arXiv subjects

Sumanta Mukherjee

Publications and source records attributed to Sumanta Mukherjee.

23 records · Page 2Linked to original sources

arXiv2020

Explainable AI based Interventions for Pre-season Decision Making in Fashion Retail

Future of sustainable fashion lies in adoption of AI for a better understanding of consumer shopping behaviour and using this understanding to further optimize product design, development and sourcing to finally reduce the probability of overproducing inventory. Explainability and interpretability are highly effective in increasing the adoption of AI based tools in creative domains like fashion. In a fashion house, stakeholders like buyers, merchandisers and financial planners have a more quantitative approach towards decision making with primary goals of high sales and reduced dead inventory. Whereas, designers have a more intuitive approach based on observing market trends, social media and runways shows. Our goal is to build an explainable new product forecasting tool with capabilities of interventional analysis such that all the stakeholders (with competing goals) can participate in collaborative decision making process of new product design, development and launch.

cs.CY↗

arXiv2019

Semi-Supervised Method using Gaussian Random Fields for Boilerplate Removal in Web Browsers

Boilerplate removal refers to the problem of removing noisy content from a webpage such as ads and extracting relevant content that can be used by various services. This can be useful in several features in web browsers such as ad blocking, accessibility tools such as read out loud, translation, summarization etc. In order to create a training dataset to train a model for boilerplate detection and removal, labeling or tagging webpage data manually can be tedious and time consuming. Hence, a semi-supervised model, in which some of the webpage elements are labeled manually and labels for others are inferred based on some parameters, can be useful. In this paper we present a solution for extraction of relevant content from a webpage that relies on semi-supervised learning using Gaussian Random Fields. We first represent the webpage as a graph, with text elements as nodes and the edge weights representing similarity between nodes. After this, we label a few nodes in the graph using heuristics and label the remaining nodes by a weighted measure of similarity to the already labeled nodes. We describe the system architecture and a few preliminary results on a dataset of webpages.

stat.ML↗

arXiv2014

Microscopic Origin of Charged Impurity Scattering and Flicker Noise in MoS2 field-effect Transistors

Scattering of charge carriers and flicker noise in electrical transport are the central performance limiting factors in electronic devices, but their microscopic origin in molybdenum disulphide~(MoS$_2$)-based field effect transistors remains poorly understood. Here, we show that both carrier scattering and low-frequency $1/f$ noise in mechanically exfoliated ultra-thin MoS$_2$ layers are determined by the localized trap states located within the MoS$_2$ channel itself. The trap states not only act as Coulomb scattering centers that determine transport in both equilibrium ($eV< k_BT$) and non-equilibrium ($eV>k_BT$) regimes, where $V$ and $T$ are the source drain bias and temperature respectively, but also exchange carriers with the channel to produce the conductivity noise. The internal origin of the trap states was further confirmed by studying noise in MoS$_2$ films deposited on crystalline boron nitride substrates. Possible origin and nature of the trap states is also discussed.

cond-mat.mes-hall↗

arXiv2013

Interface characterization of Co2MnGe/Rh2CuSn Heusler multilayers

All-Heusler multilayer structures have been investigated by means of high kinetic x-ray photoelectron spectroscopy and x-ray magnetic circular dichroism, aiming to address the amount of disorder and interface diffusion induced by annealing of the multilayer structure. The studied multilayers consist of ferromagnetic Co$_2$MnGe and non-magnetic Rh$_2$CuSn layers with varying thicknesses. We find that diffusion begins already at comparably low temperatures between 200 $^{\circ}$C and 250 $^{\circ}$C, where Mn appears to be most prone to diffusion. We also find evidence for a 4 Å thick magnetically dead layer that, together with the identified interlayer diffusion, are likely reasons for the small magnetoresistance found for current-perpendicular-to-plane giant magneto-resistance devices based on this all-Heusler system.

cond-mat.mtrl-sci↗

arXiv2011

Re-entrant ordering of solute in a colloidal suspension during solvent evaporation

We study the phenomenon of self-assembly of silica micro-spheres on a glass plate during evaporation of the solvent from a colloidal suspension. Our experiments unveil an interesting competition between ordering and compaction in a strongly driven, out of equilibrium system arising from a slowing down of dynamics due to an impending glass transition. A suitable choice of experimental conditions minimizing the influence of many other competing phenomena that usually complicate probing of this underlying physics is crucial for our study. A re-entrant behavior in the order-disorder phase diagram as a function of particle density and drying time is established and the results are explained with the help of simulations and phenomenological theory.

cond-mat.soft↗