arXiv subjects

Arun Pandey

Publications and source records attributed to Arun Pandey.

9 recordsLinked to original sources

arXiv2023

Duality in Multi-View Restricted Kernel Machines

We propose a unifying setting that combines existing restricted kernel machine methods into a single primal-dual multi-view framework for kernel principal component analysis in both supervised and unsupervised settings. We derive the primal and dual representations of the framework and relate different training and inference algorithms from a theoretical perspective. We show how to achieve full equivalence in primal and dual formulations by rescaling primal variables. Finally, we experimentally validate the equivalence and provide insight into the relationships between different methods on a number of time series data sets by recursively forecasting unseen test data and visualizing the learned features.

cs.LG↗

arXiv2023

D-Mag: a laboratory for studying plasma physics and diagnostics in strong magnetic fields

We have set up a diagnostic magnet (D-Mag) laboratory for a wide range of applications in plasma physics. It consists of a superconducting magnet for field strengths of up to 5.9 T. The main purpose is to provide an experimental environment for the development of plasma diagnostics for nuclear fusion studies and the investigation of dusty plasmas in strong magnetic fields. We describe in the article the setup and operation of the D-Mag. Some applications are presented for the development of plasma diagnostics, such as neutral pressure gauges and Langmuir probes that have to be operated in strong magnetic fields. Among the examples is the test of the long-pulse capability and stability of the diagnostic pressure gauge (DPG) for the ITER device.

physics.ins-det↗

arXiv2023

Multi-view Kernel PCA for Time series Forecasting

In this paper, we propose a kernel principal component analysis model for multi-variate time series forecasting, where the training and prediction schemes are derived from the multi-view formulation of Restricted Kernel Machines. The training problem is simply an eigenvalue decomposition of the summation of two kernel matrices corresponding to the views of the input and output data. When a linear kernel is used for the output view, it is shown that the forecasting equation takes the form of kernel ridge regression. When that kernel is non-linear, a pre-image problem has to be solved to forecast a point in the input space. We evaluate the model on several standard time series datasets, perform ablation studies, benchmark with closely related models and discuss its results.

cs.LG↗

arXiv2021

Unsupervised Energy-based Out-of-distribution Detection using Stiefel-Restricted Kernel Machine

Detecting out-of-distribution (OOD) samples is an essential requirement for the deployment of machine learning systems in the real world. Until now, research on energy-based OOD detectors has focused on the softmax confidence score from a pre-trained neural network classifier with access to class labels. In contrast, we propose an unsupervised energy-based OOD detector leveraging the Stiefel-Restricted Kernel Machine (St-RKM). Training requires minimizing an objective function with an autoencoder loss term and the RKM energy where the interconnection matrix lies on the Stiefel manifold. Further, we outline multiple energy function definitions based on the RKM framework and discuss their utility. In the experiments on standard datasets, the proposed method improves over the existing energy-based OOD detectors and deep generative models. Through several ablation studies, we further illustrate the merit of each proposed energy function on the OOD detection performance.

cs.LG↗

arXiv2020

Generative Restricted Kernel Machines: A Framework for Multi-view Generation and Disentangled Feature Learning

This paper introduces a novel framework for generative models based on Restricted Kernel Machines (RKMs) with joint multi-view generation and uncorrelated feature learning, called Gen-RKM. To enable joint multi-view generation, this mechanism uses a shared representation of data from various views. Furthermore, the model has a primal and dual formulation to incorporate both kernel-based and (deep convolutional) neural network based models within the same setting. When using neural networks as explicit feature-maps, a novel training procedure is proposed, which jointly learns the features and shared subspace representation. The latent variables are given by the eigen-decomposition of the kernel matrix, where the mutual orthogonality of eigenvectors represents the learned uncorrelated features. Experiments demonstrate the potential of the framework through qualitative and quantitative evaluation of generated samples on various standard datasets.

cs.LG↗

arXiv2020

Robust Generative Restricted Kernel Machines using Weighted Conjugate Feature Duality

Interest in generative models has grown tremendously in the past decade. However, their training performance can be adversely affected by contamination, where outliers are encoded in the representation of the model. This results in the generation of noisy data. In this paper, we introduce weighted conjugate feature duality in the framework of Restricted Kernel Machines (RKMs). The RKM formulation allows for an easy integration of methods from classical robust statistics. This formulation is used to fine-tune the latent space of generative RKMs using a weighting function based on the Minimum Covariance Determinant, which is a highly robust estimator of multivariate location and scatter. Experiments show that the weighted RKM is capable of generating clean images when contamination is present in the training data. We further show that the robust method also preserves uncorrelated feature learning through qualitative and quantitative experiments on standard datasets.

stat.ML↗

arXiv2020

Disentangled Representation Learning and Generation with Manifold Optimization

Disentanglement is a useful property in representation learning which increases the interpretability of generative models such as Variational autoencoders (VAE), Generative Adversarial Models, and their many variants. Typically in such models, an increase in disentanglement performance is traded-off with generation quality. In the context of latent space models, this work presents a representation learning framework that explicitly promotes disentanglement by encouraging orthogonal directions of variations. The proposed objective is the sum of an autoencoder error term along with a Principal Component Analysis reconstruction error in the feature space. This has an interpretation of a Restricted Kernel Machine with the eigenvector matrix-valued on the Stiefel manifold. Our analysis shows that such a construction promotes disentanglement by matching the principal directions in the latent space with the directions of orthogonal variation in data space. In an alternating minimization scheme, we use Cayley ADAM algorithm - a stochastic optimization method on the Stiefel manifold along with the ADAM optimizer. Our theoretical discussion and various experiments show that the proposed model improves over many VAE variants in terms of both generation quality and disentangled representation learning.

cs.LG↗

arXiv2018

Characterization of hydrogen plasma in a permanent ring magnet based helicon plasma source for negative ion source research

HELicon Experiment for Negative ion source (HELEN-I) with single driver is developed with a focus on the production of negative hydrogen ions. In the Helicon wave heated plasmas, very high plasma densities ($\sim10^{19} m^{-3}$) can be attained with electron temperatures as low as $\sim$1 eV in the downstream region. These conditions favor the production of negative hydrogen ions. In HELEN-I device at IPR, helicon plasma is produced using Hydrogen gas in a diverging magnetic field, created by a permanent ring magnet. RF Power ($P_{RF}$) of 800-1000W at 13.56 MHz frequency is applied to a Nagoya-III antenna to excite m = 1 helicon mode in the plasma. The plasma is confined by a multi-cusp field configuration in the expansion chamber. The transition from inductively coupled mode to Helicon mode is observed near $P_{RF}$ 700W with plasma density $\sim 10^{18} m^{-3}$ and electron temperature $\sim$ 5 eV in the driver and $\sim$ 1eV in the expansion volume. Line integrated negative hydrogen ion density is measured in the expansion chamber by employing an Optical Emission Spectroscopy (OES) diagnostic technique using $H_α/H_β$ ratio and Laser photo-detachment based Cavity Ring Down spectroscopic (CRDS) diagnostic technique. The measured value of negative hydrogen ion density is in the order of $10^{16} m^{-3}$ at 6 mTorr pressure and does not vary significantly with power in the helicon mode, pressure and downstream axial magnetic field variation. The negative ion density measurements are compared with theoretically estimated values calculated using particle balance method considering different reaction rates responsible for negative hydrogen ion creation and destruction. It is to be noted that at present Caesium (Cs) is not injected in the plasma discharge to enhance $H^-$ ion density.

physics.plasm-ph↗

arXiv2015

Social Interaction in the Flickr Social Network

Online social networking sites such as Facebook, Twitter and Flickr are among the most popular sites on the Web, providing platforms for sharing information and interacting with a large number of people. The different ways for users to interact, such as liking, retweeting and favoriting user-generated content, are among the defining and extremely popular features of these sites. While empirical studies have been done to learn about the network growth processes in these sites, few studies have focused on social interaction behaviour and the effect of social interaction on network growth. In this paper, we analyze large-scale data collected from the Flickr social network to learn about individual favoriting behaviour and examine the occurrence of link formation after a favorite is created. We do this using a systematic formulation of Flickr as a two-layer temporal multiplex network: the first layer describes the follow relationship between users and the second layer describes the social interaction between users in the form of favorite markings to photos uploaded by them. Our investigation reveals that (a) favoriting is well-described by preferential attachment, (b) over 50% of favorites are reciprocated within 10 days if at all they are reciprocated, (c) different kinds of favorites differ in how fast they are reciprocated, and (d) after a favorite is created, multiplex triangles are closed by the creation of follow links by the favoriter's followers to the favorite receiver.

cs.SI↗