Monthly Archives: October 2017

separator

Mechanical Engineering Challenges in Building a High Performance Hematology Auto-Scanner

separator

Shonit™ is a smart hematology analyser powered by advancements in artificial intelligence, image processing and cloud computing. One of the critical components of the solution is the smart scanner which enables to digitise the data efficiently for the AI models on the cloud. Today, I am going to talk about challenges involved in achieving the accuracy […]

Read More

SHONIT™ : A powerful yet simple tool for Pathologists

separator

I am a Consultant Pathologist and  since the last 9 years I have been working in the field of Hematology . In this last decade, I have seen several paradigm shift in the field of diagnostic pathology including a shift from conventional microscopy to automated analysers. Automation has taken several quantum leaps, right from its […]

Read More

An attempt to understand deep neural nets using the Information Bottleneck theory

separator

At the very core of Deep Learning lies a silent demon called the optimization problem; While deep learning has pretty much started reshaping businesses, business processes and our lives in various ways, many of us still don’t have a clear understanding about the algorithm(s) governing it. This article is an attempt to understand how Deep […]

Read More

Mysteries of the ROC curve

separator

Most Data Scientists would have come across the ROC (receiver operating characteristic) curve. It is often used to evaluate the performance of a binary classifier by measuring AUC (area under the curve) and to find an optimal probability threshold to get to the sweet spot in the sensitivity-specificity tradeoff. However, both theoretically and practically there […]

Read More