Source author record

Amit Choudhary

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

ResearcherUnclaimed source record

Catalog footprint

What is connected

2works
3topics
3close collaborators

Actions

Connect this record

Log in to claim

Research graph

See the researcher in context

Open full explorer

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.

Published work

2 published item(s)

preprint2015arXiv

Theoretical analysis of nanoparticle-induced homeotropic alignment in nematic liquid crystals

A theoretical analysis of homeotropic alignment induced by nanoparticles (NPs) in a nematic liquid crystal (NLC) sample cell is presented. It is found that such alignment on the surface of a NP causes a change in the orientation of the molecular director near the surface, which in turn induces variations in the elastic constants and free energy. The induced NLC properties allow coupling between nearby NPs, mediated by the NLC molecules. The rotation of the coupled NPs close to the substrate tends to induce a long-range orientation of the NLC molecular director, leading to modification in the alignment at the interface of NLC and substrate which induces the orientation from homogeneous (planar) to homeotropic (vertical) in the bulk material.

preprint2011arXiv

Improving the character recognition efficiency of feed forward BP neural network

This work is focused on improving the character recognition capability of feed-forward back-propagation neural network by using one, two and three hidden layers and the modified additional momentum term. 182 English letters were collected for this work and the equivalent binary matrix form of these characters was applied to the neural network as training patterns. While the network was getting trained, the connection weights were modified at each epoch of learning. For each training sample, the error surface was examined for minima by computing the gradient descent. We started the experiment by using one hidden layer and the number of hidden layers was increased up to three and it has been observed that accuracy of the network was increased with low mean square error but at the cost of training time. The recognition accuracy was improved further when modified additional momentum term was used.