ML?
mobile legends
ML?
Coined by Arthur Samuel from IBM in 1958, he defined ML as
"The field of study that gives computers the ability to learn without being explicitly programmed."
Despite it being short and sweet, this can be problematic. 🙉
Learning is an extremely difficult task. To teach learning, it implies that we already know what learning is. Well, you probably believe that we know what it is, but to peep into this rabbit hole, we are probably the only animals that are doing this right (besides the animals in the zoos performing various stunts). In fact, we are so good at this that we are able to learn from our own's learning. We are the outliers, and not the norm.
Secondly, learning has been intuitive to us from birth. It may be inane to drive this, but we weren't taught learning. We were merely taught of various subject matters from young. To those whom did not possess (or even display) this unknown ability of learning will thus be labelled as having learning disabilities.
So, what then do we know about learning?
Well, nothing significant. Hence, it can be very empty to claim that we can ever teach a computer to learn. Heck, it can be even difficult to program a robot to perform something of our desire.
Next, the more favored and popular definition across the board is from Tom M. Mitchell, with
A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P if its performance at tasks in T, as measured by P, improves with experience E.
To break it down:
Experience (E) : Input (Data)
Tasks (T) : Output (Patterns/Trends)
Performance (P) : Minimize error
Now, we have something to work with! 🤤
Materials
This GitBook will an elaboration in reference to Caltech ML Lecture 1. Before you carry on, it is recommended to
Do take a look at the lecture video 🐱👤
I've attempted to annotate the slides with as much as information as possible, do have a look! ✍
Wikipedia is good, do use it often 👍
Last updated
