ML vs DL vs DS and where they meet

Machine Learning
Machine learning is a branch of artificial intelligence (Ai) that is founded on the creation of algorithms that can process data and learn on their own. The entire approach relies on the fact that it is more efficient to teach a computer how to learn, rather than program it to perform each and every one of the required tasks that are part of a larger goal.
There are myriad applications for machine learning, and it is easy to find a few examples that are growing in popularity. The first is the rise of the virtual assistant, such as Amazon Alexa or Apple Siri. These systems employ learning algorithms to fine-tune or personalize the results of requests from individual users. As the system learns more about the user’s habits, it can better handle requests that contain ambiguity.
Another popular application is face recognition, where a still picture can be used as input into a system that will identify the people depicted within it. Social media services such as Facebook are capable of analyzing pictures and naming friends in a photograph. Similar algorithms are used, for example, to find and suggest people that you may know, or what jobs you may be a good candidate for.

Deep Learning
Deep learning algorithms are a branch off the broader field of machine learning that uses neural networks to solve problems. A neural network is a framework that combines various machine learning algorithms for solving certain types of tasks. A deep learning system is essentially a very large neural network that is trained using a very large amount of data.
There are different types of deep learning architectures, and it is not uncommon to hear about the use of a recurrent neural network (RNN) or a convolutional neural network (CNN). What is less often discussed are the internal workings. The word “deep” refers to the number of layers or points of transformation, that is contained within the framework. As the input traverses these layers it is made more abstract, terminating in the output layer. It is at this stage a prediction is made based on the original input.

Deep learning is currently being used in many complex tasks. One well-known example is Google Translate, which is capable of translating written text between more than 100 languages. Looking forward, deep learning will be applied in technologies such as: finance, autonomous vehicles, and healthcare.

Data Science
Data science is not a single technique or approach. Rather, it is a catch-all term that refers to several disciplines. This includes machine learning, data mining, data analytics, and statistics. Moreover, it encompasses the tasks related to working with big data, such as the process of extracting, transforming, and loading (ETL) data into storage repositories.
The primary goal of data science is to make sense of data. Gaining this understanding is a multi-step process. Depending on the specifics of a particular project, this may include the collection and processing of large amounts of data. If on the other hand the loading of data has already been completed, it is also squarely within the domain of data science to perform predictive analytics using tools such as machine learning algorithms and deep learning neural networks.

The Intersection of Machine Learning and Data Science
With machine learning fully contained within the field of data science, it is worthwhile to consider its role in the bigger picture. Because data science is multidisciplinary, it draws upon many tools that are outside of the machine learning scope. While pattern recognition and other data mining algorithms are common tasks that are undertaken by a data scientist, they also engage in other work that includes the use of visualization and applied statistics.
A data scientist will make use of tools for collecting, cleaning, transforming, and storing data. Regardless of the process or tools used, these steps are performed ahead of the analytics. Once the data has been fully pre-processed and is ready for analysis, the machine learning algorithms can be called upon to build predictive models for regression or classification tasks.

When the machine learning phase is complete, the job of the data scientist continues. Predictive models will be compared and analyzed, and the results reported. Furthermore, the models themselves may be part of the next stage in the exploratory or analytical process. All of this remains within the scope of data science.

Where Deep Learning Meets Machine Learning
Noted earlier is that deep learning is a subset of machine learning. In fact, the terms are sometimes used interchangeably, and this is because they function similarly. The difference from a practical standpoint lies in their capability, which affects their overall contribution to the model.
A traditional machine learning algorithm requires input from the user to help guide the process. For example, heuristics can be programmed to assign a score related to how good a potential solution is. If the model performs inadequately then it is generally up to the user to adjust the relevant parameters and try again. In a deep learning system, this adjustment is not required. The algorithms are capable of scoring the results on their own and make adjustments accordingly.
Google’s AlphaGo project involves a deep learning system that was tasked with learning the board game, Go. It has garnered international attention after successfully competing against world-ranked players. To the surprise of many, AlphaGo made use of new and inventive moves that have transformed the way the game is being played. The system is credited by some as bringing many new elements to the game, and inspiring players of all levels to vary their personal play style.

In a nutshell, data science represents the entire process of finding meaning in data. Machine learning algorithms are often used to assist in this search because they are capable of learning from data. Deep learning is a sub-field of machine learning but has improved capabilities. Many experts agree that deep learning has the potential to become the backbone of true artificial intelligence or strong AI.