Viewing profile — kastnerkyle

The direct counter-argument to "worst representation" is usually "representation with fewest assumptions", waveform as shown here is getting close. Though recording environment, eq…

On the loops / sampling front: I always thought RAVE [0][1][2] was a very interesting approach, that really embraces latent spaces and sample/stretch type approaches in the wavefor…

Research into "pure" unconditional generation can often lead to gains in the conditional setting. See literally any GAN research, VQ-VAE, VAE, diffusion, etc - all started from "un…

This work is another classic in the "neural nets meet spreadsheets" genre [0]. Really helps visualize what is going on in (at least some) latent spaces. [0] https://vusd.github.io/…

This same technique, extended can work well for detecting plagiarism from the underlying corpus as well, by tracking a trie of "good" completions in the n-gram sense, and a longer …

There was a really nice post on doing this kind of thing with CRF back in 2015 [0]. Open source data, and code on github. Also a nice tutorial on structured prediction using CRF ty…

What since Adam? Learning rate scales / schedules? I cannot think of many big massive changes since ~2014, most of the setups from that era (grad clip + medium-ish LR, some ramp up…

I disagree with this. Binarized MNIST samples of any reasonable quality are (still) tricky to get right without a hierarchical system (read: VQ-VAE tokens or some such encoder spac…

Previously TortoiseTTS was associated with PlayHT in some way, although the exact connection is a bit vague [0]. From the descriptions here it sounds a lot like AudioLM / SPEAR TTS…

Maybe 'Image Quilting for Texture Synthesis and Transfer', Efros and Freeman [0]? There's some neural / patch blends from 2016 that I always thought were interesting (CNN-MRF) [1],…

For any models where you get access to the code (counting HF in this), definitely a lot of tooling available - and this tool looks really nice hadn't seen it before! Hope similar l…

Discrete diffusion with rewriting can work well. It feels loosely similar to backtracking, if you assume n_steps large enough - need to be able to rewrite any non-provided position…

You might be interested in this thread from 2019 here on HN as well, I linked to a line of work on constrained sequence generation from Sony CSL that is pretty nifty https://news.y…

On this point, for images much of the structure is encoded in the phase . This fact was very surprising to me when I learned it. See the images in this stack overflow for an exampl…

I really like the descriptions from SUNDAE ( https://arxiv.org/abs/2112.06749 ) if you have some background about general neural net style modeling, and generally find the multinom…

I've been trying to follow your project since I heard about it - but hadn't read the link you posted here. Really thorough, really good description of many of the quagmires (with p…

https://www.theatlantic.com/technology/archive/2020/02/whats... is an interesting related article on this topic

Also, if you are seeking more melodies / catchy tunes to model NES-MDB ( https://github.com/chrisdonahue/nesmdb ) might fit the bill. The modeling paper using that dataset + transf…

I've had pretty good luck recently with a mix of SUNDAE ( https://arxiv.org/abs/2112.06749 ) and coconet ( https://arxiv.org/abs/1903.07227 ) and/or Music Transformer based interna…

Yes - the crux is just to add some logic and throw out beams which don't match your constraint, then rank candidates based on sequence probability. You can roll-back the generation…

There are many flavors of beam search - I have found that for adding explicit checks and constraints (for example rhyme constraints or certain pivot words) the resulting proposals …

Tricks in beam search to force rhyme schemes, or techniques like constrained markov chains (c.f. https://redylan.neocities.org/#/how-it-works/ and https://github.com/gabrielebarbie…

Much less audio is potentially needed for TTS than ASR, however the spread and quality of the TTS dataset is critical which is one reason why just training on ASR datasets "in reve…

Worth noting that a big chunk of the core TTS code here is built on tools from other researchers like Ryuichi Yamamoto and Keith Ito, and they have great implementations to check o…

This type of concatenation I first saw in Alex Graves' work on "Generating Sequences With Recurrent Neural Networks", including his unpublished TTS demo [1]. Biasing with part of a…