I'm really fascinated by apps like Spotify and Shazam, and would like to pursue a career related to music information retrieval (MIR) and recommender systems. It seems like an esoteric field, especially since deep learning these days seems to primarily be focused on image and text data. Maybe we can connect and form a study group! Maybe we can share interesting papers (e.g. from ISMIR and ACM RecSys conferences), work on projects together, and share study materials.

Lab is a library that help improve your machine learning workflow by organizing and tracking experiments. I have posted updates to the project on this subreddit before. We've received some valuable feedback directly on this subreddit and later from users who found out about the project here. These feedback has helped us improve the project. Here's some of the updates to the project and we are glad if you find them useful. Please let us know if you have any suggestions or feedback.

For my next job, I am tasked to build a rig for my ML project, centered on Computer Vision. Given that at my former places of employment, said material was always purchased by my employers before I joined the project, I have simply no idea of what it would be necessary to buy. However, I have yet no idea of the extent of the database, as there may be several phases of data acquisition. But I still need some good GPU to not lose too much time during the training process.

I'm relatively new to machine learning (been working with it professionally for about a year) and started working on personal projects around January just for fun. One project that I started in February recently started showing some huge promise, and it could cut training time for Keras models by roughly 20% in most cases, based on my experiments. I'm looking to write a paper on the topic the project implements, but I'm stumped. This is the first research project I've worked on and I don't have any connections with professors or anything at the college I'll be attending. Does anyone have any tips for me to help get this project and paper farther off the ground?

Let me know what you think; I want to do more of these beginners' focused tutorials, just to get more of the general public interested in AI for video. What would you want to see in this simplified form?

In VAE, samples are generated from the latents variables (mean and variance) during training only. Afterwards the encoder and latent are not used anymore and the generation is done by using a random variable as input for the decoder.

What I mean is that, can I use a simple Image classification algorithm connected in the end with a multi-dimensional RNN at the end that based on my definition of required size, say 300x500 produces 300x500 pixel values? Does Anyone know simple semantic segmentation Fully convolutional method without upsampling?

I am building a PC that is going to be used for training neural networks. My GPU is going to be the RTX 3080 (when it releases in Q3 2020) and I have two options for the CPU. Either the i5-9600k or the Ryzen Threadripper 1900x (both are the same price). However, I cannot decide which one to purchase. The 1900x has 8 cores instead of the 6 core 9600k, but the i5-9600k can overclock to 5ghz while the Threadripper cannot.

On Linear Algebra, it's rather rusty as it was a sudden leap into proof-heavy classes, so I didn't quite get what was going on most of the time. I have gone through an intro class on ML and Optimization as well (Rather superficial concepts without much exercises). I am going through a gap year - don't wanna be paying a hefty amount for zoom-classes next study year - so will be class-free for approximately 14 months. Personally aiming to devote 5 hours a day, 6 days a week.

When Camille François, a longstanding expert on disinformation, sent an email to her team late last year, many were perplexed. Her message began by raising some seemingly valid concerns: that online disinformation -- the deliberate spreading of false narratives typically designed to sow mayhem -- "could get out of control and become a huge threat to democratic norms". But the text from the chief innovation officer at social media intelligence group Graphika soon became rather more wacky. Disinformation, it read, is the "grey goo of the internet", a reference to a nightmarish, end-of-the world scenario in molecular nanotechnology. The solution the email proposed was to make a "holographic holographic hologram". The bizarre email was not actually written by François, but by computer code; she had created the message -- from her basement -- using text-generating artificial intelligence technology.