arXiv subjects

K. Krishnamurthy

Publications and source records attributed to K. Krishnamurthy.

6 recordsLinked to original sources

arXiv2024

Few-Shot Transfer Learning for Individualized Braking Intent Detection on Neuromorphic Hardware

Objective: This work explores use of a few-shot transfer learning method to train and implement a convolutional spiking neural network (CSNN) on a BrainChip Akida AKD1000 neuromorphic system-on-chip for developing individual-level, instead of traditionally used group-level, models using electroencephalographic data. Main Results: Efficacy of the above methodology to develop individual-specific braking intention predictive models by rapidly adapting the group-level model in as few as three training epochs while achieving at least 90% accuracy, true positive rate and true negative rate is presented. Further, results show the energy-efficiency of the neuromorphic hardware through a power reduction of over 97% with only a $1.3* increase in latency when using the Akida AKD1000 processor for network inference compared to an Intel Xeon central processing unit. Similar results were obtained in a subsequent ablation study using a subset of five out of 19 channels.

cs.NE↗

arXiv2022

Convolutional Spiking Neural Networks for Detecting Anticipatory Brain Potentials Using Electroencephalogram

Spiking neural networks (SNNs) are receiving increased attention because they mimic synaptic connections in biological systems and produce spike trains, which can be approximated by binary values for computational efficiency. Recently, the addition of convolutional layers to combine the feature extraction power of convolutional networks with the computational efficiency of SNNs has been introduced. This paper studies the feasibility of using a convolutional spiking neural network (CSNN) to detect anticipatory slow cortical potentials (SCPs) related to braking intention in human participants using an electroencephalogram (EEG). Data was collected during an experiment wherein participants operated a remote-controlled vehicle on a testbed designed to simulate an urban environment. Participants were alerted to an incoming braking event via an audio countdown to elicit anticipatory potentials that were measured using an EEG. The CSNN's performance was compared to a standard CNN, EEGNet and three graph neural networks via 10-fold cross-validation. The CSNN outperformed all the other neural networks, and had a predictive accuracy of 99.06 percent with a true positive rate of 98.50 percent, a true negative rate of 99.20 percent and an F1-score of 0.98. Performance of the CSNN was comparable to the CNN in an ablation study using a subset of EEG channels that localized SCPs. Classification performance of the CSNN degraded only slightly when the floating-point EEG data were converted into spike trains via delta modulation to mimic synaptic connections.

cs.NE↗

arXiv2021

Online-Learning Deep Neuro-Adaptive Dynamic Inversion Controller for Model Free Control

Adaptive methods are popular within the control literature due to the flexibility and forgiveness they offer in the area of modelling. Neural network adaptive control is favorable specifically for the powerful nature of the machine learning algorithm to approximate unknown functions and for the ability to relax certain constraints within traditional adaptive control. Deep neural networks are large framework networks with vastly superior approximation characteristics than their shallow counterparts. However, implementing a deep neural network can be difficult due to size specific complications such as vanishing/exploding gradients in training. In this paper, a neuro-adaptive controller is implemented featuring a deep neural network trained on a new weight update law that escapes the vanishing/exploding gradient problem by only incorporating the sign of the gradient. The type of controller designed is an adaptive dynamic inversion controller utilizing a modified state observer in a secondary estimation loop to train the network. The deep neural network learns the entire plant model on-line, creating a controller that is completely model free. The controller design is tested in simulation on a 2 link planar robot arm. The controller is able to learn the nonlinear plant quickly and displays good performance in the tracking control problem.

eess.SY↗

arXiv2014

Nonuniform Expansion of the Youngest Galactic Supernova Remnant G1.9+0.3

We report measurements of X-ray expansion of the youngest Galactic supernova remnant, G1.9+0.3, using Chandra observations in 2007, 2009, and 2011. The measured rates strongly deviate from uniform expansion, decreasing radially by about 60% along the X-ray bright SE-NW axis from 0.84% +/- 0.06% per yr to 0.52% +/- 0.03% per yr. This corresponds to undecelerated ages of 120-190 yr, confirming the young age of G1.9+0.3, and implying a significant deceleration of the blast wave. The synchrotron-dominated X-ray emission brightens at a rate of 1.9% +/- 0.4% per yr. We identify bright outer and inner rims with the blast wave and reverse shock, respectively. Sharp density gradients in either ejecta or ambient medium are required to produce the sudden deceleration of the reverse shock or the blast wave implied by the large spread in expansion ages. The blast wave could have been decelerated recently by an encounter with a modest density discontinuity in the ambient medium, such as found at a wind termination shock, requiring strong mass loss in the progenitor. Alternatively, the reverse shock might have encountered an order-of-magnitude density discontinuity within the ejecta, such as found in pulsating delayed-detonation Type Ia models. We demonstrate that the blast wave is much more decelerated than the reverse shock in these models for remnants at ages similar to G1.9+0.3. Similar effects may also be produced by dense shells possibly associated with high-velocity features in Type Ia spectra. Accounting for the asymmetry of G1.9+0.3 will require more realistic 3D Type Ia models.

astro-ph.HE↗

arXiv2013

Supernova Ejecta in the Youngest Galactic Supernova Remnant G1.9+0.3

G1.9+0.3 is the youngest known Galactic supernova remnant (SNR), with an estimated supernova (SN) explosion date of about 1900, and most likely located near the Galactic Center. Only the outermost ejecta layers with free-expansion velocities larger than about 18,000 km/s have been shocked so far in this dynamically young, likely Type Ia SNR. A long (980 ks) Chandra observation in 2011 allowed spatially-resolved spectroscopy of heavy-element ejecta. We denoised Chandra data with the spatio-spectral method of Krishnamurthy et al., and used a wavelet-based technique to spatially localize thermal emission produced by intermediate-mass elements (IMEs: Si and S) and iron. The spatial distribution of both IMEs and Fe is extremely asymmetric, with the strongest ejecta emission in the northern rim. Fe Kalpha emission is particularly prominent there, and fits with thermal models indicate strongly oversolar Fe abundances. In a localized, outlying region in the northern rim, IMEs are less abundant than Fe, indicating that undiluted Fe-group elements (including 56Ni) with velocities larger than 18,000 km/s were ejected by this SN. But in the inner west rim, we find Si- and S-rich ejecta without any traces of Fe, so high-velocity products of O-burning were also ejected. G1.9+0.3 appears similar to energetic Type Ia SNe such as SN 2010jn where iron-group elements at such high free-expansion velocities have been recently detected. The pronounced asymmetry in the ejecta distribution and abundance inhomogeneities are best explained by a strongly asymmetric SN explosion, similar to those produced in some recent 3D delayed-detonation Type Ia models.

astro-ph.HE↗

arXiv2011

Expansion of the Youngest Galactic Supernova Remnant G1.9+0.3

We present a measurement of the expansion and brightening of G1.9+0.3, the youngest Galactic supernova remnant, comparing Chandra X-ray images obtained in 2007 and 2009. A simple uniform expansion model describes the data well, giving an expansion rate of 0.642 +/- 0.049 % yr^-1, and a flux increase of 1.7 +/- 1.0 % yr^-1. Without deceleration, the remnant age would then be 156 +/- 11 yr, consistent with earlier results. Since deceleration must have occurred, this age is an upper limit; we estimate an age of about 110 yr, or an explosion date of about 1900. The flux increase is comparable to reported increases at radio wavelengths. G1.9+0.3 is the only Galactic supernova remnant increasing in flux, with implications for the physics of electron acceleration in shock waves

astro-ph.GA↗