thatsadude

Could be a coup or that Sam colluded with M$

vec(ABC)=kron(C.T,A)vec(C) is all your need for matrix calculus!

I'm concerned about who AGI will serve? Rich people surely!

I’m speechless as a regular user. This is so dumb, sure Twitter will die within a year or twos.

The problem of PoE is that it's too complex for causal middle age gamers like me. I have no time to invest in learning PoE after every patch.

AI Engineering is broad. There are a lot of things to learn and a lot of mistakes you have to made yourself.

For applied ML, my tips are: make sure you learn the dark side of BatchNorm and Dropout, start with simple and elegant baselines instead of complex SOTA algorithms, spend more time on understanding your data than trying algorithms, be aware that SOTA in a related task will often suck at your task, be data driven. Also, most of your ideas will not work but you have to try and conduct experiments carefully.

What? I pruned my comments before leaving reddit. If they restore my comments, does this mean they don’t respect me as a user at all!

Well, don’t kill Firebase next please.

Few reasons:

1. Speech distortion is extremely detrimental to ASR models while human listeners may not be able to notice. Noise reduction models such as RNNNoise and DeepFilterNet try to reduce "perceptible" noises. And doing that will create imperceptible distortion which ASR models does not like at all.

2. Many noise reduction models apply on raw spectrograms or ERB band such as RNNNoise and DeepFilterNet (equivalent rectangular bands). On the other hands, ASR models mostly run on melspectrograms. This mismatch tends to create problem. I have seen many papers from Google claimed that reducing noise in melspectrogram often helps their keyword spotting.

The golden standard is ITU-T P.808 subject test https://github.com/microsoft/P.808 Of course, running a subjective test is expensive, and hence, there are objective scores such as DNSMOS and UTMOS which use neural networks to predict P808 values.

Much much better, I really recommend DeepFilterNet, it's the most well-rounded open-source AI noise suppression tool out there. Big caveat, it won't help ASR model, e.g., Whisper.

Source: I work in this research area.

Not all spectral methods have such artifact. The type of artifacts you mention happens when you need to do phase retrieval or try to reconstruct waveforms from melspectrogram. Deepfilternet does spectral masking on the complex spectrogram so there is no need for phase retrieval.

> What exactly makes anyone think that they can detect an LLM that is outputting text? The notion seems absurd yet it keeps coming up.

Publish or perish!

Good read, I often have to run code step by step to catch accidental double promotion in dsp functions.

> I'm getting a bit tired of people putting their class projects or quick engineering projects on arXiv.

I got downvoted when I expressed similar opinion with regards to MiniGPT4. I guess HN crowd value usefulness more than real contribution.

How is this a paper, they simply fine-tune the fully-connected layer of BLIP-2 https://arxiv.org/pdf/2301.12597.pdf

That's a layman explanation. ShotSpotter is likely a passive radar system. In recent years, you can combine signal processing and supervised learn (neuralnet) to get better direction-of-arrival estimations.

You should see MobileOne and FastVIT from Apple

I fear for my team, really do. We work in audio processing tech. As a small team of five people residing in a third world country, we will never have a chance to compete with big companies because it looks like ML-based audio processing will soon be a commodity. Sad but it's a reality!

Personally, Safari left a bad taste in my mouth. We had to kill a product line because bugs in Safari's AudioWorklet. I don't think audio processing on Web is viable until Safari plays nicely with WebAudio.

Time should be shamed for publishing this.

Google published papers but has anybody be able to replicate their results?

It’s AGI to me!

At least they publish their papers!

Well, when you have thousands of Matlab functions you wrote over a decade, you would rather use Octave instead of rewriting them in Python. I also suppose that Octave is slower than Matlab because OpenBLAS is inferior to MKL.

I am really sad about this

Have being using Ruff for a week, it's so awesome that I'm surprised that VSCode extension for Ruff has only 8500 users.

Art!

The ability to ask the right questions!

I was at your position. I dragged on a journal and my thesis for 2 years. My advisor literally begged me to get my dissertation done at times. He complained I was a perfectionist and told me all good advice you have here every month for those two years. It didn't work.

Eventually, he told me "How about you ``take leave`` and write whatever and bring it to me". I took one week leave, stayed at home and start writing. Somehow, I finished my first thesis draft in a month after that then I graduated. Those two years was the lowest years in my life emotionally.

I don't have any advice for you because every person is different. This is a battle you need to fight by yourself. Start writing!