The final destination to intuitively understand word embeddings… finally

Image for post
Image for post
Photo by Jon Tyson on Unsplash

Yo reader! I am Manik. What’s up?.

Hope you’re doing great and working hard for your goals. If not, it’s never late. Start now, at this very moment.

With this piece of information, you’ll walk away with a clear explanation on Sequence and Text processing for Deep Neural Networks which includes:

  1. OneHot Encoding with keras.
  2. What are word embeddings and their advantage over One-Hot encoding?
  3. What are word embeddings trying to say?
  4. A complete example of converting raw text to word embeddings in keras with an LSTM and GRU layer.

if you want to learn about LSTMs, you can go…


Image for post
Image for post
Gamma Ray burst! source: Google

Hello Stardust! Today we’ll see mathematical reason behind exploding and vanishing gradient problem but first let’s understand the problem in a nutshell.

“Usually, when we train a Deep model using through backprop using Gradient Descent, we calculate the gradient of the output w.r.t to weight matrices and then subtract it from respective weight matrices to make its(matrix’s) values more accurate to give correct output”

But what if the gradient becomes negligible?

When the gradient becomes negligible, subtracting it from original matrix doesn’t makes any sense and hence the model stops learning. This problem is called as Vanishing Gradient Problem.

We’ll first visualise the problem practically in our mind. We’ll train a Deep Learning Model with MNIST(you know this) dataset with 1,2,4 and 5 hidden layers and see the effect of using different architecture on the output(accuracy doesn’t increase always! 😵). …


Photo by Nasa on UnSplash
Photo by Nasa on UnSplash
Photo by NASA on Unsplash

Trust is fine but let’s use blockchain

Let’s dive deep to understand how blockchain maintains ‘trust’:

  • What is a Smart contract?
  • How Smart Contracts work?
  • Peculiar Smart Contracts from V Systems.
  • Hash Time lock contract
  • Non fungible contracts
  • Payment channel Contracts
  • Comparison between various smart contracts

What Is a Smart Contract?

A smart contract is a self-executing contract with the terms of the agreement between buyer and seller being directly written into lines of code. The code and the agreements contained therein exist across a distributed, decentralized blockchain network. The code controls the execution, and transactions are trackable and irreversible.

Smart contracts permit trusted transactions and agreements to be carried out among disparate, anonymous parties without the need for a central authority, legal system, or external enforcement mechanism. …


Image for post
Image for post

Hey there! What’s up? Congratulations! We just sold our data with ‘I agree’. Remember, anytime you clicked on ‘I agree’ button just because you didn’t want to read a 600 pages terms and conditions which if you’d read you won’t understand anyway and moved on to use that software. It’s alright you didn’t sell your data, you just sold your identity😐.

But what am I talking about?

Alright cool! Let’s see what’s TCP/IP and the ‘selling procedure’. ⚡️

TCP/IP, or the Transmission Control Protocol/Internet Protocol, is a suite of communication protocols used to interconnect network devices on the internet.

TCP/IP specifies how data is exchanged over the internet by providing end-to-end communications that identify how it should be broken into packets, addressed, transmitted, routed and received at the destination. TCP/IP requires little central management, and it is designed to make networks reliable, with the ability to recover automatically from the failure of any device on the network. …


a whole new world of possibility!

Image for post
Image for post

•Blockchain for database storage. 😍

Instead of storing data in native database forms we may use blockchain to store data which as a result becomes highly secure and decentralized(essence of blockchain).

👉 In general, traditional user accounts can be substituted by public-private keys and addresses in blockchains. Traditional databases are subject to strong access control, almost all data is restricted to authenticated accounts. Moreover, account creation is also of a centralized model in traditional databases, by which a database administrator grants the user an account for access. …


1 — The problem of very deep neural networks

The main benefit of a very deep network is that it can represent very complex functions. It can also learn features at many different levels of abstraction, from edges (at the lower layers) to very complex features (at the deeper layers). However, using a deeper network doesn’t always help. A huge barrier to training them is vanishing gradients: very deep networks often have a gradient signal that goes to zero quickly, thus making gradient descent unbearably slow. …


High level frameworks and APIs make it a lot easy for us to implement such a complex architecture but may be implementing them from scratch gives us the ground truth intuition of how actually ConvNets work.

Image for post
Image for post

- Outline of the Article

We’ll be implementing the building blocks of a convolutional neural network! Each function we’ll implement will have detailed instructions that will walk you through the steps needed:

  • Convolution forward
  • Pooling forward

We’ll use DLS jupyter notebooks to execute our modules. Check out DLS here. The fact is it comes with pre-installed libraries and frameworks required for Deep Learning. …


MIT researchers have been able to detect if you’re in depression using AI.

Image for post
Image for post

Hey everyone here! How are you all! Hell yeah !—

You’re a King/Queen ⚡️⚡️

Let’s dive now!

‘‘ The standard method for screening and diagnosing depression is the Patient Health Questionnaire (PHQ), which has been designed by the American Psychological Association . The questionnaire packages the DSM-IV depression criteria into a brief self-report instrument which asks whether the individual finds pleasure in doing things, feels down, tired, has a poor appetite, trouble concentrating, is slow or fidgety, and/or struggles to sleep.


What if I say you that your IoT devices are deceiving you?

Image for post
Image for post

Botnets

A botnet is an ‘internet’ of compromised systems which are controlled by the ‘herder’(owner of botnet). The systems can be compromised by any kind of malware which is executed in your system and allows someone else to control your system. Your may look to be working fine but in actual it may not!

About

Manik Soni

❤ Indian ❤ | SDE @ Amazon | Former Deep Learning Research Intern @ Bennett University | ACM ICPC 2017, 2018 Regionalist |

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