Getting Onboard the ML and AI bandwagon

Source: Deep Learning on Medium

We are in the midst of a Technological Disruption and the trend is bringing in the Next Revolution.

So much lies on the shoulders of the young turks emerging into this space, the next generation, our kids. Especially for those about to leave school soon or just have left their schools, this is the time to understand and harness what the revolution means to the society.

The current trends can be categorized into BigData, Cloud, IOT, Robotics, AR/VR and Machine Learning/Deep Learning/ AI. Here, I will be focusing more on the later.

To put things into context, IOT deals with the connectivity of the devices, vehicles and appliances via the internet to interact and exchange operational and functional data. As the data is growing huge beyond Petabyte scale, on-premise systems are becoming incapable of supporting this huge deluge. Here enters the Cloud. Major competitors are AWS, MS Azure and GCP.

Now how to handle the deluge of the data? This is where BigData enters the scenario. Data from Social Media, IOT sensors/devices, Enterprise data, Text data all are ingested collected, in different file formats, into a governed Data Lake, which also stores the Metadata of the data, without which it has no meaning. The Data Lake in itself is pretty complex. There are different forms in which data arrives, like batch mode, Near-real time mode (NRT) or is streamed Realtime. Different ways to store the data exist as well, based on frequency of read and write, latency of read depends on the functional or usage aspects of the data. This data is again used by multiple use-cases, for marketing, reporting and business dashboards, Analytical workloads and Advanced Analytics.

Robotics, AR/VR (Augumented/Virtual Reality) and even 3D Printing, are some of the emerging trends, and have garnered a lot of traction in the innovation space.

Coming to Machine Learning (ML), it is a subset of AI. It is the approach to predict outcomes/inferences from a comparatively small dataset with features identified through feature engineering. Deep Learning is a subset of ML. It requires huge amount of data for the learning and is compute intensive, ie., requires powerful CPUs or GPUs. Artificial Intelligence is the science of building Intelligent machines, such that machines can think like humans. Alongside these developments, it becomes highly important to engage into topics like Machine Morality. AI/ML requires a fair understanding of Mathematics and Statistics to get deeper into it.

So if you are an enthusiast and this interests you, get on to it. Though there are multiples courses and learning paths, paid and free, I will focus on the free sources required to get you started and create an expert out of you, at your own flexibility. Some good industry-renowned paid online courses are offered by Coursera, EDx (University courses), Udacity, and Direct online courses from Foreign Universities like MIT and Harvard. So as mentioned, here are the free learning paths below.

Maths and Statistics

Ml/ DL/ AI Courseware

  • Courses at Fast.AI
  • Kaggle Learning
  • EDX (Enrolment without a certificate)
  • Udemy (Courses are around Rs. 700–1500)
  • Youtube channels for further help like Analytics University, Siraj Raval

Blogs and Webinars

  • Kaggle
  • Medium
  • TowardsDatascience
  • AnalyticsVidhya
  • DataScienceCentral
  • And many more

If you still aren’t feeling motivated enough to pick your laptop up, search for the Kaggle Competitions on your phone and you might just fast track yourself up!

My note is, especially, for the Kids of today and the Adult Enthusiasts, but others can as well make an effort towards understanding what’s happening around us. So, Happy Learning!