Humans Vs Machines
Lunar Lander from open ai gym crashes. Image created by Vinayak Patel

Humans Vs Machines

It's fascinating to see how humans and machines have certain common elements, especially in unsupervised machine learning.

Let's look at a few of them here.

Reward

Both humans and machines, let's refer to them as instances hereon, work their entire life under some form of tangible or intangible reward system.

These rewards either incentivise or penalise actions of these instances, and therefore are perceived to be the most effective way to guide and provide direction.

In most case, instances are greedy towards positive rewards, aka positive reinforcement.

Lunar Lander from open ai gym Landing perfectly. Image created by Vinayak Patel

Environment

Environment plays a very crucial part in providing reliable mode for information gathering or observation and learning opportunity.

Everyone appreciates a noiseless environment but instances in general just love to be in noiseless environment as they provide more accurate and reliable information.?

Any major change or discrepancy in the environment, can be referred to as noise here, which creates a drastic impact on both humans and machines, while they try to adapt as per their capacity and ability.

It's cheap or less expensive to produce noise in the environment, and more expensive to observe and learn from such environment.

We observe such unsustainable & destructive behaviour of creating noise in our world today where noise creators generally obtains a competitive edge, as they leave behind colossal debris of destruction around them.

Animation showing car making a success run in open ai environment. Animation created by Vinayak Patel.

Time

Though on different scale, both of them respect the dimension of time, i.e one tends to work on micro & nano scale, while we (humans) can hardly work on normal time scale :(

Thus machines are able to beat us every single TIME.

To be continued...


要查看或添加评论,请登录

Vinayak Patel的更多文章

社区洞察

其他会员也浏览了