We propose a new representation learning solution for the classification of cognitive load based on electroencephalogram (EEG). Our method integrates both time and frequency domains by first passing the raw EEG signals through the convolutional encoder to obtain the time domain representations. Next, we measure the power spectral density for all five EEG frequency bands and generate the channel power values as 2-D images referred to as multispectral topography maps. These multispectral topography maps are then fed to a separate encoder to obtain the representations in frequency domain. Our solution employs a multidomain attention module that maps these domain-specific embeddings onto a shared embedding space to emphasize more on important interdomain relationships to enhance the representations for cognitive load classification. In addition, we incorporate an orthogonal projection constraint during the training of our method to effectively increase the interclass distances while improving intraclass clustering. This enhancement allows efficient discrimination between different cognitive states and aids in better grouping of similar states within the feature space. We validate the effectiveness of our model through extensive experiments on two public EEG datasets, CL-Drive and CLARE for cognitive load classification. Our results demonstrate the superiority of our multidomain approach over the traditional single-domain techniques. Moreover, we conduct ablation and sensitivity analyses to assess the impact of various components of our method. Finally, robustness experiments on different amounts of added noise demonstrate the stability of our method compared to other state-of-the-art solutions.

Despite considerable advancements in deep learning, optimizing respiratory sound classification (RSC) models remains challenging. This is partly due to the bias from inconsistent respiratory sound recording processes and imbalanced representation of demographics, which leads to poor performance when a model trained with the dataset is applied to real-world use cases. RSC datasets usually include various metadata attributes describing certain aspects of the data, such as environmental and demographic factors. To address the issues caused by bias, we take advantage of the metadata provided by RSC datasets and explore approaches for metadata-guided domain adaptation. We thoroughly evaluate the effect of various metadata attributes and their combinations on a simple metadata-guided approach, but also introduce a more advanced method that adaptively rescales the suitable metadata combinations to improve domain adaptation during training. The findings indicate a robust reduction in domain dependency and improvement in detection accuracy on both ICBHI and our own dataset. Specifically, the implementation of our proposed methods led to an improved score of 84.97%, which signifies a substantial enhancement of 7.37% compared to the baseline model.

Existing approaches for learning representations of time-series keep the temporal arrangement of the time-steps intact with the presumption that the original order is the most optimal for learning. However, non-adjacent sections of real-world time-series may have strong dependencies. Accordingly, we raise the question: Is there an alternative arrangement for time-series which could enable more effective representation learning? To address this, we propose a simple plug-and-play neural network layer called Segment, Shuffle, and Stitch (S3) designed to improve representation learning in time-series models. S3 works by creating non-overlapping segments from the original sequence and shuffling them in a learned manner that is optimal for the task at hand. It then re-attaches the shuffled segments back together and performs a learned weighted sum with the original input to capture both the newly shuffled sequence along with the original sequence. S3 is modular and can be stacked to achieve different levels of granularity, and can be added to many forms of neural architectures including CNNs or Transformers with negligible computation overhead. Through extensive experiments on several datasets and state-of-the-art baselines, we show that incorporating S3 results in significant improvements for the tasks of time-series classification, forecasting, and anomaly detection, improving performance on certain datasets by up to 68\%. We also show that S3 makes the learning more stable with a smoother training loss curve and loss landscape compared to the original baseline. The code is available at https://github.com/shivam-grover/S3-TimeSeries.

Atrial fibrillation (AF) is the most prevalent cardiac arrhythmia. The atrial beat is irregular during AF, which causes blood flow hardly. This may cause blood clot formation and cardioembolic strokes. Computer-aided devices may assist cardiologists in diagnosing heart rhythm disorders better. From this viewpoint, we attempt to identify the premature atrial complexes (PACs) to predict the occurrence of AF by using electrocardiogram (ECG) spectrograms. Convolutional neural networks (CNN) models such as ResNet and Wide-ResNet are used to predict the prelude of AF. Regularization constraints are used to deal with the imbalanced and small number of samples in the minority premature AF class. Sensitivity regularization investigates small variations in premature AF samples. It highlights more representative features that distinguish the PACs from the normal rhythm. On the other hand, orthogonality regularization removes the interference between negatively correlated feature weights. It places constraints on capturing similar patterns with slight differences. This constraint allows convergence to a better feature representation with fewer weight redundancies. We propose a combination of sensitivity and orthogonality penalty terms to the cost function of ResNet to decrease the overfitting and obtain a superior representation. The re-sampling class distribution method is also utilized to mitigate the issue of imbalanced data. The proposed method shows better AF prediction for highly imbalanced data with a small number of samples.

Palmprint recognition is a challenging problem, mainly due to low quality of the patterns, variation in focal lens distance, large nonlinear deformations caused by contactless image acquisition system, and computational complexity for the large image size of typical palmprints. This paper proposes a new contactless biometric system using features of palm texture extracted from the single hand image acquired from a digital camera. In this work, we propose to apply convolutional neural network (CNN) for palmprint recognition. The results demonstrate that the extracted local and general features using CNN are invariant to image rotation, translation, and scale variations.