Searcharxiv⌕ Search

arXiv subjects

Raj Patel

Publications and source records attributed to Raj Patel.

20 records · Page 2Linked to original sources

Estimator Vectors: OOV Word Embeddings based on Subword and Context Clue Estimates

Semantic representations of words have been successfully extracted from unlabeled corpuses using neural network models like word2vec. These representations are generally high quality and are computationally inexpensive to train, making them popular. However, these approaches generally fail to approximate out of vocabulary (OOV) words, a task humans can do quite easily, using word roots and context clues. This paper proposes a neural network model that learns high quality word representations, subword representations, and context clue representations jointly. Learning all three types of representations together enhances the learning of each, leading to enriched word vectors, along with strong estimates for OOV words, via the combination of the corresponding context clue and subword embeddings. Our model, called Estimator Vectors (EV), learns strong word embeddings and is competitive with state of the art methods for OOV estimation.

cs.CL↗

Nanoscale electromechanics to measure thermal conductivity, expansion and interfacial losses

We study the effect of localized Joule heating on the mechanical properties of doubly clamped nanowires under tensile stress. Local heating results in systematic variation of the resonant frequency; these frequency changes result from thermal stresses that depend on temperature dependent thermal conductivity and expansion coefficient. The change in sign of the linear expansion coefficient of InAs is reflected in the resonant response of the system near a bath temperature of 20 K. Using finite element simulations to model the experimentally observed frequency shifts, we show that the thermal conductivity of a nanowire can be approximated in the 10-60 K temperature range by the empirical form $κ=b$T W/mK, where the value of $b$ for a nanowire was found to be $b=0.035$ W/mK$^2$, significantly lower than bulk values. Also, local heating allows us to independently vary the temperature of the nanowire relative to the clamping points pinned to the bath temperature. We suggest a loss mechanism (dissipation $\sim10^{-4}-10^{-5}$) originating from the interfacial clamping losses between the metal and the semiconductor nanostructure.

cond-mat.mes-hall↗