Study improves voice conversion model with Mel-spectrogram augmentation.
problem Insufficient speech pairs data for training sequence-to-sequence voice conversion models.
method Experimented with Mel-spectrogram augmentation using SpecAugment policies and proposed new augmentation policies.
result Time axis warping policies showed better performance in training the voice conversion model.
iSTFTNet speeds up mel-spectrogram vocoders without sacrificing quality.
problem Efficiently converting mel-spectrograms to speech with minimal computation.
method Replaces convolutional layers with iSTFT after frequency dimension reduction.
result Significant reduction in computational cost with comparable quality.
CycleGAN-VC3 improves CycleGAN-VCs for mel-spectrogram conversion.
problem Ambiguity in CycleGAN-VC/VC2 effectiveness for mel-spectrogram conversion.
method Proposes CycleGAN-VC3 with time-frequency adaptive normalization (TFAN).
result CycleGAN-VC3 outperforms or matches CycleGAN-VC2 for mel-spectrogram conversion.
SpecGrad improves neural vocoder sound quality by adapting diffusion noise to log-mel spectrogram.
problem Improving neural vocoder sound quality, especially in high-frequency bands.
method Adapting the diffusion noise distribution to the conditioning log-mel spectrogram through time-varying filtering.
result SpecGrad generates higher-fidelity speech waveform than conventional DDPM-based neural vocoders.
MaskCycleGAN-VC improves voice conversion without parallel data.
problem Limited ability to convert mel-spectrogram data without parallel data.
method Integrates a novel auxiliary task called filling in frames (FIF) to learn time-frequency structures.
result MaskCycleGAN-VC outperforms existing methods with similar model size.
End-to-end neural network based approaches to audio modelling are generally outperformed by models trained on high-level data representations. In this paper we present preliminary work that shows the feasibility of training the first layers of a deep convolutional neural network (CNN) model to learn the commonly-used l…
Convolutional neural network (CNN) architectures have originated and revolutionized machine learning for images. In order to take advantage of CNNs in predictive modeling with audio data, standard FFT-based signal processing methods are often applied to convert the raw audio waveforms into an image-like representations…
In this paper we propose WaveGlow: a flow-based network capable of generating high quality speech from mel-spectrograms. WaveGlow combines insights from Glow and WaveNet in order to provide fast, efficient and high-quality audio synthesis, without the need for auto-regression. WaveGlow is implemented using only a singl…
DiffWave generates high-fidelity audio waveforms efficiently.
problem Conditional and unconditional audio waveform generation.
method Non-autoregressive diffusion model using Markov chain synthesis.
result DiffWave produces high-quality audios in various tasks.
VPFD uses vocoder features for adversarial training in VC.
problem Adversarial training on waveform data is time-consuming and memory-intensive.
method VPFD employs vocoder features for adversarial training.
result VPFD achieves VC performance comparable to waveform discriminators with reduced training time and memory.
WaveGrad generates high-fidelity audio using gradient estimation.
problem Generating high-fidelity audio efficiently.
method Conditional model using score matching and diffusion models, iteratively refining a Gaussian white noise signal.
result WaveGrad can generate high-fidelity audio samples using as few as six iterations.
Paper aims to find joint representation between vocal tract geometry and speech sound acoustics.
problem Finding a joint latent representation between articulatory and acoustic domains for vowel sounds.
method Invertible neural network models, convolutional autoencoder, normalizing flows, semi-supervised learning.
result Satisfactory performance in articulatory-to-acoustic and acoustic-to-articulatory mapping.
Neural network-based vocoders have recently demonstrated the powerful ability to synthesize high-quality speech. These models usually generate samples by conditioning on spectral features, such as Mel-spectrogram and fundamental frequency, which is crucial to speech synthesis. However, the feature extraction procession…
We investigated the training of a shared model for both text-to-speech (TTS) and voice conversion (VC) tasks. We propose using an extended model architecture of Tacotron, that is a multi-source sequence-to-sequence model with a dual attention mechanism as the shared model for both the TTS and VC tasks. This model can a…
Grad-TTS models speech from text using diffusion probabilistic techniques.
problem Creating high-quality speech from text input.
method Score-based decoder with stochastic differential equations for noise-to-speech transformation.
result Grad-TTS produces mel-spectrograms from text input with competitive quality.
We present a voice conversion solution using recurrent sequence to sequence modeling for DNNs. Our solution takes advantage of recent advances in attention based modeling in the fields of Neural Machine Translation (NMT), Text-to-Speech (TTS) and Automatic Speech Recognition (ASR). The problem consists of converting be…
In the past, Acoustic Scene Classification systems have been based on hand crafting audio features that are input to a classifier. Nowadays, the common trend is to adopt data driven techniques, e.g., deep learning, where audio representations are learned from data. In this paper, we propose a system that consists of a …
Study examines how data augmentation impacts optimization in linear regression.
problem Understanding how data augmentation schedules affect optimization in linear regression.
method Analyzed the effect of augmentation on optimization in linear regression with MSE loss, using classical convex optimization and recent work on implicit bias.
result Proved that under certain joint schedules for learning rate and augmentation scheme, augmented gradient descent converges and characterized the resulting minimum.
Data augmentation doesn't improve robustness, contrary to belief.
problem The effectiveness of data augmentation in improving model robustness is questioned.
method Taking a Domain Generalization viewpoint, the study examines the robustness of augmented representations.
result Augmented representations are not robust to distortions used during training.
This paper improves auto-augment efficiency by sharing augmentation weights.
problem Efficient evaluation of augmentation policies for model training.
method Augmentation-Wise Weight Sharing (AWS) to create a fast yet accurate proxy task.
result Augmentation policies found achieve superior accuracies compared to existing methods.
Simple policy search outperforms advanced learnable test-time augmentation techniques.
problem Improving predictive performance through test-time data augmentation.
method Greedy policy search (GPS) for learning test-time augmentation policies.
result Augmentation policies learned with GPS achieve superior predictive performance and robustness.
New measures quantify how data augmentation improves model performance.
problem Understanding the effectiveness of data augmentation in deep learning.
method Introduced Affinity and Diversity measures to quantify augmentation performance.
result Augmentation performance is best achieved by optimizing both Affinity and Diversity.
CNNs encode data augmentation transformations, especially in early layers.
problem Whether neural network features encode data augmentation transformations.
method Systematic approach using pre-trained vision models to predict augmentation transformations.
result Neural network features encode data augmentation transformations, especially in early layers.
WeMix improves data augmentation by correcting bias in deep learning.
problem Data augmentation's effectiveness is limited by data bias.
method Developed AugDrop and MixLoss algorithms to correct data bias.
result WeMix improves data augmentation performance through bias correction.
A study on optimizing data augmentation weights for improved test-time predictions.
problem Improving robustness of predictions during testing with data augmentation methods.
method A weighted Test-Time Augmentation (TTA) approach based on variational Bayesian framework to optimize weights.
result Optimizing weights suppresses unwanted data augmentations and improves prediction performance.
Data augmentation can achieve the same statistical benefits as full augmentation up to an approximation error.
problem Data augmentation in learning problems
method Using Fourier analysis and representation theory of finite groups
result Partial data augmentation achieves the same minimax rates as full augmentation
Study fully augmented links in thickened torus, generalizing S3 results.
problem Classify and describe geometric properties of fully augmented links in thickened torus.
method Geometric analysis and decomposition of link complements into ideal right-angled torihedra.
result Proves Volume Density Conjecture for fully augmented links in thickened torus.
Data augmentation has been widely applied as an effective methodology to improve generalization in particular when training deep neural networks. Recently, researchers proposed a few intensive data augmentation techniques, which indeed improved accuracy, yet we notice that these methods augment data have also caused a …
A key challenge in leveraging data augmentation for neural network training is choosing an effective augmentation policy from a large search space of candidate operations. Properly chosen augmentation policies can lead to significant generalization improvements; however, state-of-the-art approaches such as AutoAugment …
SapAugment learns adaptive augmentation policies for better model training.
problem Fixed data augmentation methods often apply the same augmentation to all samples, ignoring sample difficulty.
method SapAugment adapts augmentation parameters based on training loss, learning a sample-adaptive policy.
result SapAugment achieves up to 21% relative reduction in word error rate on LibriSpeech dataset.
Automatically learns optimal data augmentation for image classification.
problem Finding optimal data augmentation hyperparameters is computationally demanding and requires domain knowledge.
method Proposes an online bilevel optimization framework to learn data augmentation parameters directly.
result Jointly trained method achieves comparable or better classification accuracy than hand-crafted data augmentation without an external validation loop.
Graph data augmentation improves GNN performance in node classification.
problem Improving generalizability of graph neural networks (GNNs) in semi-supervised node classification.
method Introduces GAug framework for graph data augmentation using neural edge predictors.
result GAug framework improves GNN-based node classification performance across various architectures and datasets.
This work investigates image augmentations for GAN training, improving image quality.
problem Improving the accuracy and robustness of GAN models for image synthesis.
method Systematic study of various image augmentation techniques for GAN training.
result Vanilla GANs can achieve state-of-the-art generation quality with image augmentations.
A method to prevent image representation collapse through data-dependent augmentation.
problem Representation collapse due to image augmentations that damage information.
method Formalizing a stochastic encoding process with a tug-of-war between corruption and preserved information, using infoMax objective.
result Learning a data-dependent distribution of augmentations to avoid representation collapse.
We describe a correspondence between augmentations and certain representations of the knot group. The correspondence makes the 2-variable augmentation polynomial into a generalization of the classical A-polynomial. It also associates to an augmentation a rank, which is bounded by the bridge number and shares its beha…
New attacks show data augmentation may not improve privacy.
problem Measuring privacy risk in models trained with data augmentation.
method Formulated membership inference as a set classification problem, designed input permutation invariant features.
result Proposed approach universally outperforms original methods on models trained with data augmentation.
Expert augmentation improves hybrid model generalization.
problem Limited generalization of hybrid models outside training distribution.
method Introducing expert augmentation to improve hybrid model performance.
result Expert augmentation improves generalization of hybrid models.
Augmentation improves machine learning model performance on small datasets.
problem Suboptimal generalization performance of machine learning models on small datasets.
method Data augmentation to increase sample size and diversity.
result Augmentation improves AUC by 15.55% on average for small datasets.
This work characterizes how data augmentation shapes neural representations.
problem Understanding the impact of data augmentation on neural network representations.
method Embedding neural network hidden representations into a metric space invariant to transformations, analyzing shape-space trajectories.
result Increasing data augmentation strength leads to well-behaved trajectories in the embedded space, and different augmentation types steer representations in distinct directions.
ADA augments data using AR replicas for robust regression.
problem Improving robustness in nonlinear over-parametrized regression.
method Extends Anchor regression (AR) for data augmentation, using replicas of modified samples.
result ADA provides more robust regression predictions compared to state-of-the-art solutions.
Augmentations and sheaves linked for Legendrian graphs.
problem Understanding categorical Legendrian isotopy invariants.
method Equivalence between augmentation category and DG category of sheaves.
result Proved 'augmentations are sheaves' for Legendrian graphs.
We construct the augmentation representation. It is a representation of the fundamental group of the link complement associated to an augmentation of the framed cord algebra. This construction connects representations of two link invariants of different types. We also study properties of the augmentation representation…
Counterfactual data augmentations may not ensure OOD robustness if performed by a context-guessing machine.
problem Deep learning models lack out-of-distribution robustness due to reliance on spurious features.
method Theoretical analysis and demonstration of counterfactual data augmentations performed by a context-guessing machine.
result Counterfactual data augmentations by a context-guessing machine do not lead to robust OOD classifiers.
Transfer learning and data augmentation improve stock classification performance.
problem Challenges in stock classification due to noise and volatility.
method Pre-trained model on S&P500 index features, transfer learning to new models, data augmentation on feature space.
result Augmentation on feature space leads to 20% increase in risk-adjusted returns.
This research investigates reliable local explanations for machine listening models.
problem Generating reliable local explanations for machine listening models.
method Investigates the sensitivity of SoundLIME explanations to input perturbations and proposes a novel method for identifying suitable content types.
result SoundLIME explanations are sensitive to the content in occluded input regions, and the average magnitude of input mel-spectrogram bins is the most suitable content type for temporal explanations.
Develops a statistical framework for self-supervised representation learning using data augmentation.
problem Lack of theoretical understanding of data augmentation in nonlinear settings.
method Augmentation invariant manifold learning framework and stochastic optimization algorithm.
result Improves downstream analysis by exploiting manifold's geometric structure and invariant property of augmented data.
Data augmentation impacts adversarial risk; careful application recommended.
problem Understanding how data augmentation affects adversarial risk in deep learning.
method Empirical analysis using three measures of adversarial risk.
result Data augmentation does not always improve adversarial risk; augmented data influences models more.
Fully augmented links have dense volume densities but discrete in certain ranges.
problem Characterizing the volume density spectrum of fully augmented links.
method Analyzing the ratio of volume to the number of augmentations.
result The set of FAL volume densities is dense in $[2\voct, 10\vtet)$ but discrete in $[\voct,2\voct)$.