AKAbhishek Kadam

Abhishek Kadam

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Abhishek Kadam has a total of 27 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Analog and Temporary On-chip Memory for ANN Training and Inference

    by Shreyas Deshmukh, Raghav Singhal, Shruti Landge, Vivek Saraswat, Anmol Biswas, Abhishek Kadam, Ajay K. Singh, Sreenivas Subramoney, Laxmeesha Somappa, Maryam Shojaei Baghini, et al.

    On-chip training at the edge becomes a primary requisite for real-time and security-sensitive artificial neural network (ANN) applications. In-memory computation (IMC) techniques have been proposed to facilitate data-intensive computational operations in ANNs. IMC-based multiply-accumulate (MAC) accelerates ANN training but suffers from significant communication overhead between the MAC engine and the off-chip storage for the intermediate data. This article proposes an analog temporary on-chip memory (ATOM) to store this intermediate data during ANN training. The ANN training architecture with the proposed ATOM has two significant advantages. First, the energy required to store intermediate data is scaled down by \(\sim\) 40 \(\times\) due to the on-chip and analog nature of the memory. Second, the proposed architecture avoids power and area-consuming analog-to-digital converters (ADCs) between neural network stages. The ATOM cell measurements are carried out from 20 fabricated chips, and the impact of ATOM characteristics on ANN system performance accuracy is analyzed. This article shows significant latency improvement of \(\sim\) 9 \(\times\) and area savings of \(\sim\) 5 \(\times\) for intermediate data storage compared to the on-chip SRAM during ANN training’s forward and backward pass operations. An improvement in the area and latency will be beneficial to instrument the area- and energy-efficient hardware system for on-chip ANN applications.

  • Design Space and Variability Analysis of SOI MOSFET for Ultralow-Power Band-to-Band Tunneling Neurons

    by Jay Sonawane, Shubham Patil, Abhishek Kadam, Ajay Kumar Singh, Sandip Lashkare, Veeresh Deshpande, Udayan Ganguly

    Large spiking neural networks (SNNs) require ultralow power and low variability hardware for neuromorphic computing applications. Recently, a band-to-band tunneling (BTBT)-based integrator was proposed, enabling the sub-kHz operation of neurons with area and energy efficiency. For an ultralow-power implementation of such neurons, a very low BTBT current is needed, so minimizing current without degrading neuronal properties is essential. Low variability is needed in the ultralow current integrator to avoid network performance degradation in a large BTBT neuron-based SNN. This work addresses device optimization to achieve low BTBT current. We conducted design space and variability analysis in technology computer-aided design (TCAD), utilizing a well-calibrated TCAD deck with experimental data from GlobalFoundries (GFs) 32 nm partially depleted silicon-on-insulator (PD-SOI) MOSFET. First, we discuss the physics-based explanation of the tunneling mechanism. Second, we explore the impact of device design parameters on SOI MOSFET performance, highlighting parameter sensitivities to tunneling current. With device parameters’ optimization, we demonstrate a $\sim$ 20 $\times$ reduction in BTBT current compared to the experimental data. Finally, a variability analysis that includes the effects of random dopant fluctuations (RDFs), oxide thickness variation (OTV), and channel–oxide interface traps ( D $_{\text{IT}}$ ) in the BTBT, subthreshold (SS), and ON regimes of operation is shown. The BTBT regime shows the highest sensitivity to OTV, with variability increasing by up to 25 $\times$ compared to the SS regime. In contrast, RDF and D $_{\text{IT}}$ variability resulted in a 1.25 $\times$ to $\sim$ 10 $\times$ lower coefficient of variation (CV) in the BTBT regime than in the SS regime, indicating better resilience to these sources of variability. The study provides essential design guidelines to enable energy-efficient neuromorphic computing, achieving biologically plausible sub-kHz spiking frequencies.


  • Open Access

    One-Time Programmable Memory for Ultra-Low Power ANN Inference Accelerator With Security Against Thermal Fault Injection

    by Shreyas Deshmukh, Ankit Bende, Diti Sanghai, Vivek Saraswat, Anmol Biswas, Abhishek Kadam, Shubham Patil, Ajay Kumar Singh, Veeresh Deshpande, Udayan Ganguly

    Memory technologies for in-memory computing (IMC) based neural network (NN) operations in edge devices face two primary challenges. First, these technologies often exhibit an excessive average power due to the high currents in the memory’s low and high resistance state (LRS and HRS). Second, their vulnerability to tampering compromises device reliability in security-sensitive Internet of Things (IoT) applications. To tackle these challenges, we propose the use of one-time programmable memory (OTPM) for IMC-based thermal attack-secured and ultra-low powered NN accelerator with write disablement to prevent re-programming/tampering. In this article, first, we show that HRS for OTP memory is exceptionally high ( 190 MΩ), but the higher current in LRS ( 100 mA) presents a challenge of high average power. We address this challenge by introducing a select transistor gate voltage biasing scheme with the 1T-1OTPM bit cell to reduce LRS current significantly while still trading off the excessive HRS/LRS ratio of 1.5 M to about 1 k. This reduces average power by 1000× without degrading CIFAR-10 classification performance. Second, we evaluate the impact of OTPM size scaling on HRS and LRS distributions. Though scaling reduces HRS current, we show that moderately scaled 10F×10F OTPM bit cells produce the lowest maximum current in HRS compared to the more aggressively scaled 1F×1F, which suffers from significant process-induced variability. Third, we show the robustness to temperature, which enables security from thermal attacks.


  • Area and Energy-Efficient Quantum Tunneling-Based Thermal Sensor on 45nm RFSOI Technology

    by Shubham Patil, Abhishek Kadam, Jay Sonawane, Shreyas Deshmukh, R. Gaurav, Ajay Kumar Singh, Sandip Lashkare, Veeresh Deshpande, Laxmeesha Somappa, Udayan Ganguly

    The compact and energy-efficient integrated temperature sensors are crucial for temperature monitoring and control. In this work, we propose a novel area and energy-efficient band-to-band tunneling (BTBT)-based oscillator for temperature sensing applications. It utilizes oscillation frequency as a metric for temperature sensing. The linearity in BTBT current with temperature can enable sensing with a simple readout mechanism. The initially designed circuit is simulated and analyzed in TCAD using mixed-mode simulation, followed by the fabrication of the proposed circuit using GF 45nm RFSOI technology. Measurements show the BTBT oscillator’s quasi-linear response as a function of temperature. Our proposed work enables the area ( $0.32~\mu \text { m}^{{2}}$ ) and energy-efficient temperature sensor (2.5 fJ/cycle) for the energy and area-constraint edge applications.

  • Reliability of Tunneling Regime for Silicon on Insulator-Based Neuron

    by Shubham Patil, Abhishek Kadam, Rashmi Saikia, Jay Sonawane, Karansingh Thakor, Ajay Kumar Singh, R. Gaurav, Nihar Ranjan Mohapatra, Sandip Lashkare, Veeresh Deshpande, et al.

    Low-power operations are essential for implementing large spiking neural networks (SNNs) in real-world applications. An area and energy-efficient demonstration of a functional liquid state machine (LSM) for spoken word detection using the band-to-band-tunneling (BTBT)-based neuron was proposed earlier. For a product-level demonstration of the BTBT regime operating chips, the variability and reliability of neurons emerge as noteworthy concerns for neuromorphic processors due to the potential implications for performance degradation over time. In this work, we characterize and compare the reliability of partially depleted (PD) silicon on insulator (SOI) transistors in the BTBT and on-regime ( ${I}_{\text {ON}}$ ) regimes. The drain current fractional degradation ( ${\Delta }{I}_{\text {D}}/{I}_{\text {D}}$ ) increases with an increase in voltage and thermal stress in the BTBT and ${I}_{\text {ON}}$ regime. The reaction-diffusion-drift (RDD) framework is used to estimate device degradation under operating bias conditions. At operating bias, a ~17% fractional degradation is observed in the BTBT regime operation for ten years, comparable to the degradation observed in the ${I}_{\text {ON}}$ regime. Finally, we analyzed the impact of device degradation on the SNN performance.


  • A 42.3 μm² Band to Band Tunneling-Based Oscillator Enabled Temperature to Digital Converter With Resolution FoM of 0.16 pJK² for Embedded Temperature Sensing

    by Abhishek A. Kadam, Shubham Patil, Ajay K. Singh, Maryam Shojaei Baghini, Udayan Ganguly, Laxmeesha Somappa

    In advanced high-speed integrated systems, the widely distributed and proliferation of temperature sensors to detect hotspots improve the robustness and reliability of the system by preventing overheating. Low area and low energy consumption are essential for integrated temperature sensors in such applications. The band-to-band tunneling (BTBT) current in PDSOI technology increases linearly with temperature. We demonstrate a BTBT-based temperature-dependent oscillator. The fabricated oscillator has a low footprint (42.3 μm2, 10× less than state-of-the-art temperature sensing cores) and enables low energy consuming temperature to digital converter (0.32 nJ energy/conversion) in GF45RFSOI technology. The proposed oscillator facilitates an area and energy-efficient temperature sensor (20 °C to 90 °C) with a simple counter-based digital readout with a best-in-class resolution FoM of 0.16 pJK2.


  • Open Access

    A Compact Low Power Multi-mode Spiking Neuron using Band to Band Tunneling

    by Abhishek A. Kadam, Ajay K Singh, Laxmeesha Somappa, Maryam Shojaei Baghini, Udayan Ganguly

    Efficient and compact neurons with low power consumption are crucial when designing large-scale spiking neural networks (SNNs) for hardware implementation. Many architectures in the literature showcase different spike patterns associated with biological neurons. However, using bulky capacitors to generate the different time constants related to complex neuron patterns makes these circuits area inefficient. This paper presents a band-to-band-tunneling (BTBT) based energy-efficient and compact neuron capable of producing various spike patterns. The BTBT region's extremely low current enables different time constants while eliminating the need of bulky capacitors. The circuit is based on the Izhikevich neuron model. The proposed circuit is designed in Silicon on Insulator technology to exhibit important firing patterns observed in the biological cortex, viz. regular spiking, fast-spiking, and chattering, and it is fine-tuned for efficient operation at low subthreshold voltages. This circuit utilizes only 129 μm 2 area and consumes only 6.7 fJ energy per spike (approximately 40% lower area and energy per spike than state-of-the-art multi-mode neurons) in G45RFSOI technology.

  • Open Access

    A 42.3 μ m 2 Band to Band Tunneling-based Thermillator Enabled Temperature to Digital Converter with Resolution FoM of 0.16 p J K 2 for Embedded Temperature Sensing

    by Abhishek A. Kadam

    In advanced high-speed integrated systems, the widely distributed and proliferation of temperature sensors to detect hotspots improve the robustness and reliability of the system by preventing overheating. Low area and low energy consumption are essential for integrated temperature sensors in such applications. The band-to-band tunneling (BTBT) current in PDSOI technology increases linearly with temperature. We demonstrate a BTBT-based quasi-linear temperature-dependent oscillator, thermillator. The fabricated thermillator has a low footprint (42.3 μ m 2 , 10× less than state-of-the-art temperature sensing cores) and enables low energy consuming temperature to digital converter (0.33 nJ energy/conversion) in GF45RFSOI technology. The proposed thermillator facilitates a low power and low footprint temperature sensor (0 o C to 85 o C ) with a simple counter-based digital readout with a best-in-class resolution FoM of 0.16 p J K 2 .

  • Process-Voltage-Temperature Variability Estimation of Tunneling Current for Band-to-Band-Tunneling-Based Neuron

    by Shubham Patil, Anand Sharma, R. Gaurav, Abhishek Kadam, Ajay Kumar Singh, Sandip Lashkare, Nihar Ranjan Mohapatra, Udayan Ganguly

    Compact and energy-efficient synapse and neurons are essential to realize the full potential of neuromorphic computing. In addition, a low variability is indeed needed for neurons in deep neural networks for higher accuracy. Further, process ( P), voltage ( V), and temperature ( T) (PVT) variation are essential considerations for low-power circuits as performance impact and compensation complexities are added costs. Recently, band-to-band tunneling (BTBT) neuron has been demonstrated to operate successfully in a network to enable a liquid state machine (LSM). A comparison of the PVT with competing modes of operation (e.g., BTBT versus subthreshold and above threshold) of the same transistor is a critical factor in assessing performance. In this work, we demonstrate the PVT variation impact on the BTBT regime and benchmark the operation against the subthreshold regime (SS) and ON-regime (I $_{\biosc{on}}$ ) of partially depleted silicon-on-insulator MOSFET. It is shown that the ON-state regime offers the lowest variability but dissipates higher power, hence not usable for low-power sources. Among the BTBT and SS regimes, which can enable the low-power neuron, the BTBT regime has shown $\sim$ 3 ${\times}$ variability reduction ( ${\bm{\sigma}_{{\textit{I}}{_{{\textit{D}}}}}{/}\bm{\mu }_{{\textit{I}}{_{{\textit{D}}}}}}$ ) compared to the SS regime, considering the cumulative PVT variability. The improvement is due to the well-known weaker P, V, and T dependence of BTBT versus SS. We show that the BTBT variation is uncorrelated with mutually correlated SS and I $_{{\biosc{on}}}$ operation—indicating its different origin from the mechanism and location perspectives. Hence, the BTBT regime is promising for low-current, low-power, and low device-to-device (D2D) variability neuron operation.


  • A Compact and Ultra-Low-Power Low-Pass Filter Based on Band-to-Band Tunneling Effect

    by Abhishek A. Kadam, Ajay K. Singh, Laxmeesha Somappa, Maryam Shojaei Baghini, Udayan Ganguly

    Continuous-time filters with low-frequency cutoff are crucial building blocks in analog signal processing circuits for speech processing and biomedical applications. The design of an integrated continuous-time filter with a cutoff frequency spanning from sub-Hz to a few kHz is constrained by the ultra-low power and area minimization requirement. In this article, we have experimentally demonstrated a low pass filter response using only a single partially depleted (PD) silicon on insulator (SOI) transistor. The proposed novel filter is based on band to band tunnelling (BTBT). This low pass filter is on-chip tunable to provide a wide cutoff frequency ranging from 2 Hz to 20 kHz, with a silicon footprint of 9 μm2 and the maximum power consumption of 0.6 nW in Global Foundries’ (GF) 45 nm RFSOI technology. The proposed tunneling-based filter is less prone to process variations and mismatches as compared to traditional integrated analog filters. The proposed BTBT-based passive RC filter with spurious free dynamic range (SFDR) of 29 dbc, outperforms previously reported filters in the literature in terms of power consumption and area. The ultra-low power consumption and area efficiency make this proposed low pass filter suitable for the multichannel analog front-end low-frequency filters for noise-tolerant neural network-based applications such as speech recognition and electrophysiological signal recognition.