Source author record

Vipul Bansal

Vipul Bansal appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Machine Learning cond-mat.mtrl-sci cs.CY

Catalog footprint

What is connected

3works

3topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

Activation Functions: Dive into an optimal activation function

Activation functions have come up as one of the essential components of neural networks. The choice of adequate activation function can impact the accuracy of these methods. In this study, we experiment for finding an optimal activation function by defining it as a weighted sum of existing activation functions and then further optimizing these weights while training the network. The study uses three activation functions, ReLU, tanh, and sin, over three popular image datasets, MNIST, FashionMNIST, and KMNIST. We observe that the ReLU activation function can easily overlook other activation functions. Also, we see that initial layers prefer to have ReLU or LeakyReLU type of activation functions, but deeper layers tend to prefer more convergent activation functions.

preprint2020arXiv

Computational Models for Academic Performance Estimation

Evaluation of students' performance for the completion of courses has been a major problem for both students and faculties during the work-from-home period in this COVID pandemic situation. To this end, this paper presents an in-depth analysis of deep learning and machine learning approaches for the formulation of an automated students' performance estimation system that works on partially available students' academic records. Our main contributions are (a) a large dataset with fifteen courses (shared publicly for academic research) (b) statistical analysis and ablations on the estimation problem for this dataset (c) predictive analysis through deep learning approaches and comparison with other arts and machine learning algorithms. Unlike previous approaches that rely on feature engineering or logical function deduction, our approach is fully data-driven and thus highly generic with better performance across different prediction tasks.

preprint2010arXiv

Absence of Morphotropic Phase Boundary Effects in BiFeO3-PbTiO3 Thin Films Grown via a Chemical Multilayer Deposition Method

Here, we report the unusual behaviour shown by the (BiFeO3)1-x-(PbTiO3)x (BF-xPT) films prepared using a multilayer deposition approach by chemical solution deposition method. Thin film samples of various compositions were prepared by depositing several bilayers of BF and PT precursors by varying the BF or PT layer thicknesses. X-ray diffraction showed that final samples of all compositions show mixing of the two compounds resulting in a single phase mixture, also confirmed by transmission electron microscopy. In contrast to bulk equilibrium compositions, our samples show a monoclinic (MA type) structure suggesting disappearance of morphotropic phase boundary (MPB) about x = 0.30 as observed in the bulk. This is accompanied by the lack of any enhancement of remnant polarization at MPB as shown by the ferroelectric measurements. Magnetic measurements show that the magnetization of the samples increases with increasing BF content. Significant magnetization of the samples indicates melting of spin spirals in the BF-xPT arising from random distribution of iron atoms across the film. Absence of Fe2+ ions in the films was corroborated by X-ray photoelectron spectroscopy measurements. The results illustrate that used thin film processing methodology significantly changes the structural evolution in contrast to predictions from the equilibrium phase diagram as well as modify the functional characteristics of BP-xPT system dramatically.