machine-learning Tag

David Dao has created a great list of scary/sexist/racist/concerning AI projects to raise awareness of its misuses in society:

Artificial intelligence in its current state is unfair, easily susceptible to attacks and notoriously difficult to control. Often, AI systems and predictions amplify existing systematic biases even when the data is balanced. Nevertheless, more and more concerning the uses of AI technology are appearing in the wild. This list aims to track all of them. We hope that Awful AI can be a platform to spur discussion for the development of possible preventive technology (to fight back!).

We need socially aware, empathetic, well-rounded people from all kinds of backgrounds to get more involved in AI because the projects on this list are showing us that the current path we’re heading down is completely unacceptable.

Microsoft just announced that they’re exclusively licensing GPT-3.

I knew Microsoft has really been investing in ML/AI through Azure, and now it looks like the only way you’ll be able to use GPT-3 is through Azure services. I’m not sure how I feel about this yet.

From VentureBeat:

The implications of the licensing agreement weren’t immediately clear, but Microsoft says that OpenAI will continue to offer GPT-3 and other models via its Azure-hosted API, launched in June. (To date, the API, which remains in beta, has received tens of thousands of applications, according to OpenAI.) Microsoft plans to leverage the capabilities of GPT-3 in its own products, services, and experiences and to continue to work with OpenAI to keep commercialize the firm’s AI research.

I started using this app called Krisp over the weekend, and I’m def impressed.

I was using Discord over the weekend and tried the built in noise suppression. The description for the feature explains that it uses an embedded version of Krisp for it. The app uses machine learning to figure out which sounds are voices, and filters out everything else.

I live right above a night club (which sadly has been silent for the past few months) and next to the N train (which seems to run exactly the same as it always has) and New Yorkers are still New Yorkers so voice calls have been hard with all the noise. I’ve been using Krisp for my Discord, FaceTime, and Hangout calls over the weekend and it completely blocks out all of it.

I’m impressed with how it works but I’m also into machine learning (ML) and all the stuff people are calling AI these days so it’s even more interesting to me. I’ve also been reading about NVIDIA’s version of this, which works through their RTX graphics cards.

I’ve been wanting to start using ML in my projects for a while but I never came up with good enough applications for it. Voice detection is a perfect use case. I think this might have given me a reason to try playing around with it again.

Okay, my (very) limited understanding of quantum computing is that it could be exponentially more powerful than classical computing, but right now there are some problems like total running time (milliseconds? nanoseconds?) and very noisy data (considering that’s just quantum mechanics work). This project is machine learning (the thing people have been calling AI for a while) ON quantum computers to help with noisy data, and a bridge between quantum data and classical data.

I can’t tell if this is the beginning of the end of the human race OR just the only way we’ll ever get useful data out of quantum computing at all.

https://ai.googleblog.com/2020/03/announcing-tensorflow-quantum-open.html