In recent years, there has been a big leap in machine learning due to the capabilities of graphic processing units (GPUs) evolving at lightning speed. Even though machine learning is never going to be exactly human-like, we are getting closer in a few areas, like imagenet large scale challenge, where the error rates are on par with or better than humans. A recent announcement of SAP human resources implementing bias filters is another great example of machine learning performing better than humans. Machine learning promises a next wave of innovations for every application, business process, and interaction of the enterprise with its customers.
Broadly, you can classify machine learning algorithms by:
- Rote learning – Memorizing facts or programmed
- Learning from instruction – Acquiring knowledge
- Learning from deduction – Inferring from its knowledge and leading to useful conclusions
- Learning by analogy – Transforming existing knowledge to form new knowledge
- Learning by induction – Making conclusions based on information collected (all statistical learning is about inductive reasoning)
- Adoptive learning – Automatically adjusting numerical parameters in a neural network
The building blocks of next-generation machine learning that I see are:
What do companies want from machine learning?
Feedback from various customers and considerable research insights all suggest that when it comes to machine learning and artificial intelligence, these are the top three value-adds enterprises expect:
- Improvement of customer experience
- Uplift of existing products and services
- Disruption with new business models
To disrupt, companies need to have a “machine learning-first” mentality to differentiate them in the marketplace.
Learn more about machine learning and artificial intelligence:
For more on predictive analytics, read the other blogs in our Predictive Thursday series.