End-to-end models perform better with learned log-scaled mel-spectrogram features.
problem End-to-end neural network models struggle with performance compared to models using high-level data representations.
method Trained first layers of a CNN model on log-scaled mel-spectrogram transformation and then used these learned features to initialize an end-to-end CNN classifier.
result Convergence and performance on ESC-50 dataset are similar to a model trained on pre-processed log-scaled mel-spectrogram features.
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.
A simple fusion of deep and shallow learning improves acoustic scene classification.
problem Improving acoustic scene classification accuracy.
method Combining a deep learning approach and a feature engineering approach using a late fusion strategy.
result The fused system achieves 72.8% classification accuracy, outperforming individual methods.
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.
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.
A novel method for learning DAGs from positive-valued data.
problem Causal discovery from observational data of positive-valued variables.
method Hybrid Moment-Ratio Scoring (H-MRS) algorithm combining moment-based scoring and log-scale regression.
result H-MRS integrates log-scale Ridge regression for moment-ratio estimation with a greedy ordering procedure based on raw-scale moment ratios, followed by Elastic Net-based parent selection.
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.
We introduce a new functional measure of tail dependence for weakly dependent (asymptotically independent) random vectors, termed weak tail dependence function. The new measure is defined at the level of copulas and we compute it for several copula families such as the Gaussian copula, copulas of a class of Gaussian mi…
Solution for voice conversion with limited data using hierarchical seq2seq and attention models.
problem Voice conversion between speakers with limited parallel audio pairs.
method Hierarchical sequence to sequence model with attention-based decoder, trained on single speaker dataset.
result Improved voice conversion quality using mel spectrograms and wavenet vocoder.
Estimates spectral projections restricted to uniformly embedded submanifolds.
problem Estimating spectral projections on submanifolds of manifolds with nonpositive curvature.
method Estimates the L2(M)oLq(Σ) norm of spectral projection operators. result Sharp spectral projection estimates for small spectral windows.
Study finds long memory in some emerging Asian stocks but not in developed markets.
problem Evaluating stock market efficiency in emerging vs developed markets.
method Improved wavelet estimator of long range dependence.
result Emerging Asian markets show more long memory in stock returns than developed markets.
Study shows cliff-learning in transfer learning from foundation models.
problem Data-scaling of transfer learning from foundation models in low data regimes.
method Investigation of cliff-learning phenomenon through foundation-model analysis and toy models.
result Cliff-learning reflects compatibility between priors and tasks.
WaveGlow generates high-quality speech from spectrograms.
problem Speech synthesis quality and efficiency.
method WaveGlow combines Glow and WaveNet insights, using a single network and cost function for efficient, high-quality audio synthesis.
result WaveGlow produces audio samples at over 500 kHz, matching WaveNet quality.
S&P 500 index data sampled at one-minute intervals over the course of 11.5 years (January 1989- May 2000) is analyzed, and in particular the Hurst parameter over segments of stationarity (the time period over which the Hurst parameter is almost constant) is estimated. An asymptotically unbiased and efficient estimator …
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.
The so-called Pareto-Levy or power-law distribution has been successfully used as a model to describe probabilities associated to extreme variations of worldwide stock markets indexes data and it has the form Pr(X>x) x∗∗(−alpha)forgamma<x<infinity.Theselectionofthethresholdparametergamma from empirical d…
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.
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.
Unified model for financial derivatives pricing with stochastic interest rates.
problem Pricing and hedging financial derivatives with stochastic interest rates.
method Volterra Stein-Stein model with correlated Gaussian Volterra processes.
result Explicit formulas for bond and cap/floor pricing, and characteristic function for log-forward index.
A method for vectorizing persistence diagrams simplifies topological data analysis.
problem Challenges in integrating persistence diagrams into machine learning pipelines.
method Quantized Persistence and Integral transforms of Diagrams (Qupid) using binning and discrete transforms.
result Qupid preserves highly competitive performances compared to state-of-the-art methods across various classification tasks.
RawNet synthesizes speech without relying on human-designed features.
problem Speech synthesis models rely heavily on human-designed spectral features.
method End-to-end neural vocoder using auto-encoder structure, learning features and audio recovery directly from raw waveform.
result RawNet achieves better speech quality with a simplified model architecture and faster inference speed.
Joint training model for TTS and VC tasks using Tacotron and WaveNet.
problem Training a shared model for text-to-speech and voice conversion.
method Extended Tacotron model with dual attention mechanism for shared tasks, WaveNet for waveform generation.
result Joint training of a shared model achieves both TTS and VC tasks efficiently.
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.
Study compares new audio representation methods for limited data music retrieval.
problem Improving machine learning for audio data with limited training data.
method Investigated mel-spectrogram and Mel scattering representations, and augmented target loss function.
result All proposed methods outperform standard mel-spectrogram when using limited data.
Paper extends ICA to ISA with auxiliary variables for better speech representation learning.
problem Learning unsupervised speech representations with independent subspaces.
method Theoretical framework of nonlinear ISA with auxiliary variables.
result Proposes an algorithm to learn speech representations with independent subspaces.
Being able to predict whether a song can be a hit has impor- tant applications in the music industry. Although it is true that the popularity of a song can be greatly affected by exter- nal factors such as social and commercial influences, to which degree audio features computed from musical signals (whom we regard as …
PriorGrad improves speech synthesis models by using data-dependent adaptive priors.
problem Inefficiency in denoising diffusion models due to mismatch between prior and data distributions.
method Proposes PriorGrad, an adaptive prior derived from data statistics based on conditional information.
result PriorGrad achieves faster convergence and superior performance in speech synthesis models.
We present a sparse knowledge gradient (SpKG) algorithm for adaptively selecting the targeted regions within a large RNA molecule to identify which regions are most amenable to interactions with other molecules. Experimentally, such regions can be inferred from fluorescence measurements obtained by binding a complement…
Novel fusion of autoencoders predicts sleepiness from speech.
problem Predicting sleepiness from speech recordings.
method Attention-based and recurrent sequence to sequence autoencoders for unsupervised representation learning.
result Fusion of autoencoders' representations achieves higher correlation with sleepiness scales.
Model predicts tornado damage with high accuracy.
problem Accurate prediction of tornado-induced property damage.
method Zero-inflated neural networks.
result Models predict both occurrence and amount of damage.
Study of historic stock returns distributions, highlighting asymmetry and outliers.
problem Understanding the asymmetry in accumulated gains and losses in stock returns over time.
method Analyzing decades-long historic distributions of S&P500 returns, comparing gains and losses, using statistical U-tests and fitting log-log scale linearly.
result The mean of de-trended distributions increases linearly with the number of days of accumulation, and the overall skew is negative, indicating heavier tails of losses.
End-to-end sound event detection uses learned time-frequency representations.
problem Sound event detection using standard features like mel spectrogram.
method End-to-end approach with a feedforward layer block and convolutional recurrent network.
result The learned time-frequency representations improve performance over fixed features.
Study finds 'Dragon Kings' in stock market volatility during major economic crises.
problem Identifying significant deviations from normal market volatility.
method Analyzed S&P500 index volatility, categorized as Black Swans, Dragon Kings, or Negative Dragon Kings, using modified Generalized Beta and Generalized Beta Prime distributions.
result Observed 'potential' Dragon Kings that eventually turn into Negative Dragon Kings, with more pronounced phenomenon as time averaging increases.
Model predicts drug response in tumors using genomic profiles.
problem Challenges in translating genomic insights to tumor-specific drug response predictions.
method Deep neural network (DNN) model trained on mutation and expression profiles of cancer cell lines.
result Achieved mean squared error of 1.96 for predicting IC50 values of 265 drugs.
Deep learning extracts terrain texture covariates for geostatistical modeling.
problem Improving prediction accuracy in geostatistical modeling using terrain texture data.
method Deep learning approach to automatically derive optimal terrain texture covariates from SRTM 90m DEM.
result Deep learning-derived covariates have strong explanatory power (R-squared around 0.6) for geochemical data.
Deep CNN-RNN model classifies breathing sounds for respiratory disease diagnosis.
problem Automated diagnosis of respiratory diseases using wearable devices.
method Patient-specific model tuning and local log quantization of weights.
result 71.81% accuracy on leave-one-out validation with patient-specific data.
Enhances ASC using time- and frequency-liked CNNs and bilinear pooling.
problem Improving acoustic scene classification accuracy.
method Harmonic and percussive source separation, two-stream CNN architecture, bilinear pooling.
result Improved accuracy on DCASE 2019 sub task 1a dataset.