Huy Phan

Shubhr Singh, Christian J. Steinmetz, Emmanouil Benetos et al.

Deep learning models such as CNNs and Transformers have achieved impressive performance for end-to-end audio tagging. Recent works have shown that despite stacking multiple layers, the receptive field of CNNs remains severely limited. Transformers on the other hand are able to map global context through self-attention, but treat the spectrogram as a sequence of patches which is not flexible enough to capture irregular audio objects. In this work, we treat the spectrogram in a more flexible way by considering it as graph structure and process it with a novel graph neural architecture called ATGNN. ATGNN not only combines the capability of CNNs with the global information sharing ability of Graph Neural Networks, but also maps semantic relationships between learnable class embeddings and corresponding spectrogram regions. We evaluate ATGNN on two audio tagging tasks, where it achieves 0.585 mAP on the FSD50K dataset and 0.335 mAP on the AudioSet-balanced dataset, achieving comparable results to Transformer based models with significantly lower number of learnable parameters.

Phairot Autthasan, Rattanaphon Chaisaen, Huy Phan et al.

Recent advances in deep learning (DL) have significantly impacted motor imagery (MI)-based brain-computer interface (BCI) systems, enhancing the decoding of electroencephalography (EEG) signals. However, most studies struggle to identify discriminative patterns across subjects during MI tasks, limiting MI classification performance. In this article, we propose MixNet, a novel classification framework designed to overcome this limitation by utilizing spectral-spatial signals from MI data, along with a multitask learning architecture named MIN2Net, for classification. Here, the spectral-spatial signals are generated using the filter-bank common spatial patterns (FBCSPs) method on MI data. Since the multitask learning architecture is used for the classification task, the learning in each task may exhibit different generalization rates and potential overfitting across tasks. To address this issue, we implement adaptive gradient blending, simultaneously regulating multiple loss weights and adjusting the learning pace for each task based on its generalization/overfitting tendencies. Experimental results on six benchmark data sets of different data sizes demonstrate that MixNet consistently outperforms all state-of-the-art algorithms in subject-dependent and -independent settings. Finally, the low-density EEG MI classification results show that MixNet outperforms all state-of-the-art algorithms, offering promising implications for Internet of Thing (IoT) applications, such as lightweight and portable EEG wearable devices based on low-density montages.

Phu X. Nguyen, Konstantinos Kontras, Wei Dai et al.

Electrocardiograms (ECGs) are widely used non-invasive measurements of cardiac activity and play a central role in clinical diagnosis. Recent multimodal approaches align ECG signals with clinical reports to incorporate diagnostic semantics, but clinical reports often fail to preserve the rich physiological structure of ECG waveforms, particularly across multiple levels of abstraction ranging from coarse diagnostic categories to fine-grained morphology. To address this limitation, we formulate ECG representation learning from an information-theoretic perspective and derive a tractable objective that jointly preserves signal structure and integrates clinical semantics. Based on this principle, we propose \textbf{MERIT} (Multimodal ECG Representation via Information Theory), a dual-branch pretraining framework combining masked ECG modeling with ECG--text contrastive alignment. Extensive experiments on PTB-XL and additional benchmarks demonstrate consistent improvements over prior methods, including gains exceeding $3%$ F1 on PTB-XL All and $5%$ F1 on SubClass classification. In zero-shot evaluation, MERIT further improves performance by up to $ +2.66\%$ AUC and $ +2.11\%$ F1 on PTB-XL SubClass, while also demonstrating robustness under multiple distribution-shift settings. Moreover, leveraging the learned ECG representations for ECG-conditioned clinical text generation with large language models improves text quality across several metrics, including ROUGE and METEOR. Together, these results demonstrate that MERIT learns more informative and clinically meaningful ECG representations, particularly for fine-grained clinical applications.

Huy Phan, Oliver Y. Chén, Philipp Koch et al.

Background: Despite recent significant progress in the development of automatic sleep staging methods, building a good model still remains a big challenge for sleep studies with a small cohort due to the data-variability and data-inefficiency issues. This work presents a deep transfer learning approach to overcome these issues and enable transferring knowledge from a large dataset to a small cohort for automatic sleep staging. Methods: We start from a generic end-to-end deep learning framework for sequence-to-sequence sleep staging and derive two networks as the means for transfer learning. The networks are first trained in the source domain (i.e. the large database). The pretrained networks are then finetuned in the target domain (i.e. the small cohort) to complete knowledge transfer. We employ the Montreal Archive of Sleep Studies (MASS) database consisting of 200 subjects as the source domain and study deep transfer learning on three different target domains: the Sleep Cassette subset and the Sleep Telemetry subset of the Sleep-EDF Expanded database, and the Surrey-cEEGrid database. The target domains are purposely adopted to cover different degrees of data mismatch to the source domains. Results: Our experimental results show significant performance improvement on automatic sleep staging on the target domains achieved with the proposed deep transfer learning approach. Conclusions: These results suggest the efficacy of the proposed approach in addressing the above-mentioned data-variability and data-inefficiency issues. Significance: As a consequence, it would enable one to improve the quality of automatic sleep staging models when the amount of data is relatively small. The source code and the pretrained models are available at http://github.com/pquochuy/sleep_transfer_learning.

Thi Ngoc Tho Nguyen, Douglas L. Jones, Karn N. Watcharasupat et al.

Polyphonic sound event localization and detection (SELD) has many practical applications in acoustic sensing and monitoring. However, the development of real-time SELD has been limited by the demanding computational requirement of most recent SELD systems. In this work, we introduce SALSA-Lite, a fast and effective feature for polyphonic SELD using microphone array inputs. SALSA-Lite is a lightweight variation of a previously proposed SALSA feature for polyphonic SELD. SALSA, which stands for Spatial Cue-Augmented Log-Spectrogram, consists of multichannel log-spectrograms stacked channelwise with the normalized principal eigenvectors of the spectrotemporally corresponding spatial covariance matrices. In contrast to SALSA, which uses eigenvector-based spatial features, SALSA-Lite uses normalized inter-channel phase differences as spatial features, allowing a 30-fold speedup compared to the original SALSA feature. Experimental results on the TAU-NIGENS Spatial Sound Events 2021 dataset showed that the SALSA-Lite feature achieved competitive performance compared to the full SALSA feature, and significantly outperformed the traditional feature set of multichannel log-mel spectrograms with generalized cross-correlation spectra. Specifically, using SALSA-Lite features increased localization-dependent F1 score and class-dependent localization recall by 15% and 5%, respectively, compared to using multichannel log-mel spectrograms with generalized cross-correlation spectra.

Marco Comunità, Huy Phan, Joshua D. Reiss

Footsteps are among the most ubiquitous sound effects in multimedia applications. There is substantial research into understanding the acoustic features and developing synthesis models for footstep sound effects. In this paper, we present a first attempt at adopting neural synthesis for this task. We implemented two GAN-based architectures and compared the results with real recordings as well as six traditional sound synthesis methods. Our architectures reached realism scores as high as recorded samples, showing encouraging results for the task at hand.

Huy Phan, Alfred Mertins, Mathias Baumert

Background: Despite the tremendous progress recently made towards automatic sleep staging in adults, it is currently unknown if the most advanced algorithms generalize to the pediatric population, which displays distinctive characteristics in overnight polysomnography (PSG). Methods: To answer the question, in this work, we conduct a large-scale comparative study on the state-of-the-art deep learning methods for pediatric automatic sleep staging. Six different deep neural networks with diverging features are adopted to evaluate a sample of more than 1,200 children across a wide spectrum of obstructive sleep apnea (OSA) severity. Results: Our experimental results show that the individual performance of automated pediatric sleep stagers when evaluated on new subjects is equivalent to the expert-level one reported on adults. Combining the six stagers into ensemble models further boosts the staging accuracy, reaching an overall accuracy of 88.8%, a Cohen's kappa of 0.852, and a macro F1-score of 85.8%. At the same time, the ensemble models lead to reduced predictive uncertainty. The results also show that the studied algorithms and their ensembles are robust to concept drift when the training and test data were recorded seven months apart and after clinical intervention. Conclusion: However, we show that the improvements in the staging performance are not necessarily clinically significant although the ensemble models lead to more favorable clinical measures than the six standalone models. Significance: Detailed analyses further demonstrate "almost perfect" agreement between the automatic stagers to one another and their similar patterns on the staging errors, suggesting little room for improvement.

Huy Phan, Kaare Mikkelsen, Oliver Y. Chén et al.

Background: Black-box skepticism is one of the main hindrances impeding deep-learning-based automatic sleep scoring from being used in clinical environments. Methods: Towards interpretability, this work proposes a sequence-to-sequence sleep-staging model, namely SleepTransformer. It is based on the transformer backbone and offers interpretability of the model's decisions at both the epoch and sequence level. We further propose a simple yet efficient method to quantify uncertainty in the model's decisions. The method, which is based on entropy, can serve as a metric for deferring low-confidence epochs to a human expert for further inspection. Results: Making sense of the transformer's self-attention scores for interpretability, at the epoch level, the attention scores are encoded as a heat map to highlight sleep-relevant features captured from the input EEG signal. At the sequence level, the attention scores are visualized as the influence of different neighboring epochs in an input sequence (i.e. the context) to recognition of a target epoch, mimicking the way manual scoring is done by human experts. Conclusion: Additionally, we demonstrate that SleepTransformer performs on par with existing methods on two databases of different sizes. Significance: Equipped with interpretability and the ability of uncertainty quantification, SleepTransformer holds promise for being integrated into clinical settings.

Huy Phan, Huy Le Nguyen, Oliver Y. Chén et al.

We propose in this work a multi-view learning approach for audio and music classification. Considering four typical low-level representations (i.e. different views) commonly used for audio and music recognition tasks, the proposed multi-view network consists of four subnetworks, each handling one input types. The learned embedding in the subnetworks are then concatenated to form the multi-view embedding for classification similar to a simple concatenation network. However, apart from the joint classification branch, the network also maintains four classification branches on the single-view embedding of the subnetworks. A novel method is then proposed to keep track of the learning behavior on the classification branches and adapt their weights to proportionally blend their gradients for network training. The weights are adapted in such a way that learning on a branch that is generalizing well will be encouraged whereas learning on a branch that is overfitting will be slowed down. Experiments on three different audio and music classification tasks show that the proposed multi-view network not only outperforms the single-view baselines but also is superior to the multi-view baselines based on concatenation and late fusion.

Phairot Autthasan, Rattanaphon Chaisaen, Thapanun Sudhawiyangkul et al.

Advances in the motor imagery (MI)-based brain-computer interfaces (BCIs) allow control of several applications by decoding neurophysiological phenomena, which are usually recorded by electroencephalography (EEG) using a non-invasive technique. Despite great advances in MI-based BCI, EEG rhythms are specific to a subject and various changes over time. These issues point to significant challenges to enhance the classification performance, especially in a subject-independent manner. To overcome these challenges, we propose MIN2Net, a novel end-to-end multi-task learning to tackle this task. We integrate deep metric learning into a multi-task autoencoder to learn a compact and discriminative latent representation from EEG and perform classification simultaneously. This approach reduces the complexity in pre-processing, results in significant performance improvement on EEG classification. Experimental results in a subject-independent manner show that MIN2Net outperforms the state-of-the-art techniques, achieving an F1-score improvement of 6.72%, and 2.23% on the SMR-BCI, and OpenBMI datasets, respectively. We demonstrate that MIN2Net improves discriminative information in the latent representation. This study indicates the possibility and practicality of using this model to develop MI-based BCI applications for new users without the need for calibration.

Lam Pham, Huy Phan, Ross King et al.

This paper presents an inception-based deep neural network for detecting lung diseases using respiratory sound input. Recordings of respiratory sound collected from patients are firstly transformed into spectrograms where both spectral and temporal information are well presented, referred to as front-end feature extraction. These spectrograms are then fed into the proposed network, referred to as back-end classification, for detecting whether patients suffer from lung-relevant diseases. Our experiments, conducted over the ICBHI benchmark meta-dataset of respiratory sound, achieve competitive ICBHI scores of 0.53/0.45 and 0.87/0.85 regarding respiratory anomaly and disease detection, respectively.

Huy Phan, Huy Le Nguyen, Oliver Y. Chén et al.

Existing generative adversarial networks (GANs) for speech enhancement solely rely on the convolution operation, which may obscure temporal dependencies across the sequence input. To remedy this issue, we propose a self-attention layer adapted from non-local attention, coupled with the convolutional and deconvolutional layers of a speech enhancement GAN (SEGAN) using raw signal input. Further, we empirically study the effect of placing the self-attention layer at the (de)convolutional layers with varying layer indices as well as at all of them when memory allows. Our experiments show that introducing self-attention to SEGAN leads to consistent improvement across the objective evaluation metrics of enhancement performance. Furthermore, applying at different (de)convolutional layers does not significantly alter performance, suggesting that it can be conveniently applied at the highest-level (de)convolutional layer with the smallest memory overhead.

Huy Phan, Lam Pham, Philipp Koch et al.

Audio event localization and detection (SELD) have been commonly tackled using multitask models. Such a model usually consists of a multi-label event classification branch with sigmoid cross-entropy loss for event activity detection and a regression branch with mean squared error loss for direction-of-arrival estimation. In this work, we propose a multitask regression model, in which both (multi-label) event detection and localization are formulated as regression problems and use the mean squared error loss homogeneously for model training. We show that the common combination of heterogeneous loss functions causes the network to underfit the data whereas the homogeneous mean squared error loss leads to better convergence and performance. Experiments on the development and validation sets of the DCASE 2020 SELD task demonstrate that the proposed system also outperforms the DCASE 2020 SELD baseline across all the detection and localization metrics, reducing the overall SELD error (the combined metric) by approximately 10% absolute.

Huy Phan, Oliver Y. Chén, Minh C. Tran et al.

Automating sleep staging is vital to scale up sleep assessment and diagnosis to serve millions experiencing sleep deprivation and disorders and enable longitudinal sleep monitoring in home environments. Learning from raw polysomnography signals and their derived time-frequency image representations has been prevalent. However, learning from multi-view inputs (e.g., both the raw signals and the time-frequency images) for sleep staging is difficult and not well understood. This work proposes a sequence-to-sequence sleep staging model, XSleepNet, that is capable of learning a joint representation from both raw signals and time-frequency images. Since different views may generalize or overfit at different rates, the proposed network is trained such that the learning pace on each view is adapted based on their generalization/overfitting behavior. In simple terms, the learning on a particular view is speeded up when it is generalizing well and slowed down when it is overfitting. View-specific generalization/overfitting measures are computed on-the-fly during the training course and used to derive weights to blend the gradients from different views. As a result, the network is able to retain the representation power of different views in the joint features which represent the underlying distribution better than those learned by each individual view alone. Furthermore, the XSleepNet architecture is principally designed to gain robustness to the amount of training data and to increase the complementarity between the input views. Experimental results on five databases of different sizes show that XSleepNet consistently outperforms the single-view baselines and the multi-view baseline with a simple fusion strategy. Finally, XSleepNet also outperforms prior sleep staging methods and improves previous state-of-the-art results on the experimental databases.

Huy Phan, Oliver Y. Chén, Lam Pham et al.

Acoustic scenes are rich and redundant in their content. In this work, we present a spatio-temporal attention pooling layer coupled with a convolutional recurrent neural network to learn from patterns that are discriminative while suppressing those that are irrelevant for acoustic scene classification. The convolutional layers in this network learn invariant features from time-frequency input. The bidirectional recurrent layers are then able to encode the temporal dynamics of the resulting convolutional features. Afterwards, a two-dimensional attention mask is formed via the outer product of the spatial and temporal attention vectors learned from two designated attention layers to weigh and pool the recurrent output into a final feature vector for classification. The network is trained with between-class examples generated from between-class data augmentation. Experiments demonstrate that the proposed method not only outperforms a strong convolutional neural network baseline but also sets new state-of-the-art performance on the LITIS Rouen dataset.

Huy Phan, Fernando Andreotti, Navin Cooray et al.

Automatic sleep staging has been often treated as a simple classification problem that aims at determining the label of individual target polysomnography (PSG) epochs one at a time. In this work, we tackle the task as a sequence-to-sequence classification problem that receives a sequence of multiple epochs as input and classifies all of their labels at once. For this purpose, we propose a hierarchical recurrent neural network named SeqSleepNet. At the epoch processing level, the network consists of a filterbank layer tailored to learn frequency-domain filters for preprocessing and an attention-based recurrent layer designed for short-term sequential modelling. At the sequence processing level, a recurrent layer placed on top of the learned epoch-wise features for long-term modelling of sequential epochs. The classification is then carried out on the output vectors at every time step of the top recurrent layer to produce the sequence of output labels. Despite being hierarchical, we present a strategy to train the network in an end-to-end fashion. We show that the proposed network outperforms state-of-the-art approaches, achieving an overall accuracy, macro F1-score, and Cohen's kappa of 87.1%, 83.3%, and 0.815 on a publicly available dataset with 200 subjects.

Huy Phan, Fernando Andreotti, Navin Cooray et al.

Correctly identifying sleep stages is important in diagnosing and treating sleep disorders. This work proposes a joint classification-and-prediction framework based on CNNs for automatic sleep staging, and, subsequently, introduces a simple yet efficient CNN architecture to power the framework. Given a single input epoch, the novel framework jointly determines its label (classification) and its neighboring epochs' labels (prediction) in the contextual output. While the proposed framework is orthogonal to the widely adopted classification schemes, which take one or multiple epochs as contextual inputs and produce a single classification decision on the target epoch, we demonstrate its advantages in several ways. First, it leverages the dependency among consecutive sleep epochs while surpassing the problems experienced with the common classification schemes. Second, even with a single model, the framework has the capacity to produce multiple decisions, which are essential in obtaining a good performance as in ensemble-of-models methods, with very little induced computational overhead. Probabilistic aggregation techniques are then proposed to leverage the availability of multiple decisions. We conducted experiments on two public datasets: Sleep-EDF Expanded with 20 subjects, and Montreal Archive of Sleep Studies dataset with 200 subjects. The proposed framework yields an overall classification accuracy of 82.3% and 83.6%, respectively. We also show that the proposed framework not only is superior to the baselines based on the common classification schemes but also outperforms existing deep-learning approaches. To our knowledge, this is the first work going beyond the standard single-output classification to consider multitask neural networks for automatic sleep staging. This framework provides avenues for further studies of different neural-network architectures for automatic sleep staging.

Huy Phan, Philipp Koch, Ian McLoughlin et al.

This paper presents a methodology for early detection of audio events from audio streams. Early detection is the ability to infer an ongoing event during its initial stage. The proposed system consists of a novel inference step coupled with dual parallel tailored-loss deep neural networks (DNNs). The DNNs share a similar architecture except for their loss functions, i.e. weighted loss and multitask loss, which are designed to efficiently cope with issues common to audio event detection. The inference step is newly introduced to make use of the network outputs for recognizing ongoing events. The monotonicity of the detection function is required for reliable early detection, and will also be proved. Experiments on the ITC-Irst database show that the proposed system achieves state-of-the-art detection performance. Furthermore, even partial events are sufficient to achieve good performance similar to that obtained when an entire event is observed, enabling early event detection.

Huy Phan, Martin Krawczyk-Becker, Timo Gerkmann et al.

This report presents our audio event detection system submitted for Task 2, "Detection of rare sound events", of DCASE 2017 challenge. The proposed system is based on convolutional neural networks (CNNs) and deep neural networks (DNNs) coupled with novel weighted and multi-task loss functions and state-of-the-art phase-aware signal enhancement. The loss functions are tailored for audio event detection in audio streams. The weighted loss is designed to tackle the common issue of imbalanced data in background/foreground classification while the multi-task loss enables the networks to simultaneously model the class distribution and the temporal structures of the target events for recognition. Our proposed systems significantly outperform the challenge baseline, improving F-score from 72.7% to 90.0% and reducing detection error rate from 0.53 to 0.18 on average on the development data. On the evaluation data, our submission obtains an average F1-score of 88.3% and an error rate of 0.22 which are significantly better than those obtained by the DCASE baseline (i.e. an F1-score of 64.1% and an error rate of 0.64).

Huy Phan, Philipp Koch, Fabrice Katzberg et al.

We introduce in this work an efficient approach for audio scene classification using deep recurrent neural networks. An audio scene is firstly transformed into a sequence of high-level label tree embedding feature vectors. The vector sequence is then divided into multiple subsequences on which a deep GRU-based recurrent neural network is trained for sequence-to-label classification. The global predicted label for the entire sequence is finally obtained via aggregation of subsequence classification outputs. We will show that our approach obtains an F1-score of 97.7% on the LITIS Rouen dataset, which is the largest dataset publicly available for the task. Compared to the best previously reported result on the dataset, our approach is able to reduce the relative classification error by 35.3%.

Huy Phan, Philipp Koch, Fabrice Katzberg et al.

There is a common observation that audio event classification is easier to deal with than detection. So far, this observation has been accepted as a fact and we lack of a careful analysis. In this paper, we reason the rationale behind this fact and, more importantly, leverage them to benefit the audio event detection task. We present an improved detection pipeline in which a verification step is appended to augment a detection system. This step employs a high-quality event classifier to postprocess the benign event hypotheses outputted by the detection system and reject false alarms. To demonstrate the effectiveness of the proposed pipeline, we implement and pair up different event detectors based on the most common detection schemes and various event classifiers, ranging from the standard bag-of-words model to the state-of-the-art bank-of-regressors one. Experimental results on the ITC-Irst dataset show significant improvements to detection performance. More importantly, these improvements are consistent for all detector-classifier combinations.

Lars Hertel, Huy Phan, Alfred Mertins

We trained a deep all-convolutional neural network with masked global pooling to perform single-label classification for acoustic scene classification and multi-label classification for domestic audio tagging in the DCASE-2016 contest. Our network achieved an average accuracy of 84.5% on the four-fold cross-validation for acoustic scene recognition, compared to the provided baseline of 72.5%, and an average equal error rate of 0.17 for domestic audio tagging, compared to the baseline of 0.21. The network therefore improves the baselines by a relative amount of 17% and 19%, respectively. The network only consists of convolutional layers to extract features from the short-time Fourier transform and one global pooling layer to combine those features. It particularly possesses neither fully-connected layers, besides the fully-connected output layer, nor dropout layers.

Huy Phan, Lars Hertel, Marco Maass et al.

This report describes our submissions to Task2 and Task3 of the DCASE 2016 challenge. The systems aim at dealing with the detection of overlapping audio events in continuous streams, where the detectors are based on random decision forests. The proposed forests are jointly trained for classification and regression simultaneously. Initially, the training is classification-oriented to encourage the trees to select discriminative features from overlapping mixtures to separate positive audio segments from the negative ones. The regression phase is then carried out to let the positive audio segments vote for the event onsets and offsets, and therefore model the temporal structure of audio events. One random decision forest is specifically trained for each event category of interest. Experimental results on the development data show that our systems significantly outperform the baseline on the Task2 evaluation while they are inferior to the baseline in the Task3 evaluation.

Huy Phan, Lars Hertel, Marco Maass et al.

We present in this paper an efficient approach for acoustic scene classification by exploring the structure of class labels. Given a set of class labels, a category taxonomy is automatically learned by collectively optimizing a clustering of the labels into multiple meta-classes in a tree structure. An acoustic scene instance is then embedded into a low-dimensional feature representation which consists of the likelihoods that it belongs to the meta-classes. We demonstrate state-of-the-art results on two different datasets for the acoustic scene classification task, including the DCASE 2013 and LITIS Rouen datasets.

Huy Phan, Marco Maass, Lars Hertel et al.

We introduce a new learned descriptor for audio signals which is efficient for event representation. The entries of the descriptor are produced by evaluating a set of regressors on the input signal. The regressors are class-specific and trained using the random regression forests framework. Given an input signal, each regressor estimates the onset and offset positions of the target event. The estimation confidence scores output by a regressor are then used to quantify how the target event aligns with the temporal structure of the corresponding category. Our proposed descriptor has two advantages. First, it is compact, i.e. the dimensionality of the descriptor is equal to the number of event classes. Second, we show that even simple linear classification models, trained on our descriptor, yield better accuracies on audio event classification task than not only the nonlinear baselines but also the state-of-the-art results.

Lars Hertel, Huy Phan, Alfred Mertins

Recognizing acoustic events is an intricate problem for a machine and an emerging field of research. Deep neural networks achieve convincing results and are currently the state-of-the-art approach for many tasks. One advantage is their implicit feature learning, opposite to an explicit feature extraction of the input signal. In this work, we analyzed whether more discriminative features can be learned from either the time-domain or the frequency-domain representation of the audio signal. For this purpose, we trained multiple deep networks with different architectures on the Freiburg-106 and ESC-10 datasets. Our results show that feature learning from the frequency domain is superior to the time domain. Moreover, additionally using convolution and pooling layers, to explore local structures of the audio signal, significantly improves the recognition performance and achieves state-of-the-art results.