Vinayak Abrol

Vinayak Abrol

Auto-generated profile, based on publicly available information

Vinayak Abrol

Indraprastha Institute of Information Technology Delhi, New Delhi, India

Research Focus & Metrics

Academic Career Length

14Years (2013-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 June 2020 in IEEE/ACM Transactions on Audio, Speech, and Language Processing
Verified
Journal Article
Closed access
VAVinayak Abrol
Vinayak Abrol
PSPulkit Sharma
Pulkit Sharma
Recent advancements in modeling speech and audio signals using deep neural networks have shown that systems learning both features and the classifier can be built directly from raw signal. However, the performance of such end-to-end systems for acoustic scene classification (ASC) task is still not at par with conventional systems built using spectral features. In this work, we propose a raw waveform based end-to-end ASC system using convolutional neural network. In contrast to the existing studies using a non-hierarchical model, our framework leverages the hierarchical relations between acoustic categories to improve the classification performance. To this aim, our multi-task model is trained with coarse and fine labels that correspond to different levels of abstraction. In order to ensure consistency in the encoded information via label hierarchy, the proposed framework uses a prototypical model. Such a model ensures that the learned representations at least match to one of the global categorical learned prototypes. We also employed a statistical pooling layer to aggregate hidden representations over multiple frames of the input audio signal. The statistics (mean and standard deviation) are concatenated together to form a fixed-length audio embedding. This aggregation is done via an attention module so as to guide the model's attention even in presence of relatively short or transient acoustic events. Further, the proposed framework incorporate two parallel feature processing pipelines to achieve different resolutions for extracting important acoustic cues. Various experiments on publicly available datasets are performed to demonstrate the effectiveness of the proposed framework for ASC task. Additional transfer learning experiments showed the proposed model's adaptation capability to unseen data. Network analysis and visualizations demonstrate the importance of individual modules and their impact on overall representation learning for ASC task.

30 August 2021 in IEEE Journal of Biomedical and Health Informatics
Verified
Journal Article
Closed access
PSPulkit Sharma
Pulkit Sharma
FSFarah E. Shamout
Farah E. Shamout
VAVinayak Abrol
Vinayak Abrol
DCDavid A. Clifton
David A. Clifton
Clinical time-series data retrieved from electronic medical records are widely used to build predictive models of adverse events to support resource management. Such data is often sparse and irregularly-sampled, which makes it challenging to use many common machine learning methods. Missing values may be interpolated by carrying the last value forward, or through linear regression. Gaussian process (GP) regression is also used for performing imputation, and often re-sampling of time-series at regular intervals. The use of GPs can require extensive, and likely adhoc, investigation to determine model structure, such as an appropriate covariance function. This can be challenging for multivariate real-world clinical data, in which time-series variables exhibit different dynamics to one another. In this work, we construct generative models to estimate missing values in clinical time-series data using a neural latent variable model, known as a Neural Process (NP). The NP model employs a conditional prior distribution in the latent space to learn global uncertainty in the data by modelling variations at a local level. In contrast to conventional generative modelling, this prior is not fixed and is itself learned during the training process. Thus, NP model provides the flexibility to adapt to the dynamics of the available clinical data. We propose a variant of the NP framework for efficient modelling of the mutual information between the latent and input spaces, ensuring meaningful learned priors. Experiments using the MIMIC III dataset demonstrate the effectiveness of the proposed approach as compared to conventional methods.

8 July 2020 in IEEE Transactions on Multimedia
Verified
Journal Article
Closed access
VAVinayak Abrol
Vinayak Abrol
PSPulkit Sharma
Pulkit Sharma
APArijit Patra
Arijit Patra
In this paper we propose a conditional generative modelling (CGM) approach for unsupervised disentangled representation learning using variational autoencoder (VAE). CGM employs a multimodal/categorical conditional prior distribution in the latent space to learn global uncertainty in data by modelling the variations at local level. Thus, the proposed framework enforces the model to independently estimate the inherent patterns within each category, which improves the interpretability of the latent representations learned by the VAE model. The evidence lower bound objective for training the generative model is maximized using a mutual information criterion between the global latent categorical code and the encoded inputs. Further, the approach has a built-in mechanism for bounding the information flow between the encoder and the decoder which addresses the problems of posterior collapse in conventional VAE models. Experiments on a variety of datasets demonstrate that our objective can learn disentangled representations and the proposed approach achieves competitive results in various task such as generative modelling, image classification and image denoising.

25 August 2013
Verified
Conference Paper
Closed access
VAVinayak Abrol
Vinayak Abrol
PSPulkit Sharma
Pulkit Sharma
ASAnil Kumar Sao
Anil Kumar Sao

30 June 2021 in Optimization and Engineering
Verified
Journal Article
Open access
PDF file
GUGiuseppe Ughi
Giuseppe Ughi
VAVinayak Abrol
Vinayak Abrol
JTJared Tanner
Jared Tanner
We perform a comprehensive study on the performance of derivative free optimization (DFO) algorithms for the generation of targeted black-box adversarial attacks on Deep Neural Network (DNN) classifiers assuming the perturbation energy is bounded by an $$\ell _\infty$$ ℓ ∞ constraint and the number of queries to the network is limited. This paper considers four pre-existing state-of-the-art DFO-based algorithms along with a further developed algorithm built on BOBYQA, a model-based DFO method. We compare these algorithms in a variety of settings according to the fraction of images that they successfully misclassify given a maximum number of queries to the DNN. The experiments disclose how the likelihood of finding an adversarial example depends on both the algorithm used and the setting of the attack; algorithms limiting the search of adversarial example to the vertices of the $$\ell ^\infty$$ ℓ ∞ constraint work particularly well without structural defenses, while the presented BOBYQA based algorithm works better for especially small perturbation energies. This variance in performance highlights the importance of new algorithms being compared to the state-of-the-art in a variety of settings, and the effectiveness of adversarial defenses being tested using as wide a range of algorithms as possible.

8 November 2024 in IEEE Sensors Letters
Verified
Journal Article
Closed access
RBRantu Buragohain
Rantu Buragohain
JAJejariya Ajaybhai
Jejariya Ajaybhai
KNKaran Nathwani
Karan Nathwani
VAVinayak Abrol
Vinayak Abrol
Electroencephalogram (EEG) signal classification is of utmost importance in brain-computer interface (BCI) systems. However, the inherent complex properties of EEG signals pose a challenge in their analysis and modeling. This paper proposes a novel approach of integrating wavelet scattering transform (WST) with convolutional neural network (CNN) for classifying motor imagery (MI) via EEG signals (referred as WST-CNN), capable of extracting distinctive characteristics in signals even when the data is limited. In this architecture, the first layer is non-trainable WST features with fixed initializations in WST-CNN. Furthermore, WSTs are robust to local perturbations in data, especially in the form of translation invariance, and resilient to deformations, thereby enhancing the network's reliability. The performance of the proposed idea is evaluated on the DBCIE dataset for three different scenarios: left-arm (LA) movement, right-arm (RA) movement, and simultaneous movement of both arms (BA). The BCI Competition IV-2a dataset was also employed to validate the proposed concept across four distinct motor imagery (MI) tasks, like movements in: left-hand (LH), right-hand (RH), feet (FT), and tongue (T). The classifications' performance was evaluated in terms of accuracy (η), sensitivity (S e ), specificity (S p ), and weighted F1-score, which reached up to 92.72%, 92.72%, 97.57%, and 92.75% for classifying LH, RH, FT, and T on the BCI Competition IV-2a dataset and 89.19%, 89.19%, 94.60%, and 89.33% for classifying LA, RA, and BA, on the DBCIE dataset, respectively.