conclusions about road conditions, like identifying congestion and getting first responders to the scenes of accidents faster.

NTT DOCOMO, Japan’s largest mobile service provider, launched a demand forecasting service for taxi operators. The service collects real-time people density from mobile phones and runs data analytics with a deep learning model on TensorFlow. The model has the ability to predict how many possible riders could be waiting in each block or street in the next 30 minutes, with 93-95% accuracy.


Paired with image classification models, TensorFlow can help to analyze satellite / aerial / street view imagery of buildings to deliver risk-related details of property.

AXA is working on complementing the traditional list of relevant risk factors (e.g., driver age) through new items such as real-time vehicle diagnostics and maintenance results. The collected data is processes through the open-source deep-learning framework TensorFlow.

Some of the popular companies that use TensorFlow are, Airbnb, AMD, Dropbox, Uber, Coca-Cola, etc.


As TensorFlow is an open source library, we will see many more innovative use cases soon.This will influence one another and contribute to Machine Learning technology.