Data Science Basics for Executives

Machine Learning vs Deep Learning

So you are impressed with what you have heard so far about the latest and greatest deep learning and you would like to use it to solve your business problem. Well… the good news is you can — the bad news is that it might not help you solve your problem.

Let’s take a step back and try to understand the relationship between artificial intelligence, machine learning, and deep learning:

AI or Artificial Intelligence is a large universe of concepts that includes all ML or machine learning. In the same vein, ML encompasses all that is deep learning, or DL. Speaking in strict terms, one would define these in a way similar to:

Artificial intelligence: the intelligence displayed by machines that simulate human and animal intelligence.

Machine learning: a set of methods by which artificial intelligence systems learn by extracting patterns from data without being explicitly programmed. In machine learning, human intervention is needed to extract features/inputs.

Deep learning: a type of machine learning that requires fewer data preprocessing by humans and can often produce more accurate results than traditional machine learning approaches. It uses a layered approach where each hidden layer extracts some features until it can use the prediction or classification features. In general, these approaches are much more expensive than general machine learning — and take longer to be executed.

Speaking from a pragmatic viewpoint, I would say that AI is a technology that imitates living systems. Machine learning is how we train machines to imitate living systems. Deep learning is machine learning that can learn deeper insights with fewer assumptions but at a greater cost. I caution, though, that when someone sees the term AI, AI Engineer, etc., they are NOT doing AI as discussed above and what you may find written about in a Sci-Fi novel. Still, they are doing what is called narrow AI — applying a model or set of models to emulate a single action — such as answering questions.

So how do you know when to use machine learning vs. deep learning?

Use machine learning when:

1. Features/independent variables can be easily extracted for the problem.
2. We are trying to solve simple prediction, classification, time series, recommendation, sentiment analysis, etc.
3. We have fewer data available for training and testing.

Use deep learning when:

1) Solving problems related to computing vision, voice and speech recognition, natural language processing (NLP), and generation (NLG).

2) It's hard to extract features/ independent variables. For example, image features are pixel values. It's hard to identify edges, sides, etc. Deep learning takes care of feature extraction, with each hidden layer extracting some features.

3) We have a lot of data for training. The deep learning models overfit on training with less amount of data.

4) We have accelerated hardware compute available. The training of deep learning models needs a lot of processing power for training.

Final Thoughts

Both machine learning and deep learning solutions can be used to solve business problems.

While the inclination is always to use newer and more sophisticated solutions, it's important to analyze first the problem's complexity, available training data, and available compute for training.

An important point to remember while working on machine learning solutions — it’s not about using the best and latest algorithm to build a model. It’s still about solving a business problem.

A successful machine learning project’s foundation is understanding the business problem, quality of data chosen, and best-fit architecture to industrialize the model on-prem or in the cloud.

Now is the time to invest in machine learning solutions. At Hashmap, an NTT Data Company, we have the right expertise and tools to guide and help businesses succeed with machine learning solutions and initiatives. Reach out to us here.

Reach Out | Hashmap

Other Tools and Content For You

Data Science for Executives

Hashmap on Tap | Hashmap Podcast

Snowflake Utilities & Accelerators | Do more with Snowflake | Hashmap

Strategy and the Cloud — Delegate, Delegate, Delegate

HASHMAP

Jhimli Bora is a Cloud and Data Engineer with Hashmap, an NTT Data Company, providing Data, Cloud, IoT, and AI/ML solutions and consulting expertise across industries with a group of innovative technologists and domain experts accelerating high-value business outcomes for our customers. Connect with her on LinkedIn.