I feel we need a revamp of our education system and perhaps even adult training. In our education system, we dogmatised our students NOT TO make mistakes. We punish them quite badly if they make mistakes, and if they made too many mistakes, and score low marks, the system together with the society may “stigmatise and condemn” them as irrecoverable “idiots”. As such, you see the strong pressure from parents, especially in Asian societies in making sure the children never make mistakes in our lives. Especially in Asian societies, we see tuition industry being every so vibrant. However, this post is not talking about the tuition industry and with this context, let us put a pin here.
For those of you who are familiar with machine learning, you probably know when we train any machine learning models, we have to define the loss/error function. Why is that the case? Because we know that the model will definitely make mistakes (i.e. low accuracy) at the start given the lack of information/examples so we wanted to measure it and know how much it has improved across the training examples.
The model will not be accurate in estimating the outcomes especially at the start of the model training as it needs more examples so as to find and estimate the relationships (between target and features) better. But more importantly, we tell the model whether it made mistakes in estimation and in the case of continuous target, we can even share how big the mistakes are. The model will then learn from these mistakes/errors and improve on its guess-timate. And after many iterations of updates in its parameters through learning from the mistakes and correct “answers” it made, it become more accurate.
If you have benefited much from my posts and like my newsletter thus far, consider making a “book" donation. Link at the bottom of the newsletter or here. :)
So here is the thing, if we are to say that most of these machine learning algorithms are used to imitate how humans learn, then why aren’t our education system build to allow mistakes to be made, feedback to students and guide them to use these mistakes to improve their learning then?
Through my training classes, I also observed there are participants who carried the “habit” of avoiding mistakes over to their own learning, in the end, paralysed them from moving forward. They forgot to learn, to experiment, to learn from mistakes and in the end missed a lot of opportunities to advance themselves. This will be detrimental to their individual career and especially in the Knowledge Economy where learning efficiency and innovation is going to count the most.
The two most important points I want to drive at for this post is:
We need to learn to experiment (which usually lead to mistakes) but take precautions to reduce foreseeable mistakes. Find learning environments that allows you to experiment and also minimize the impact made from mistakes.
Have a learning process, record any mistakes made and self-assessed the reason for the mistakes and how it can be avoided.
I think it is time we dropped the legacy of avoiding mistakes at all costs and rather be smart about mistake-making and have a process to learn from them instead. :)
Your thoughts? Please share them in the comments. :)