The luring of machine learning

Big data is a new name for new form of tasks data scientists apply to pour some reason out of the big mud of numbers or text. Machine learning trend is mainly naming scientists for providing insights for those with the data.

I don't think that the “deep learning” new trend will prevail. Much of the current enthusiasm for big data focuses on technologies that make taming it possible, Hadoop, Spark and related open-source tools, cloud computing, and data visualization. While those are important breakthroughs, at least as important are the people with the skill set to put them to good use. On this front the shortage of data scientists and lack of breakthrough algorithms is becoming a serious constraint in some sectors.

If capitalizing on big data depends on hiring scarce data scientists, then the challenge for managers is to learn how to identify that talent and make it productive.  More than anything, what data scientists do is make discoveries while swimming in data. Data scientists realize that they face technical limitations, but they don’t allow that to slow their search for novel solutions.

At the end of the day data scientists using machine learning techniques will not prevail for replacing the need for the human factor. ML is not magic. it is hard to train a ML algorithm and build a good predictive model out of a relatively small samples or to give a good prediction on a future input value. it’s a fact! How do you make an 70-80% accurate model useful? Until we're replaced by robots, this is going to be the machine learning challenge of the next decade.