Introduction
AI and Machine Learning are poised to make significant changes to the way many businesses operate. AI is transiting from just a research topic to the early stages of enterprise adoption. Tech giants like Google, Facebook and Microsoft have placed huge bets on Artificial Intelligence and Machine Learning and are already using it in their products. They have set up massive cloud platforms to enable other businesses to run their machine learning models at scale. As an AI practitioner, over the last couple of years I have seen an exponential increase in interest in applying AI to a variety of domains. Artificial intelligence (AI) in business is rapidly becoming a commonly-used competitive tool. Clearly, companies are past debating the pros and cons of AI. That puts AI in the short-list of technologies that your company should not just be watching, but actively exploring how to take advantage of. And business leaders agree based on the results of a Forrester survey below.
Difference between AI and Machine Learning
Many people including myself, use the terms AI and Machine learning quite interchangeably but this is not correct. And the term deep learning also gets thrown in the mix frequently! However they mean different things. The image below explains this well.
AI — Is the field dealing with making computer systems intelligent.
Machine Learning — Machine Learning is the learning in which machine can learn on its own without being explicitly programmed. It is an application of AI that provide system the ability to automatically learn and improve from experience.
Machine Learning is the main way in which AI is being advanced right now and hence the two terms are used interchangeably but AI is broader than Machine Learning. You can read more about their differences at this link.
Deep Learning — Deep Learning is Machine Learning done through neural networks. The design of neural network algorithms that can learn non linear relations between data has fuelled the advance in machine learning and hence AI over the last 5 years.
Why Machine Learning has become so powerful and popular
Advances in deep neural networks i.e deep learning has fuelled the boom in use of Machine Learning and AI over the last 5 years. Which begs the question — how and why now?
I think there are 3 main factors behind the massive adoption of deep learning over the last 5 years. All these factors have played together, boosting each other.
- Increase in the amount of data collected — Deep Neural Networks need a lot of data to shine. ML, prior to deep learning era was supervised and unsupervised techniques with big focus on feature engineering. Deep Learning has changed that. Neural Networks are automatic feature extractors learning important relations in data themselves, but they are data hungry. With the rise of internet, decrease in storage cost, cameras on cellphones etc, the amount of structured and unstructured data collected has boomed.
- Decrease in cost of computing — GPUs have allowed us to run powerful models on big data in a reasonable amount of time. This is critical. Computing power available per dollar has increased fairly evenly by a factor of ten roughly every four years in the last quarter of a century (a phenomenon sometimes called ‘price-performance Moore’s Law‘). The compute abilities of GPU and cost per FLOP for GPUs has also been decreasing year over year.
- Massive research in Algorithms — As there were early glimpses into the potential of neural networks by some breakthrough results through Hinton, Bengio etc, the machine learning community has increased the amount of research going into algorithms. Researchers have created powerful algorithms for solving complex problems like object detection, semantic segmentation, automatic summarization, machine translation etc.
All three factors above are still in play now and will continue to be for the next many years which leads me to believe that AI revolution that has been started is not just a boom/bust but will be fundamentally changing to the human civilisation.
Services we offer at Deep Learning Analytics
At Deep Learning Analytics, we work closely with our clients to help them deploy powerful machine learning models into their business. We have experience working with data in different formats like images, video clips, text, audio, time series etc and have worked on many different use cases. You can read more about our services at link. Contact us to learn more on how you can use AI to take your business to the next level