NN#7 — Neural Networks Decoded: Concepts Over Code

Author(s): RSD Studio.ai

Originally published on Towards AI.

This member-only story is on us. Upgrade to access all of Medium.

Till now we have studied various concepts introducing us to how neural networks are made and trained. Before moving on to more advanced concepts, we need one final chesspiece to close this intro. And that is once training is done how can we evaluate performance of our model. We have only seen loss functions but we cannot use them to define performance as different loss functions give us different values and they cannot be used as standard metrics to compare several models. For this, various different metrics are used based on type or category of neural networks. You might have seen them in news when new models are publicized like AP@50, latency e.t.c.

In this article, you will get an overview of these metrics completing our introduction to neural networks and enabling us to take the next steps!

You might have noticed that I have talked about “metrics” (not one!). That is because there cannot be one universal metric to know everything and many are used to get a broader picture of our model so that we can make a tradeoff between various aspects.

Why can’t we… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI