Applied Machine Learning is a Meritocracy

When making a start in a new field it is common to feel overwhelmed.

You may lack confidence or feel as though you are not good enough or that you are lacking some prerequisite.

You will explore these issues in this post and learn that such feelings can lead to actions that can consume a lot of time and resources and leave you feeling disappointed in yourself.

You will learn that there are many paths through the field of machine learning and that like programming, it is a meritocracy.

Dangerous Thinking

Feeling like you are not good enough or that you are lacking some skill that you think you must have before you can make a start in machine learning is dangerous thinking.

I think it is dangerous because it can lead you down paths that consume a lot of time, money and resources that may not be required.

For example, you may feel like you need to have a grounding in mathematics or or computer science or some other subject that you are not really passionate about. You may decide that you simply must:

  • Get a Degree: An undergraduate or postgraduate degree giving you a formal education in machine learning including any prerequisites defined by the institution.
  • Take a Mathematics Course: A course, free or otherwise that will teach you linear algebra or calculus.
  • Read a Textbook: A post graduate level textbook on machine learning that presupposes a level of prior training in probability theory and linear algebra.

The risk is that you feel like you need to achieve some minimum skill level before you can get started.

You defer getting started in machine learning to start learning that skill that you think is required, it is difficult. Really difficult.

Because it is hard and you are not passionate about it, you are more likely to throw in the towel, meaning you continue to deny yourself the permission to start in machine learning.

This path can work for some, but it’s exceedingly difficult.

It does not have to be so hard.

Many Paths

There are many paths through the field of machine learning.

There is a place for the degree, the maths course and the textbook, they might be further along the path for you, or on a different path.

many paths

Many Paths
Photo by keepitsurreal, some rights reserved

Machine learning is a multidisciplinary field, meaning that you have people coming to it with backgrounds all across the fields of science and engineering.

It also means that there is no archetype for the “machine learning practitioner“, although I do think programmers have an extraordinary opportunity in the field.

It is a relatively new field and much of the documentation is in the form of research papers and textbooks produced by academics. As such this colors the perception of the field as highly academic. This is the reason why there is a focus on theory over application and the perceived need for training required by academics like research methods and higher degrees.

The technicians path is applied and to start with experimentation and process, and maybe some programming. Be confident in this approach, it is valid, effective and a path followed by countless fellow programmers. Know your limitations and play to your strengths and the skills you have already acquired.

Meritocracy

Like the field of software development, the application of machine learning is a meritocracy. A meritocracy is a structure under which participants are valued based on their contributions or demonstrated achievement (merit).

Business, clients and employers care about your credentials, but only in as much as the results you have demonstrated you can achieve and that you can achieve for them. Degrees, other awards and working for fortune 500 companies are a symbols that can be used by others to short cut this determination, but that is all.

As a meritocracy, you must demonstrate you have merit. If you are looking for your skill to be recognized by others, then you must demonstrate and promote it. You can do that by participating in projects, competitions and completing small open projects and using outputs from such efforts as advertisements of your capability to your self and others.

In this post you learned that feeling overwhelmed in a natural feeling when starting a new technical discipline. You learned that these feelings of inadequacy can lead to dangers thoughts such as expending large time and resource costs pursuing a degree or education you think you need to have before you can get started.

You learned that there are many paths through the field of machine learning and that the empirical path of the programmer as the technician is valued. You also learned that machine learning like programming is a meritocracy and that if you persist and do good work, it will can be recognized and acknowledged.

Next Step

So what is the next step?

Perhaps this path of machine learning practitioner is for you:

24 Responses to Applied Machine Learning is a Meritocracy

  1. Justin January 23, 2014 at 12:52 am #

    Very good words of advice for people that want in but are standing on the sidelines. This can be applied to many areas as more and more fields are becoming about what you’ve done rather than your academic credentials. This includes all of IT, machine learning/data analysis, and robotics. I chose to go the formal education route, but I have seen many succesful people in CS and machine learning in particular with no degree or coursework in a related field. One of the best predictive modelers I know has a Bachelor’s degree in History. To anyone reading that is wondering what it takes to do machine learning: a strong desire is all you need to get started.

    • jasonb January 23, 2014 at 8:00 am #

      “a strong desire is all you need to get started”

      Right on Justin!

  2. Sid January 30, 2014 at 12:38 am #

    Excellent articles applicable widely to life in general. Kindly attend to typos.

  3. SaFa December 9, 2014 at 8:50 am #

    There is a french proverbe that resumes your post : “Vouloir c’est pouvoir”, which means: if you want it you can do it. all you need is passion and perseverance.

  4. Juichiro March 18, 2015 at 9:27 am #

    “Take a Mathematics Course: A course, free or otherwise that will teach you linear algebra or calculus.”

    That was me!

  5. Raimonds Vanags November 5, 2015 at 12:28 am #

    This is a really motivating post. Though I have a question: “Is there a difference in wages between those who have a degree and those who don’t even when their skills in machine learning are quite equal?”

    • Gilles Delpech February 22, 2018 at 12:17 am #

      If the employer is honest and competent there is no difference in salary. Employers currently prefer motivated and competent people than those who are career-oriented graduates. In France this is the current trend although graduates still have better wages than self-taught more motivated. But that’s starting to change.

  6. Jim Kitzmiller October 22, 2016 at 5:29 am #

    Jason, thank you for your unwavering encouragement. It’s become obvious that you really care about your students.

  7. Jesús Martínez February 21, 2018 at 6:10 am #

    Thanks for this encouraging post, Jason! As most things in life, ML is certainly a meritocracy. That’s why a college degree or an online course can take you so far. In the end, it’s what you can prove that’ll open the doors to success!

    • Jason Brownlee February 21, 2018 at 6:43 am #

      Absolutely!

      The challenge is to keep this in mind when others try to play the “credentials are the only authority” game.

  8. Okoh Emmanuel December 12, 2018 at 10:26 pm #

    You just nailed it,many thanks.

  9. Far McKon February 2, 2019 at 1:13 am #

    Justin;
    I’m sure you’re heart is in the right place, but to the millions of people who even the basics of education and core nutrition has failed since a young age, who are passed over to answer questions or assumed to know less, until the too have those assumptions internalized, this is all just trash.

    We don’t have an even starting ground, and the invisible hand of wealth and prejudice is already tipping the scale of ‘the ability to contribute’ by Kindergarden or first grade, and ins extremely hard to escape.

    Is there a hard, formal, qualifier for these positions, or these industries? No. And I’ve happily hired non-traditional people over and over again, with the same mixed results as formally educated people.

    Is the field really based on merit? Sure, but only as far as our society itself balances and allows fair chances to succeed. HINT: it does not.

    I’m glad to see the encouraging post, and I’m glad your rooting for the under dog. But they really don’t have a fair chance the deck is stacked against them. People are culture are biased against will have to work 2x work, to have 1/2 the accomplishments in any part of our society even ML as a ‘meritocracy’ field . To think otherwise is quite naive.

  10. nzo February 2, 2019 at 1:48 am #

    Or waiting hype down 😉

  11. Vasile Mironeasa March 26, 2019 at 7:20 am #

    Hi Jason !

    I say Jason because I consider you a very good friend . I am an old man and I want learn Machine learning for my nephews, they are young boys and I have know that they need to learn TI for the future. I must do thousands thanks to you because you help me to understand how work Machine Learning and maybe one day after times you shall be thankfull for your endeavour to help other peoples free. I think you Jason are a great man and a great friend !
    My respect to you Jason !

    • Jason Brownlee March 26, 2019 at 8:14 am #

      Thanks, it’s great to have you here Vasile!

      I’ll do my best to help.

  12. Vikram V June 1, 2019 at 12:04 pm #

    Hi Mr. Jason, I am a biology graduate, I’m planning to build my career in AI / ML… Pls guide me, like
    1.what kind of programming languages should I learn to understand and get used,
    2. is mathematics background compulsory to learn AI/ML
    3.as I can find many wings in AI and ML, pls suggest what among it best suits for a person like me without mathematics background

    Thank you

Leave a Reply