Bifurcation Diagram of the Logistic Map showing complexity evolving from simplicity: (credit:

Here’s a list of things that I learnt over the last 1 year all of which has had a profound impact on my thought process.

Relationships matter

This is one of those things that’s there in every leadership book and honestly who would disagree with this? Inadvertently, I’ve been an individual contributor for most of my life and therefore have never had the opportunity to think deeply about the importance of relationships and the dynamics of a team.

However, this year I feel like I’ve finally understood WHY relationships matter. Apart from the fact that nobody likes working with a douchebag, relationships…

Image by Gerd Altmann from Pixabay

Well designed products make you say — “Wow, they [company/product designers] have really thought about everything!”. This is the pinnacle that every product team strives to reach, to convert users into fans. Every article I’ve come across or PM I’ve spoken to, mentions that the path to achieve this is simply empathizing with your users.

Now, empathizing with users is undeniably hard — else we would be surrounded by beautifully designed products. To empathize you need to put yourself in the shoes of your users and have a very deep understanding of the problems and pain points they go through…

PM Modi addressing the nation about the lock-down

While watching PM Modi’s national address on Tuesday night, I couldn’t help but analyse the Prime Minister’s lock-down strategy from a product management lens. Here are 4 lessons to take away:

1. Minimum Viable Product/Pilot Program for gathering data

By now it’s pretty obvious that the voluntary curfew on Sunday was an attempt to understand how the public would react to a lock-down, what kind of measures the state and central government would need to take to enforce it and how to ensure that the essential services keep running.

This could very well have been the world’s largest MVP or Pilot experiment. In a situation where time is…

Photo by Hosea Georgeson on Unsplash

Convolutions have become a fundamental part of modern neural networks because of their ability to capture local information and reduce the number of parameters with weight sharing. Since almost all vision-based models (and a few NLP-models) use convolutions of one form or the other, it’s obvious that we would like to make these operations as fast as possible.

To emphasis the need for fast convolutions, here’s a profiler output of a simple network with a single 2D convolution layer followed by a Fully Connected layer:

TSNE Plot of MNIST Pseudo-Labeling

A few days ago I came across Yoshua Bengio’s reply to a Quora question — Why is Unsupervised Learning important?”. Here’s an excerpt of his reply:

To climb the AI ladder with supervised learning may require “teaching” the computer all the concepts that matter to us by showing tons of examples where these concepts occur. This is not how humans learn: yes, thanks to language we get some examples illustrating new named concepts that are given to us, but the bulk of what we observe does not come labeled, at least initially.

His reply makes a lot of sense both…

Stop overfitting!

As the saying goes, in this era of deep learning “data is the new oil”. However, unless you work for a Google, a Facebook or some other tech giant, getting access to adequate data can be a tough task. This is especially true for small companies operating in niche domains or personal projects that you or I might have.

In this blog, we’ll simulate a scenario where we only have access to a very small dataset and explore this concept at length. …

Anirudh Shenoy

Product @ Yellow Messenger | Breaking down Machine Learning & Data Science topics into simple concepts | | Views are personal |

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store