David McCall’s Post

A good thing done by AI: turning 5,000 public domain ebooks into audio books that sound like they were read by humans. I’ve sampled a couple, and the quality is quite impressive. Books here: aka.ms/audiobook Academic paper here: https://lnkd.in/e2PhuaUG

Listen to masterpieces thanks to artificial intelligence 📚

Following on Ethan Mollick's post sharing an article on Text-to-Speech (TTS) models (converting written books into audio), the "voice" they produce is a synthetic one created by the model itself, and it's #not a #copy of any specific #human voice. TTS models are trained on text data in the sense that they learn to convert written text into spoken words. TTS models are trained on a large dataset of text, often containing transcripts of spoken language. This text data is used to teach the model how words and phrases are pronounced. TTS models learn the #phonetic characteristics of words and sentences. They understand the #pronounciation of individual #words and how they should be connected when spoken together. TTS models also learn #prosody, which includes aspects like #intonation, #stress, and #rhythm in #speech. This helps them generate speech that #sounds more #natural and #expressive. The #voice in TTS models is usually not a specific human voice. Instead, it's a #simulated #voice created by the model based on the phonetic and prosodic information it has learned. These models have a neutral or generic voice that does #not #replicate any particular person's voice. When you input text into a TTS system, it uses the knowledge it has learned from the training data to #synthesise #speech. It combines phonetic information, prosody, and other #linguistic #rules to generate the audio. Some TTS systems allow you to choose from different voices. These voices are variations of the model's synthetic voice, with different prosodic patterns, accents, or speaking styles. They are not recordings of human voices but rather modifications of the model's synthetic voice. #audio #audiobooks #narrator #speech #syntheticdata

#COPYRIGHT* A good thing done by AI: turning 5,000 public domain ebooks into audio books that sound like they were read by humans. I’ve sampled a couple, and the quality is quite impressive. Books here: aka.ms/audiobook Academic paper here: https://lnkd.in/e2PhuaUG

These are some of the early examples of the type of utopia we could create if we get AI right, open access to knowledge, for everyone! Amazing stuff. Thanks Ethan Mollick for sharing #AI #Audiobooks #ProjectGutenberg

The lines from the book called ' Ikigai'. Telling us about how our mind is focused in some ways. In this paragraph they mention the lines given by Einstein for being focused by reference of relativity. You must read! #Ikigai #read #books #selfgrowth #growth

📚🍁 Finished my summer reading list! Now, I'm looking forward to your fall books recommendations, especially interested in strategic thinking and AI topics. Please share your favorite books, authors! Thank you 📖🍂 #books #reading #StrategicThinking #AIbooks

My book "With "Oscar" and Nearby" is on its last-mile way. #AI #NN #MLT #OlegLubske #digitalpublishing #books #WithOscarAndNearby #education #worldcinemahistory #hollywoodhistory #filmcelebrities

Have you heard of this Exciting News! You need to Discover the world of AI and language models with our FREE PDF eBook Whether you're a tech enthusiast or a curious mind, this eBook offers insights into the fascinating world of AI-powered language generation. Don't miss out – click the link below to download your copy today for FREE! https://selar.co/q6627j #AI #LanguageModels #Ebook.

AI should be able to rewrite a book in the style of your favourite author. Imagine taking a dry book with a good idea and getting your favourite author to rewrite it.