A new multi-layer attention mechanism improves speech keyword recognition accuracy.
problem Inaccurate attention weights in LSTM networks for speech keyword recognition.
method Introducing information from layers prior to feature extraction into attention weights calculations.
result The proposed multi-layer attention mechanism leads to more accurate attention weights and improved keyword spotting performance.
Sparse representations improve network robustness and stability.
problem The benefits of sparse representations in artificial networks.
method Analysis of sparse networks with sparse weights and activations, simulations on MNIST and Google Speech Command Dataset.
result Sparse networks show significantly improved robustness and stability compared to dense networks.
Paper optimizes KWS models using NAS and quantization for limited resources.
problem Developing efficient keyword spotting models in resource-constrained environments.
method Neural Architecture Search (NAS) for model structure optimization and quantization of weights and activations.
result Achieved high accuracy (95.55%) with minimal parameters and operations using NAS and quantization.
Detects adversarial examples in deep speech recognition.
problem Vulnerability of deep speech recognition systems to adversarial attacks.
method Formulated as a classification problem, generated adversarial and normal datasets, trained CNN.
result Accurately distinguishes between adversarial and normal examples for known attacks.
Despite showing state-of-the-art performance, deep learning for speech recognition remains challenging to deploy in on-device edge scenarios such as mobile and other consumer devices. Recently, there have been greater efforts in the design of small, low-footprint deep neural networks (DNNs) that are more appropriate fo…
Paper finds universal speech command perturbations that fool models.
problem Existence of universal adversarial examples in speech command classification.
method Proposed a novel analytical framework for evaluating universal perturbations and a detailed distortion measurement method.
result Universal perturbations can fool speech command classification models across different models.
New dataset for evaluating speech recognition fairness across demographics.
problem Lack of fairness metrics in speech recognition datasets.
method Developed Fair-Speech dataset with diverse demographic information.
result Helps evaluate ASR models for fairness across demographics.
In recent years, neural networks have demonstrated outstanding effectiveness in a large amount of applications.However, recent works have shown that neural networks are susceptible to adversarial examples, indicating possible flaws intrinsic to the network structures. To address this problem and improve the robustness …
Improved SNNs for speech classification with PyTorch.
problem Training convolutional spiking neural networks efficiently.
method Supervised training using backpropagation through time, surrogate gradient, and novel connections.
result Achieved nearly state-of-the-art accuracy on speech classification task.
Simplified plug-in loss approximates EDL for reliable uncertainty estimation.
problem Efficient and reliable uncertainty estimation in real-world sensor-based learning systems.
method Approximate Dirichlet expected objectives with plug-in losses evaluated at the Dirichlet mean.
result Plug-in losses provide comparable predictive accuracy and selective prediction performance to classical EDL, while being simpler to implement.
Improved TTS style transfer across disjoint datasets with adversarial cycle consistency.
problem Suboptimal TTS style transfer on disjoint datasets with underrepresented styles.
method Adversarial cycle consistency training with paired and unpaired triplets.
result 78% improvement in style transfer with minimal reduction in fidelity and naturalness.
Large deep neural networks are powerful, but exhibit undesirable behaviors such as memorization and sensitivity to adversarial examples. In this work, we propose mixup, a simple learning principle to alleviate these issues. In essence, mixup trains a neural network on convex combinations of pairs of examples and their …
Method uses Seq2Seq learning to automatically generate recovery commands for ICT systems.
problem Manual decision-making for recovery commands is time-consuming and error-prone.
method Seq2Seq neural network model trained on past logs and commands.
result The model can estimate accurate recovery commands from new failures.
Improves ASR accuracy for domain mismatch using machine translation.
problem Domain mismatch in ASR systems leads to suboptimal results.
method Machine translation to map out-of-domain ASR errors to in-domain terms.
result 7% absolute improvement in word error rate, 4 point BLEU score improvement.
Novel BCI system classifies imagined speech with high accuracy.
problem Classifying imagined speech from brain signals.
method Hierarchical deep learning with CNN and autoencoder.
result Achieved 83.42% average accuracy across six phonological tasks.
Activates speech DNNs to generate understandable examples.
problem Difficulty in understanding DNN classifications for speech.
method Activation maximization to generate speech samples.
result Activation maximization can generate understandable speech samples.
Improved human-like speech synthesis using WaveGAN extensions.
problem Generating realistic speech from limited data.
method Extensions to WaveGAN for audio generation, including preprocessing, skip connections, and progressive structures.
result Moderate improvement in human likeness (Cohen's d coefficient of 0.65) in synthetic speech samples.
The problem of identifying voice commands has always been a challenge due to the presence of noise and variability in speed, pitch, etc. We will compare the efficacies of several neural network architectures for the speech recognition problem. In particular, we will build a model to determine whether a one second audio…
AudioPaLM combines text and speech models to improve speech processing and translation.
problem Improving speech understanding and generation models.
method Unified multimodal architecture combining text and speech models.
result AudioPaLM significantly outperforms existing speech translation systems.
To solve a text-based game, an agent needs to formulate valid text commands for a given context and find the ones that lead to success. Recent attempts at solving text-based games with deep reinforcement learning have focused on the latter, i.e., learning to act optimally when valid actions are known in advance. In thi…
The paper proposes an on-line monitoring framework for continuous real-time safety/security in learning-based control systems (specifically application to a unmanned ground vehicle). We monitor validity of mappings from sensor inputs to actuator commands, controller-focused anomaly detection (CFAM), and from actuator c…
Improved language identification accuracy through signal combination methods.
problem Enhancing speech recognition accuracy across multiple languages.
method Combining low-level acoustic signals with language-specific recognizer signals using lattice-based ensemble models and deep neural networks.
result Deep neural network model outperforms lattice-based ensemble model, reducing error rate from 5.5% to 4.3%.
The problem of keyword spotting i.e. identifying keywords in a real-time audio stream is mainly solved by applying a neural network over successive sliding windows. Due to the difficulty of the task, baseline models are usually large, resulting in a high computational cost and energy consumption level. We propose a new…
Current UAV-recorded datasets are mostly limited to action recognition and object tracking, whereas the gesture signals datasets were mostly recorded in indoor spaces. Currently, there is no outdoor recorded public video dataset for UAV commanding signals. Gesture signals can be effectively used with UAVs by leveraging…
We present Listen, Attend and Spell (LAS), a neural network that learns to transcribe speech utterances to characters. Unlike traditional DNN-HMM models, this model learns all the components of a speech recognizer jointly. Our system has two components: a listener and a speller. The listener is a pyramidal recurrent ne…
Study shows current metrics for audio adversarial examples are unreliable for human perception.
problem The reliability of metrics for evaluating audio adversarial examples.
method Analytical framework and human evaluation experiment.
result Current metrics for audio adversarial examples are not reliable for human perception.
Large speech dataset for commercial use with 9.98% word error rate.
problem Creating a diverse speech recognition dataset for commercial purposes.
method Internet search for licensed audio data with transcriptions, training model on the dataset.
result Model trained on dataset achieves 9.98% word error rate on Librispeech's test-clean test set.
Large-scale datasets have played a significant role in progress of neural network and deep learning areas. YouTube-8M is such a benchmark dataset for general multi-label video classification. It was created from over 7 million YouTube videos (450,000 hours of video) and includes video labels from a vocabulary of 4716 c…
Improved anger detection in speech using transfer learning from SoundNet.
problem Detecting anger in speech with limited emotion datasets.
method Transfer learning from SoundNet, a multimodal audio classifier trained on video data.
result Improved performance and generalization on various speech emotion datasets.
Deep Neural Networks are powerful models that attained remarkable results on a variety of tasks. These models are shown to be extremely efficient when training and test data are drawn from the same distribution. However, it is not clear how a network will act when it is fed with an out-of-distribution example. In this …
ADDoG improves cross-dataset speech emotion recognition.
problem Cross-dataset speech emotion recognition failure.
method ADDoG uses an iterative approach to move representations closer together across datasets.
result ADDoG and MADDoG improve cross-dataset speech emotion recognition.
State-level minimum Bayes risk (sMBR) training has become the de facto standard for sequence-level training of speech recognition acoustic models. It has an elegant formulation using the expectation semiring, and gives large improvements in word error rate (WER) over models trained solely using cross-entropy (CE) or co…
ADReSS Challenge at INTERSPEECH 2020 benchmarks speech recognition for Alzheimer's dementia.
problem Automated recognition of Alzheimer's dementia from spontaneous speech.
method Provides a benchmark speech dataset, defines two tasks (classification and regression), and presents baseline models.
result Demonstrates the feasibility of automated speech recognition for Alzheimer's dementia.
Enhances speech quality in noisy environments using symbolic sequential modeling.
problem Improving speech quality in noisy conditions.
method Incorporates symbolic sequential modeling into speech enhancement framework.
result Significant improvement in speech quality metrics (PESQ, STOI) on TIMIT dataset.
In recent years, China, the United States and other countries, Google and other high-tech companies have increased investment in artificial intelligence. Deep learning is one of the current artificial intelligence research's key areas. This paper analyzes and summarizes the latest progress and future research direction…
Paper introduces WHAM! dataset for realistic speech separation in noisy environments.
problem Real-world speech separation challenges with noise.
method Created WHAM! dataset with real ambient noise, evaluated various speech separation architectures.
result Speech separation architectures show robustness to noise, but performance decreases.
ClovaCall introduces a new Korean call speech corpus for contact centers.
problem Lack of large-scale call-based speech corpora for Korean dialog scenarios.
method Development of a new large-scale Korean call-based speech corpus (ClovaCall) in a restaurant reservation domain.
result Validation of the dataset with ASR models shows its effectiveness.
Bitcoin's attention is linked to Google Trends data, not general uncertainty.
problem Bitcoin's correlation with Google Trends data was previously misunderstood.
method Analyzed bidirectional relationships between Bitcoin returns and Google Trends attention over six days.
result Information flows from Bitcoin volatility to Google Trends attention, not the other way.
Speech emotion recognition system using features and text.
problem Improving accuracy in emotion recognition from speech.
method Used speech features (Spectrogram, MFCC) and text, trained Deep Neural Networks.
result Combined MFCC-Text CNN model achieved highest accuracy.
Speech datasets for identifying Alzheimer's disease (AD) are generally restricted to participants performing a single task, e.g. describing an image shown to them. As a result, models trained on linguistic features derived from such datasets may not be generalizable across tasks. Building on prior work demonstrating th…
Study tackles hate speech against journalists on social media.
problem Hate speech against journalists on social media remains prevalent despite efforts.
method Defined journalist-specific hate speech, annotated tweets, trained deep learning models, and proposed an ensemble model.
result Proposed ensemble model outperforms individual models in detecting journalist-targeted hate speech.
Regularization is important for end-to-end speech models, since the models are highly flexible and easy to overfit. Data augmentation and dropout has been important for improving end-to-end models in other domains. However, they are relatively under explored for end-to-end speech models. Therefore, we investigate the e…
Improved far-field speaker verification for short utterances in noisy conditions.
problem Challenges in speaker verification on short utterances in uncontrolled noisy environments.
method Used deep neural network architectures (TDNN and ResNet) and experimented with various embedding extractors and training procedures.
result ResNet architectures outperform x-vector approach in speaker verification quality for both long and short utterances.
Google Trends data improves economic forecasts of private consumption.
problem Improving economic forecasts of private consumption.
method Machine learning techniques applied to categorized Google search data.
result Google data can identify patterns to generate a leading indicator in real time.
AV-CPL uses continuous pseudo-labels for AVSR combining labeled and unlabeled data.
problem Improving AVSR performance with labeled and unlabeled data.
method Semi-supervised method using continuous pseudo-labels generated by the same AVSR model.
result Significant improvements in VSR performance on LRS3 dataset.
Paper proposes a new speech representation benchmark and model.
problem Lack of benchmarks for comparing speech representations.
method Unsupervised triplet-loss objective for training a universal non-semantic speech representation.
result Proposed representation outperforms other models on benchmark and transfer learning tasks.
A dataset for detecting online hate speech from YouTube and Reddit comments.
problem Detecting and preventing hate speech on social media platforms.
method Created a dataset with two variants: binary and multi-label, based on YouTube and Reddit comments, using Figure-Eight crowdsourcing platform.
result Demonstrated that even a small amount of labelled data can help detect hate speech occurrences.
NIST CTS Superset offers a large dataset for telephony speaker recognition.
problem Lack of a large-scale, uniform dataset for telephony speaker recognition.
method Compilation of speech segments from multiple corpora, including Greybeard, Switchboard, and Mixer series.
result Results on the NIST 2020 CTS Speaker Recognition Challenge serve as a reference baseline.