As a part of my project, I want to train an ASR system using teacher-student learning by the encoder of a TTS system. So, I have a TTS system that takes audio as input, not text. I try to implement an article for this purpose and I can't understand the author's meaning in this section.
The description is: " The text/speech domain discriminator takes each frame of spectral or character embeddings as input and is a 4-layer FC neural network (256→512→512→2) ".
While the student network gets audio and the teacher gets text as input and both outputs are embeddings with (#minibatch, #chars of input, #embdding dim=512).
By this description, what is the input of the discriminator? input size of the discriminator is (#minibatch, #chars of input, #embdding dim=512) or (#minibatch, #embdding dim=512)?
I attached the image of the training diagram for better understanding.
I very thankful, If somebody can help me.
I'm reading the paper "Concerto: A High Concurrency Key-Value Store with Integrity
".
For memory checking, they said they used AES-based pseudo random function, which has two input and outputs hash.
enter image description here
My question is: How to implement this AES-based pseudo random function in C or Go? Is this just use AES to encrypt？ Someone can give me an example of implementation.
it is a question I have. I am not expert in this subject, so please, be kind on the answer.
I understood the homomorphic encryption process allow to read a message as if it has been decrypted, but it will do so without removing the protective layer that the encryption process placed on it.
Let's suppose the word "TESTE" is cryptographed, and a homomorphic encryption is done on that encrypted word.
My question is:
Homomorphic will understand the "meaning" of encrypted text? Homomorphic will know that the encrypted word is also "TESTE" ?
Thank you.
Let me give you a different example. I'm not sure the example can be implemented with today's systems. But it illustrates it anyway:
Party A has 10 numbers and encrypts them.
The encrypted numbers are given to party B.
Party B calculates the sum of the 10 numbers. The result is encrypted data.
Party A is given the encrypted result.
Party A decrypts the result.
The main feature is that party B does not need to decrypt the 10 numbers. Futhermore, the encryption is maintained throughout the entire sum calculation. Therefore, party B has neither knowledge of the input numbers nor of the calculated sum because all operations are carried out on the encrypted data.
I understood the homomorphic encryption process allow to read a message as if it has been decrypted.
No. Homomorphic encryption is public key encryption that allows someone to evaluate (technically circuit evaluation) on the encrypted data without accessing the data. The good side, the client can give the heavy process to the cloud without thinking that it's data is compromised as long as the scheme is not broken.
To understand the FHE, we can look at the textbook RSA, which has no padding. Textbook RSA enables multiplication that is if you multiply two the ciphertexts then decrypt you will get the multiplication of the plaintexts. So if you want to multiplication of your data on the cloud just send your data encrypted with RSA.
RSA only multiplies but there is no other operation and this called partially homomorphic.
There are other public key cryptosystems that support only one operation, for example x-or. This can be used to verify fingerprints on the cloud without revealing the data to the cloud.
If two operations are performed then it is called Fully-Homomorphic and we can build an arbitrary circuit, in theory.
The main idea is semantically encrypting your data (input for the calculation), then send the circuit (the operation you want to be performed) and send to the cloud to calculate with your public key. The cloud calculates the circuit and returns back to you. Only, you can decrypt with your private the return to get the result.
The main point is the calculation is done without accessing the data. As long as the Cryptographic primitives are not broken, no one will access your data.
Note: The breakthrough of Gentry's seminal work is finding a novel way to deal with the noise which is doubles with multiplication.
I am trying to create end to end SMS encryption application, but don't want to use standard encryption algorithm.The idea is to convert the text of message to completely different meaningful text so that over the network it doesn't seem encrypted. I am assuming messages only in English language.
In implementation part, I have first compressed the message using huffman encoding which gives me compressed stream of bits. Now for encryption I don't have any idea. Is it possible to build a dictionary of some random text or what other way can be used for getting such encryption? Decryption at other receiving end is also required.
Are you asking for a a codebook? If you have access to an English language dictionary of 65536 entries for example, you can take every 16-bits of your message as an index into this table to get a word. Good luck converting this into a real cryptosystem.
What is the difference between encoding and encryption?
Encoding transforms data into another format using a scheme that is publicly available so that it can easily be reversed.
Encryption transforms data into another format in such a way that only specific individual(s) can reverse the transformation.
For Summary -
Encoding is for maintaining data usability and uses schemes that are publicly available.
Encryption is for maintaining data confidentiality and thus the ability to reverse the transformation (keys) are limited to certain people.
More details in SOURCE
Encoding:
Purpose: The purpose of encoding is to transform data so that it can be properly (and safely) consumed by a different type of system.
Used for: Maintaining data usability i.e., to ensure that it is able to be properly consumed.
Data Retrieval Mechanism: No key and can be easily reversed provided we know what algorithm was used in encoding.
Algorithms Used: ASCII, Unicode, URL Encoding, Base64.
Example: Binary data being sent over email, or viewing special characters on a web page.
Encryption:
Purpose: The purpose of encryption is to transform data in order to keep it secret from others.
Used for: Maintaining data confidentiality i.e., to ensure the data cannot be consumed by anyone other than the intended recipient(s).
Data Retrieval Mechanism: Original data can be obtained if we know the key and encryption algorithm used.
Algorithms Used: AES, Blowfish, RSA.
Example: Sending someone a secret letter that only they should be able to read, or securely sending a password over the Internet.
Reference URL: http://danielmiessler.com/study/encoding_vs_encryption/
Encoding is the process of transforming data so that it may be transmitted without danger over a communication channel or stored without danger on a storage medium. For instance, computer hardware does not manipulate text, it merely manipulates bytes, so a text encoding is a description of how text should be transformed into bytes. Similarly, HTTP does not allow all characters to be transmitted safely, so it may be necessary to encode data using base64 (uses only letters, numbers and two safe characters).
When encoding or decoding, the emphasis is placed on everyone having the same algorithm, and that algorithm is usually well-documented, widely distributed and fairly easily implemented. Anyone is eventually able to decode encoded data.
Encryption, on the other hand, applies a transformation to a piece of data that can only be reversed with specific (and secret) knowledge of how to decrypt it. The emphasis is on making it hard for anyone but the intended recipient to read the original data. An encoding algorithm that is kept secret is a form of encryption, but quite vulnerable (it takes skill and time to devise any kind of encryption, and by definition you can't have someone else create such an encoding algorithm for you - or you would have to kill them). Instead, the most used encryption method uses secret keys : the algorithm is well-known, but the encryption and decryption process requires having the same key for both operations, and the key is then kept secret. Decrypting encrypted data is only possible with the corresponding key.
Encoding is the process of putting a sequence of characters into a special format for transmission or storage purposes
Encryption is the process of translation of data into a secret code. Encryption is the most effective way to achieve data security. To read an encrypted file, you must have access to a secret key or password that enables you to decrypt it. Unencrypted data is called plain text ; encrypted data is referred to as cipher text
Encoding is for maintaining data usability and can be reversed by employing the same algorithm that encoded the content, i.e. no key is used.
Encryption is for maintaining data confidentiality and requires the use of a key (kept secret) in order to return to plaintext.
Also there are two major terms that brings confusion in the world of security Hashing and Obfuscation
Hashing is for validating the integrity of content by detecting all modification thereof via obvious changes to the hash output.
Obfuscation is used to prevent people from understanding the meaning of something, and is often used with computer code to help prevent successful reverse engineering and/or theft of a product’s functionality.
Read more # Danielmiessler article
See encoding as a way to store or communicate data between different systems. For example, if you want to store text on a hard drive, you're going to have to find a way to convert your characters to bits. Alternatively, if all you have is a flash light, you might want to encode your text using Morse. The result is always "readable", provided you know how it's stored.
Encryption means you want to make your data unreadable, by encrypting it using an algorithm. For example, Caesar did this by substituting each letter by another. The result here is unreadable, unless you know the secret "key" with which is was encrypted.
I'd say that both operations transform information from one form to another, the difference being:
Encoding means transforming information from one form to another, in most cases it is easily reversible
Encryption means that the original information is obscured and involves encryption keys which must be supplied to the encryption / decryption process to do the transformation.
So, if it involves (symmetric or asymmetric) keys (aka a "secret"), it's encryption, otherwise it's encoding.
Encoding -》 example data is 16
Then encoding is 10000 means it's binary format or ASCII or UNCODED etc
Which can be read by any system eassily and eassy to understand it's real meaning
Encryption -》 example data is 16
Then encryprion is 3t57 or may be anything depend upon which algo is used to encryption
Which can be read by any system eassily BUT ony who can understand it's real meaning who has it's decryption key
These are little bit different from each other. The encoding used when we want to convert text in a specific computer coding technique and in the encryption we hide data between a specific key or text.
Encoding is process of transforming given set of characters in relevant accepted format, take this question's URL,
This is what we see -->
hhttps://stackoverflow.com/questions/4657416/difference-between-encoding-and-encryption
Over transmission this will be transformed to -->
https%3A%2F%2Fstackoverflow.com%2Fquestions%2F4657416%2Fdifference-between-encoding-and-encryption
^ is example of URL encoding using ASCII char set where,
: = %3A
/ = %2F
The reverse of Encoding is Decoding to original form and with given ASCII standard.
Encryption is process of converting plane text to cipher text so only authorized party can decipher it.
For example a simple HELLO is encrypted into KHOOR if just 3 characters are shifted.
p.s. Encoding (to code in some form) is form of encryption. :)
what-is-encryption
Encryption converts data to non-readable format (Possibly containing special non-readable characters).
Encoding helps to convert that data to readable format (characters) so that it can be stored for future use i.e. possibly during decryption.