Don't forget to read the paper ahead of the reading: https://
openai.com/blog/whisper/
RESEARCH
-
Whisper Paper Reading: OpenAI’s Speech Recognition Model
By
–
-
ML4Audio Paper Reading: OpenAI Whisper Speech Recognition
By
–
Come one, come all! We're back with ML 4 Audio paper readings! This time we'll unravel the mysteries behind @OpenAI Whisper and what makes it so good at speech recognition and other Seq2Seq tasks! We'll be live this Thursday, 24th Nov, 18:00 CET on @huggingface discord
-
Juris Hartmanis’ Mentorship Impact on Computer Science Student
By
–
A heartwarming story by @rrwilliams
, on Juris Hartmanis' impact on him. These words really stood out to me. Every advisor should aspire to give the same support to their students as Juris did to Ryan. https://
youtube.com/watch?v=ACxU-9
0O-ag&t=10350s
… -
Top 10 Tech Trends 2023: AI Everywhere and Metaverse Growth
By
–
10 Tech Trends 2023: #AI Everywhere & Parts of #Metaverse Take-Off! https://
forbes.com/sites/bernardm
arr/2022/11/21/the-top-10-tech-trends-in-2023-everyone-must-be-ready-for/
… @pierrepinna @sallyeaves @PawlowskiMario @Xbond49 @gvalan @SpirosMargaris @JimMarous @TheRudinGroup @HaroldSinnott @Shi4Tech @mikeflache #MachineLearning #DeepLearning #Fintech #5G #ESG -
Make with H2O: Kaggle Grandmasters Q&A and Competition Strategies
By
–
Join us next week for Make with http://
H2O.ai! In this exclusive Make session, you will have the opportunity to ask Kaggle Grandmasters data science questions, hear winning Kaggle competition strategies, and more! Register: https://
h2o.ai/events/make-wi
th-h2o/?utm_campaign=Make-With-H2O&utm_source=twitter&utm_medium=Organic-Social
… -
Transformers Finally Added: Critical AI Architecture Recognition
By
–
Thank you for adding Transformers! Many miss it which is wrong in 2022!
-
Comparison of GPT-3 and GPT-4 Parameter Counts
By
–
Currently, GPT-3 has 175 billion parameters, which is 10x faster than any of its closest competitors. GPT-4 is rumored be about 100 trillion parameters.
-
Magic3D: High-Resolution Text-to-3D Content Generation
By
–
Magic3D: High-Resolution Text-to-3D Content Creation Lin et al.: https://
arxiv.org/abs/2211.10440 #ArtificialIntelligence #DeepLearning #MachineLearning -
Tech Journalism Doesn’t Require Coding Skills
By
–
I’d say this consumer/gadget tech journalist trope is from about 10 years ago. And also you really don’t need to code to be able to competently analyse the tech industry. Just like you don’t need to code if you’re a sociologist/political scientist/ethicist studying impact of tech
-
Galton Board Demonstrates Statistical Concepts for Data Science
By
–
From Apparent Chaos to Order
— Pascal Bornet (@pascal_bornet) 21 novembre 2022
This device is a Galton Board. 3,000 beads are dropped to demonstrate several #statistical & #math concepts:
– Central Limit Theorem
– Regression to the Mean
– Normal distribution
– Binomial distribution
– Pascal's Triangle
Crt: I. Korolev#data #ai pic.twitter.com/gD5XzSBvLeFrom Apparent Chaos to Order This device is a Galton Board. 3,000 beads are dropped to demonstrate several #statistical & #math concepts:
– Central Limit Theorem
– Regression to the Mean
– Normal distribution
– Binomial distribution
– Pascal's Triangle Crt: I. Korolev
#data #ai