Rashmi Jha - Academia.edu (original) (raw)

Papers by Rashmi Jha

2019 International Conference on Communication and Electronics Systems (ICCES)

2021 IEEE International Midwest Symposium on Circuits and Systems (MWSCAS)

2021 IEEE International Midwest Symposium on Circuits and Systems (MWSCAS)

IEEE Access

Malware is an ever-present problem in the modern era and while detecting malware with AI has grow... more Malware is an ever-present problem in the modern era and while detecting malware with AI has grown as a new field of exploration, current methods are not yet mature enough for widespread adoption in terms of speed and performance. Current methods largely focus on viewing malicious assembly as an image for detection, requiring a large amount of preprocessing and making network architectures inflexible. Preprocessing malware images to one size introduces additional time to predict and makes the task of prediction more difficult. We explore a novel method for transforming executable bytecode into a video rather than an image for classification with deep, time-distributed neural networks, achieving up to 98.74% testing accuracy on 9 classes of malware, and up to 99.36% testing accuracy on a balanced set of malicious vs. benign files. The network could also classify all malware in our dataset for a false positive rate of 13%, and was also found successful in classifying only parts of an input, as well as initial success in a 0-day scenario. The network only uses the executable code and no additional information to make predictions. We then explore methods for pruning and quantizing the network so that it may be more feasible for widespread implementation, including a novel pruning method we call Node-Distance pruning. Our model is found to be competitive to current works while remaining fast, lean and flexible.

IEEE Transactions on Very Large Scale Integration (VLSI) Systems

IEEE Transactions on Electron Devices

IEEE Access

This paper reports an approach for monitoring aging and integrity of CMOS circuits through additi... more This paper reports an approach for monitoring aging and integrity of CMOS circuits through additively manufactured Resistive Random-Access Memory (ReRAM) based test structures. MgO-based ReRAM devices demonstrated excellent temperature sensing and aging modalities with simultaneous storage of sensed temperature and age as a change in the resistive state. The Process Voltage Temperature (PVT) characteristics, aging, and temperature sensitivity of MgO-ReRAM devices were experimentally studied and modeled to capture resistance distributions and temperature-based modalities. This inmemory sensing feature of ReRAM was integrated with specially designed read circuitry using 180 nm CMOS technology, to produce a measurable change in spiking-frequency over the lifetime of the ReRAM under normal aging conditions with the underlying CMOS circuits. Large feature sizes were used so these circuits can be fabricated in-house in trusted foundry. Temporal changes in temperature of underlying CMOS circuit could be captured by instantaneous change in resistive state of ReRAM with local temperature fluctuations which translated to a change in read circuit output. The characteristics of this circuit is studied in detail using simulations. Due to additive integration of ReRAM and associated circuitry, this approach for aging and integrity monitoring (AIM) ensures large spatial and accurate temporal monitoring of underlying CMOS die with minimal loss of the functional chip area for these added security features. The passive, in-memory sensing, and non-volatile nature of ReRAM also ensures low-power consumption in these circuits. The devices resistance states and material composition are specific to every device preventing reverse engineering and tampering of the devices, thus making it an attractive approach for adding customized security and trust features in advanced CMOS nodes-based circuits.

Advances in Science, Technology and Engineering Systems Journal

We present findings on classifying the class of executable code using convolutional, recurrent ne... more We present findings on classifying the class of executable code using convolutional, recurrent neural networks by creating images from only the .text section of executables and dividing them into standard-size windows, using minimal preprocessing. We achieve up to 98.24% testing accuracy on classifying 9 types of malware, and 99.50% testing accuracy on classifying malicious vs. benign code. Then, we find that a recurrent network may not entirely be necessary, opening the door for future neural network architectures.

MRS Advances

ABSTRACTIn this work, a printable tungsten disulfide (WS2) based ink is developed from readily av... more ABSTRACTIn this work, a printable tungsten disulfide (WS2) based ink is developed from readily available WS2 powder (0.6 µm average particle size), and an ink-jet printing based deposition method for a tungsten disulfide film is presented. WS2 flake coverage and bulk electrical characteristics under three different irradiance conditions are examined and discussed. Presence of excitons in the absorbance of the inks is performed by optical UV-Vis spectrometry. Metrics using the A exciton peak generated by the few-layered flakes are used to calculate the average flake lateral dimensions, the concentration of WS2 in the inks after size selection and filtering, as well as the average monolayer count of the flakes. After printing, scanning electron microscopy is used to confirm average flake lateral size and average flake area coverage, while an atomic force microscope is used to confirm flake thickness.

IEEE Electron Device Letters

Bioinspiration & biomimetics, Nov 11, 2017

Biomimetic robots have gained attention recently for various applications ranging from resource h... more Biomimetic robots have gained attention recently for various applications ranging from resource hunting to search and rescue operations during disasters. Biological species are known to intuitively learn from the environment, gather and process data, and make appropriate decisions. Such sophisticated computing capabilities in robots are difficult to achieve, especially if done in real-time with ultra- low energy consumption. Here, we present a novel memristive device based learning architecture for robots. Two terminal memristive devices with resistive switching of oxide layer are modeled in a crossbar array to develop a neuromorphic platform that can impart active real-time learning capabilities in a robot. This approach is validated by navigating a robot vehicle in an unknown environment with randomly placed obstacles. Further, the proposed scheme is compared with Reinforcement Learning based algorithms using local and global knowledge of the environment. The simulation as well as...

IEEE Journal on Emerging and Selected Topics in Circuits and Systems, 2016

2016 IEEE National Aerospace and Electronics Conference (NAECON) and Ohio Innovation Summit (OIS), 2016

2019 International Conference on Communication and Electronics Systems (ICCES)

2021 IEEE International Midwest Symposium on Circuits and Systems (MWSCAS)

2021 IEEE International Midwest Symposium on Circuits and Systems (MWSCAS)

IEEE Access

Malware is an ever-present problem in the modern era and while detecting malware with AI has grow... more Malware is an ever-present problem in the modern era and while detecting malware with AI has grown as a new field of exploration, current methods are not yet mature enough for widespread adoption in terms of speed and performance. Current methods largely focus on viewing malicious assembly as an image for detection, requiring a large amount of preprocessing and making network architectures inflexible. Preprocessing malware images to one size introduces additional time to predict and makes the task of prediction more difficult. We explore a novel method for transforming executable bytecode into a video rather than an image for classification with deep, time-distributed neural networks, achieving up to 98.74% testing accuracy on 9 classes of malware, and up to 99.36% testing accuracy on a balanced set of malicious vs. benign files. The network could also classify all malware in our dataset for a false positive rate of 13%, and was also found successful in classifying only parts of an input, as well as initial success in a 0-day scenario. The network only uses the executable code and no additional information to make predictions. We then explore methods for pruning and quantizing the network so that it may be more feasible for widespread implementation, including a novel pruning method we call Node-Distance pruning. Our model is found to be competitive to current works while remaining fast, lean and flexible.

IEEE Transactions on Very Large Scale Integration (VLSI) Systems

IEEE Transactions on Electron Devices

IEEE Access

This paper reports an approach for monitoring aging and integrity of CMOS circuits through additi... more This paper reports an approach for monitoring aging and integrity of CMOS circuits through additively manufactured Resistive Random-Access Memory (ReRAM) based test structures. MgO-based ReRAM devices demonstrated excellent temperature sensing and aging modalities with simultaneous storage of sensed temperature and age as a change in the resistive state. The Process Voltage Temperature (PVT) characteristics, aging, and temperature sensitivity of MgO-ReRAM devices were experimentally studied and modeled to capture resistance distributions and temperature-based modalities. This inmemory sensing feature of ReRAM was integrated with specially designed read circuitry using 180 nm CMOS technology, to produce a measurable change in spiking-frequency over the lifetime of the ReRAM under normal aging conditions with the underlying CMOS circuits. Large feature sizes were used so these circuits can be fabricated in-house in trusted foundry. Temporal changes in temperature of underlying CMOS circuit could be captured by instantaneous change in resistive state of ReRAM with local temperature fluctuations which translated to a change in read circuit output. The characteristics of this circuit is studied in detail using simulations. Due to additive integration of ReRAM and associated circuitry, this approach for aging and integrity monitoring (AIM) ensures large spatial and accurate temporal monitoring of underlying CMOS die with minimal loss of the functional chip area for these added security features. The passive, in-memory sensing, and non-volatile nature of ReRAM also ensures low-power consumption in these circuits. The devices resistance states and material composition are specific to every device preventing reverse engineering and tampering of the devices, thus making it an attractive approach for adding customized security and trust features in advanced CMOS nodes-based circuits.

Advances in Science, Technology and Engineering Systems Journal

We present findings on classifying the class of executable code using convolutional, recurrent ne... more We present findings on classifying the class of executable code using convolutional, recurrent neural networks by creating images from only the .text section of executables and dividing them into standard-size windows, using minimal preprocessing. We achieve up to 98.24% testing accuracy on classifying 9 types of malware, and 99.50% testing accuracy on classifying malicious vs. benign code. Then, we find that a recurrent network may not entirely be necessary, opening the door for future neural network architectures.

MRS Advances

ABSTRACTIn this work, a printable tungsten disulfide (WS2) based ink is developed from readily av... more ABSTRACTIn this work, a printable tungsten disulfide (WS2) based ink is developed from readily available WS2 powder (0.6 µm average particle size), and an ink-jet printing based deposition method for a tungsten disulfide film is presented. WS2 flake coverage and bulk electrical characteristics under three different irradiance conditions are examined and discussed. Presence of excitons in the absorbance of the inks is performed by optical UV-Vis spectrometry. Metrics using the A exciton peak generated by the few-layered flakes are used to calculate the average flake lateral dimensions, the concentration of WS2 in the inks after size selection and filtering, as well as the average monolayer count of the flakes. After printing, scanning electron microscopy is used to confirm average flake lateral size and average flake area coverage, while an atomic force microscope is used to confirm flake thickness.

IEEE Electron Device Letters

Bioinspiration & biomimetics, Nov 11, 2017

Biomimetic robots have gained attention recently for various applications ranging from resource h... more Biomimetic robots have gained attention recently for various applications ranging from resource hunting to search and rescue operations during disasters. Biological species are known to intuitively learn from the environment, gather and process data, and make appropriate decisions. Such sophisticated computing capabilities in robots are difficult to achieve, especially if done in real-time with ultra- low energy consumption. Here, we present a novel memristive device based learning architecture for robots. Two terminal memristive devices with resistive switching of oxide layer are modeled in a crossbar array to develop a neuromorphic platform that can impart active real-time learning capabilities in a robot. This approach is validated by navigating a robot vehicle in an unknown environment with randomly placed obstacles. Further, the proposed scheme is compared with Reinforcement Learning based algorithms using local and global knowledge of the environment. The simulation as well as...

IEEE Journal on Emerging and Selected Topics in Circuits and Systems, 2016

2016 IEEE National Aerospace and Electronics Conference (NAECON) and Ohio Innovation Summit (OIS), 2016