Matching Items (56)
Description
In recent years, Artificial Intelligence (AI) (e.g., Deep Neural Networks (DNNs), Transformer) has shown great success in real-world applications due to its superior performance in various cognitive tasks. The impressive performance achieved by AI models normally accompanies the cost of enormous model size and high computational complexity, which significantly hampers their implementation on resource-limited Cyber-Physical Systems (CPS), Internet-of-Things (IoT), or Edge systems due to their tightly constrained energy, computing, size, and memory budget. Thus, the urgent demand for enhancing the \textbf{Efficiency} of DNN has drawn significant research interests across various communities.
Motivated by the aforementioned concerns, this doctoral research has been mainly focusing on Enabling Deep Learning at Edge: From Efficient and Dynamic Inference to On-Device Learning. Specifically, from the inference perspective, this dissertation begins by investigating a hardware-friendly model compression method that effectively reduces the size of AI model while simultaneously achieving improved speed on edge devices. Additionally, due to the fact that diverse resource constraints of different edge devices, this dissertation further explores dynamic inference, which allows for real-time tuning of inference model size, computation, and latency to accommodate the limitations of each edge device. Regarding efficient on-device learning, this dissertation starts by analyzing memory usage during transfer learning training. Based on this analysis, a novel framework called "Reprogramming Network'' (Rep-Net) is introduced that offers a fresh perspective on the on-device transfer learning problem. The Rep-Net enables on-device transferlearning by directly learning to reprogram the intermediate features of a pre-trained model. Lastly, this dissertation studies an efficient continual learning algorithm that facilitates learning multiple tasks without the risk of forgetting previously acquired knowledge. In practice, through the exploration of task correlation, an interesting phenomenon is observed that the intermediate features are highly correlated between tasks with the self-supervised pre-trained model. Building upon this observation, a novel approach called progressive task-correlated layer freezing is proposed to gradually freeze a subset of layers with the highest correlation ratios for each task leading to training efficiency.
ContributorsYang, Li (Author) / Fan, Deliang (Thesis advisor) / Seo, Jae-Sun (Committee member) / Zhang, Junshan (Committee member) / Cao, Yu (Committee member) / Arizona State University (Publisher)
Created2023
Description
Resistive random-access memory (RRAM) or memristor, is an emerging technology used in neuromorphic computing to exceed the traditional von Neumann obstacle by merging the processing and memory units. Two-dimensional (2D) materials with non-volatile switching behavior can be used as the switching layer of RRAMs, exhibiting superior behavior compared to conventional oxide-based RRAMs. The use of 2D materials allows scaling the resistive switching layer thickness to sub-nanometer dimensions enabling devices to operate with low switching voltages and high programming speeds, offering large improvements in efficiency and performance as well as ultra-dense integration. This dissertation presents an extensive study of linear and logistic regression algorithms implemented with 1-transistor-1-resistor (1T1R) memristor crossbars arrays. For this task, a simulation platform is used that wraps circuit-level simulations of 1T1R crossbars and physics-based model of RRAM to elucidate the impact of device variability on algorithm accuracy, convergence rate, and precision. Moreover, a smart pulsing strategy is proposed for the practical implementation of synaptic weight updates that can accelerate training in real crossbar architectures.
Next, this dissertation reports on the hardware implementation of analog dot-product operation on arrays of 2D hexagonal boron nitride (h-BN) memristors. This extends beyond previous work that studied isolated device characteristics towards the application of analog neural network accelerators based on 2D memristor arrays. The wafer-level fabrication of the memristor arrays is enabled by large-area transfer of CVD-grown few-layer h-BN films. The dot-product operation shows excellent linearity and repeatability, with low read energy consumption, with minimal error and deviation over various measurement cycles. Moreover, the successful implementation of a stochastic linear and logistic regression algorithm in 2D h-BN memristor hardware is presented for the classification of noisy images. Additionally, the electrical performance of novel 2D h-BN memristor for SNN applications is extensively investigated. Then, using the experimental behavior of the h-BN memristor as the artificial synapse, an unsupervised spiking neural network (SNN) is simulated for the image classification task. A novel and simple Spike-Timing-Dependent-Plasticity (STDP)-based dropout technique is presented to enhance the recognition task of the h-BN memristor-based SNN.
ContributorsAfshari, Sahra (Author) / Sanchez Esqueda, Ivan (Thesis advisor) / Barnaby, Hugh J (Committee member) / Seo, Jae-Sun (Committee member) / Cao, Yu (Committee member) / Arizona State University (Publisher)
Created2023
Description
With the push for integration, a slew of modern switching power management circuits are operating at higher switching frequencies in order to reduce passive filter sizes. But while these switching regulators provide power conversion at high efficiencies, their output is prone to ripples due to the inherent switching behavior. These switching regulators use linear-low dropout regulators (LDOs) downstream to provide clean supplies. Typically, these LDOs have good power supply rejection (PSR) at lower frequencies but this degrades at higher frequencies. Therefore, some residual ripple is still manifested on the output. Because of this, high power supply rejection (PSR) with a wide rejection frequency band is becoming a critical requirement in linear low-dropout regulators (LDOs) used in complex systems- on-chip (SOCs).
Typical LDOs achieve higher PSR within their loop-bandwidth; however, their supply rejection performance degrades with reduced loop-gain outside their loop- bandwidth. The LDOs with external filtering capacitors may also have spectral peaking in their PSR response, causing excess system- level supply noise. This work presents an LDO design approach, which achieves a PSR of higher than 68 dB up to 2 MHz frequency and over a wide range of loads up to 250 mA. The wide PSR bandwidth is achieved using a current-mode feedforward ripple canceller (CFFRC) amplifier which provides up to 25 dB of PSR improvement. The feedforward path gain is inherently matched to the forward gain of the LDO, not requiring calibration. The LDO has a fast load transient response with a recovery time of 6.1μs and has a quiescent current of 5.6μA. For a full load transition, the LDO achieves settling with overshoot and undershoot voltages below 27.6 mV and 36.36 mV, respectively. The LDO is designed and fabricated in a 180 nm bipolar/CMOS/DMOS (BCD) technology. The CFFRC amplifier helps to achieve low quiescent power due to its inherent current mode nature, eliminating the need for supply ripple summing amplifiers and adaptive biasing.
Typical LDOs achieve higher PSR within their loop-bandwidth; however, their supply rejection performance degrades with reduced loop-gain outside their loop- bandwidth. The LDOs with external filtering capacitors may also have spectral peaking in their PSR response, causing excess system- level supply noise. This work presents an LDO design approach, which achieves a PSR of higher than 68 dB up to 2 MHz frequency and over a wide range of loads up to 250 mA. The wide PSR bandwidth is achieved using a current-mode feedforward ripple canceller (CFFRC) amplifier which provides up to 25 dB of PSR improvement. The feedforward path gain is inherently matched to the forward gain of the LDO, not requiring calibration. The LDO has a fast load transient response with a recovery time of 6.1μs and has a quiescent current of 5.6μA. For a full load transition, the LDO achieves settling with overshoot and undershoot voltages below 27.6 mV and 36.36 mV, respectively. The LDO is designed and fabricated in a 180 nm bipolar/CMOS/DMOS (BCD) technology. The CFFRC amplifier helps to achieve low quiescent power due to its inherent current mode nature, eliminating the need for supply ripple summing amplifiers and adaptive biasing.
ContributorsJoshi, Kishan (Author) / Bakkaloglu, Bertan (Thesis advisor) / Garrity, Douglas (Committee member) / Seo, Jae-Sun (Committee member) / Kitchen, Jennifer (Committee member) / Arizona State University (Publisher)
Created2020
Description
Modern-day automobiles are becoming more connected and reliant on wireless connectivity. Thus, automotive electronics can be both a cause of and highly sensitive to electromagnetic interference (EMI), and the consequences of failure can be fatal. Technology advancements in engineering have brought several features into the automotive field but at the expense of electromagnetic compatibility issues. Automotive EMC problems are the result of the emissions from electronic assemblies inside a vehicle and the susceptibility of the electronics when exposed to external EMI sources. In both cases, automotive EMC problems can cause unintended changes in the automotive system operation. Robustness to electromagnetic interference (EMI) is one of the primary design aspects of state-of-the-art automotive ICs like System Basis Chips (SBCs) which provide a wide range of analog, power regulation and digital functions on the same die. One of the primary sources of conducted EMI on the Local Interconnect Network (LIN) driver output is an integrated switching DC-DC regulator noise coupling through the parasitic substrate capacitance of the SBC. In this dissertation an adaptive active EMI cancellation technique to cancel the switching noise of the DC-DC regulator on the LIN driver output to ensure electromagnetic compatibility (EMC) is presented. The proposed active EMI cancellation circuit synthesizes a phase synchronized cancellation pulse which is then injected onto the LIN driver output using an on-chip tunable capacitor array to cancel the switching noise injected via the substrate. The proposed EMI reduction technique can track and cancel substrate noise independent of process technology and device parasitics, input voltage, duty cycle, and loading conditions of the DC-DC switching regulator. The EMI cancellation system is designed and fabricated on a 180nm Bipolar-CMOS-DMOS (BCD) process with an integrated power stage of a DC-DC buck regulator at a switching frequency of 2MHz along with an automotive LIN driver. The EMI cancellation circuit occupies an area of 0.7 mm2, which is less than 3% of the overall area in a standard SBC and consumes 12.5 mW of power and achieves 25 dB reduction of conducted EMI in the LIN driver output’s power spectrum at the switching frequency and its harmonics.
ContributorsRay, Abhishek (Author) / Bakkaloglu, Bertan (Thesis advisor) / Garrity, Douglas (Committee member) / Kitchen, Jennifer (Committee member) / Seo, Jae-Sun (Committee member) / Arizona State University (Publisher)
Created2023
Description
Among the many challenges facing circuit designers in deep sub-micron technologies, power, performance, area (PPA) and process variations are perhaps the most critical. Since existing strategies for reducing power and boosting the performance of the circuit designs have already matured to saturation, it is necessary to explore alternate unconventional strategies. This investigation focuses on using perceptrons to enhance PPA in digital circuits and starts by constructing the perceptron using a combination of complementary metal-oxide-semiconductor (CMOS) and flash technology. The use of flash enables the perceptron to have a variable delay and functionality, making them robust to process, voltage, and temperature variations. By replacing parts of an application-specific integrated circuit (ASIC) with these perceptrons, improvements of up to 30% in the area and 20% in power can be achieved without affecting performance. Furthermore, the ability to vary the delay of a perceptron enables circuit designers to fix setup and hold-time violations post-fabrication, while reprogramming the functionality enables the obfuscation of the circuits. The study extends to field-programmable gate arrays (FPGAs), showing that traditional FPGA architectures can also achieve improved PPA by replacing some Look-Up-Tables (LUTs) with perceptrons. Considering that replacing parts of traditional digital circuits provides significant improvements in PPA, a natural extension was to see whether circuits built dedicatedly using perceptrons as its compute unit would lead to improvements in energy efficiency. This was demonstrated by developing perceptron-based compute elements and constructing an architecture using these elements for Quantized Neural Network acceleration. The resulting circuit delivered up to 50 times more energy efficiency compared to a CMOS-based accelerator without using standard low-power techniques such as voltage scaling and approximate computing.
ContributorsWagle, Ankit (Author) / Vrudhula, Sarma (Thesis advisor) / Khatri, Sunil (Committee member) / Shrivastava, Aviral (Committee member) / Seo, Jae-Sun (Committee member) / Ren, Fengbo (Committee member) / Arizona State University (Publisher)
Created2023
Description
Convolutional neural networks(CNNs) achieve high accuracy on large datasets but requires significant computation and storage requirement for training/testing. While many applications demand low latency and energy-efficient processing of the images, deploying these complex algorithms on the hardware is a challenging task. This dissertation first presents a compiler-based CNN training accelerator using DDR3 and HBM2 memory. An optimized RTL library is implemented to perform training-specific tasks and an RTL compiler is developed to generate FPGA-synthesizable RTL based on user-defined constraints. High Bandwidth Memory(HBM) provides efficient off-chip communication and improves the training performance. The impact of HBM2 on CNN training workloads is analyzed and compressively compared with DDR3. For training ResNet-20/VGG-like CNNs for the CIFAR-10 dataset, the proposed CNN training accelerator on Stratix-10 GX FPGA(DDR3) demonstrates 479 GOPS performance, and on Stratix-10 MX FPGA(HBM) shows 4.5/9.7 X energy-efficiency improvement compared to Tesla V100 GPU. Next, the FPGA online learning accelerator is presented. Adopting model segmentation techniques from Progressive Segmented Training(PST), the online learning accelerator achieved a 4.2X reduction in training latency. Furthermore, this dissertation presents an 8-bit floating-point (FP8) training processor which implements (1) Highly parallel tensor cores that maintain high PE utilization, (2) Hardware-efficient channel gating for dynamic output activation sparsity (3) Dynamic weight sparsity based on group Lasso (4) Gradient skipping based on FP prediction error. The 28nm prototype chip demonstrates significant improvements in FLOPs reduction (7.3×), energy efficiency (16.4 TFLOPS/W), and overall training latency speedup (4.7×) for both supervised training and self-supervised training tasks. In addition to the training accelerators, this dissertation also presents a CNN inference accelerator on ASIC(FixyNN) and FPGA(FixyFPGA). FixyNN consists of a fixed-weight feature extractor that generates ubiquitous CNN features and a conventional programmable CNN accelerator. In the fixed-weight feature extractor, the network weights are hard-coded into hardware and used as a fixed operand for the multiplication. Experimental results demonstrate FixyNN can achieve very high energy efficiencies up to 26.6 TOPS/W, and FixyFPGA achieves $2.34\times$ higher GOPS on ImageNet classification. In summary, this dissertation comprehensively discusses novel architectures of high-performance and energy-efficient ASIC/FPGA CNN inference/training accelerators.
ContributorsKolala Venkataramaniah, Shreyas (Author) / Seo, Jae-Sun (Thesis advisor) / Cao, Yu (Committee member) / Chakrabarti, Chaitali (Committee member) / Fan, Deliang (Committee member) / Arizona State University (Publisher)
Created2022
Description
Computer vision is becoming an essential component of embedded system applications such as smartphones, wearables, autonomous systems and internet-of-things (IoT). These applications are generally deployed into environments with limited energy, memory bandwidth and computational resources. This trend is driving the development of energy-effi cient image processing solutions from sensing to computation. In this thesis, diff erent alternatives are explored to implement energy-efficient computer vision systems. First, I present a fi eld programmable gate array (FPGA) implementation of an adaptive subsampling algorithm for region-of-interest (ROI) -based object tracking. By implementing the computationally intensive sections of this algorithm on an FPGA, I aim to offl oad computing resources from energy-ineffi cient graphics processing units (GPUs) and/or general-purpose central processing units (CPUs). I also present a working system executing this algorithm in near real-time latency implemented on a standalone embedded device. Secondly, I present a neural network-based pipeline to improve the performance of event-based cameras in non-ideal optical conditions. Event-based cameras or dynamic vision sensors (DVS) are bio-inspired sensors that measure logarithmic per-pixel brightness changes in a scene. Their advantages include high dynamic range, low latency and ultra-low power when compared to standard frame-based cameras. Several tasks have been proposed to take advantage of these novel sensors but they rely on perfectly calibrated optical lenses that are in-focus. In this work I propose a methodto reconstruct events captured with an out-of-focus event-camera so they can be fed into an intensity reconstruction task. The network is trained with a dataset generated by simulating defocus blur in sequences from object tracking datasets such as LaSOT and OTB100. I also test the generalization performance of this network in scenes captured with a DAVIS event-based sensor equipped with an out-of-focus lens.
ContributorsTorres Muro, Victor Isaac (Author) / Jayasuriya, Suren (Thesis advisor) / Spanias, Andreas (Committee member) / Seo, Jae-Sun (Committee member) / Arizona State University (Publisher)
Created2022
Description
Adversarial threats of deep learning are increasingly becoming a concern due to the ubiquitous deployment of deep neural networks(DNNs) in many security-sensitive domains. Among the existing threats, adversarial weight perturbation is an emerging class of threats that attempts to perturb the weight parameters of DNNs to breach security and privacy.In this thesis, the first weight perturbation attack introduced is called Bit-Flip Attack (BFA), which can maliciously flip a small number of bits within a computer’s main memory system storing the DNN weight parameter to achieve malicious objectives. Our developed algorithm can achieve three specific attack objectives: I) Un-targeted accuracy degradation attack, ii) Targeted attack, & iii) Trojan attack. Moreover, BFA utilizes the rowhammer technique to demonstrate the bit-flip attack in an actual computer prototype. While the bit-flip attack is conducted in a white-box setting, the subsequent contribution of this thesis is to develop another novel weight perturbation attack in a black-box setting. Consequently, this thesis discusses a new study of DNN model vulnerabilities in a multi-tenant Field Programmable Gate Array (FPGA) cloud under a strict black-box framework. This newly developed attack framework injects faults in the malicious tenant by duplicating specific DNN weight packages during data transmission between off-chip memory and on-chip buffer of a victim FPGA. The proposed attack is also experimentally validated in a multi-tenant cloud FPGA prototype. In the final part, the focus shifts toward deep learning model privacy, popularly known as model extraction, that can steal partial DNN weight parameters remotely with the aid of a memory side-channel attack. In addition, a novel training algorithm is designed to utilize the partially leaked DNN weight bit information, making the model extraction attack more effective. The algorithm effectively leverages the partial leaked bit information and generates a substitute prototype of the victim model with almost identical performance to the victim.
ContributorsRakin, Adnan Siraj (Author) / Fan, Deliang (Thesis advisor) / Chakrabarti, Chaitali (Committee member) / Seo, Jae-Sun (Committee member) / Cao, Yu (Committee member) / Arizona State University (Publisher)
Created2022
Description
Deep neural networks (DNNs), as a main-stream algorithm for various AI tasks, achieve higher accuracy at the cost of increased computational complexity and model size, posing great challenges to hardware platforms. This dissertation first tackles the design challenges of resistive random-access-memory (RRAM) based in-memory computing (IMC) architectures. A new metric, model stability from the loss landscape, is proposed to help shed light on accuracy under variations and model compression and guide a novel variation-aware training (VAT) solution. The proposed method effectively improves post-mapping accuracy of multiple datasets. Next, a hybrid RRAM/SRAM IMC DNN inference accelerator is developed, that integrates an RRAM-based IMC macro, a reconfigurable SRAM-based multiply-accumulate (MAC) macro, and a programmable shifter. The hybrid IMC accelerator fully recovers the inference accuracy post the mapping. Furthermore, this dissertation researches on architectural optimizations for high IMC utilization, low on-chip communication cost, and low energy-delay product (EDP), including on-chip interconnect design, PE array utilization, and tile-to-router mapping and scheduling. The optimal choice of on-chip interconnect results in up to 6x improvement in energy-delay-area product for RRAM IMC architectures. Furthermore, the PE and NoC optimizations show up to 62% improvement in PE utilization, 78% reduction in area, and 78% lower energy-area product for a wide range of modern DNNs. Finally, this dissertation proposes a novel chiplet-based IMC benchmarking simulator, SIAM, and a heterogeneous chiplet IMC architecture to address the limitations of a monolithic DNN accelerator. SIAM utilizes model-based and cycle-accurate simulation to provide a scalable and flexible architecture. SIAM is calibrated against a published silicon result, SIMBA, from Nvidia. The heterogeneous architecture utilizes a custom mapping with a bank of big and little chiplets, and a hybrid network-on-package (NoP) to optimize the utilization, interconnect bandwidth, and energy efficiency. The proposed big-little chiplet-based RRAM IMC architecture significantly improves energy efficiency at lower area, compared to conventional GPUs. In summary, this dissertation comprehensively investigates novel methods that encompass device, circuits, architecture, packaging, and algorithm to design scalable high-performance and energy-efficient IMC architectures.
ContributorsKrishnan, Gokul (Author) / Cao, Yu (Thesis advisor) / Seo, Jae-Sun (Committee member) / Chakrabarti, Chaitali (Committee member) / Ogras, Umit Y. (Committee member) / Arizona State University (Publisher)
Created2022
Description
A remarkable phenomenon in contemporary physics is quantum scarring in classically chaoticsystems, where the wave functions tend to concentrate on classical periodic orbits. Quantum
scarring has been studied for more than four decades, but the problem of efficiently detecting
quantum scars has remained to be challenging, relying mostly on human visualization of
wave function patterns. This paper develops a machine learning approach to detecting
quantum scars in an automated and highly efficient manner. In particular, this paper exploits Meta
learning. The first step is to construct a few-shot classification algorithm, under the
requirement that the one-shot classification accuracy be larger than 90%. Then propose a
scheme based on a combination of neural networks to improve the accuracy. This paper shows that
the machine learning scheme can find the correct quantum scars from thousands images of
wave functions, without any human intervention, regardless of the symmetry of the underlying
classical system. This will be the first application of Meta learning to quantum systems. Interacting spin networks are fundamental to quantum computing. Data-based tomography oftime-independent spin networks has been achieved, but an open challenge is to ascertain the
structures of time-dependent spin networks using time series measurements taken locally
from a small subset of the spins. Physically, the dynamical evolution of a spin network under
time-dependent driving or perturbation is described by the Heisenberg equation of motion.
Motivated by this basic fact, this paper articulates a physics-enhanced machine learning framework
whose core is Heisenberg neural networks. This paper demonstrates that, from local measurements, not only the local Hamiltonian can be recovered but the Hamiltonian reflecting the interacting structure of the whole system can
also be faithfully reconstructed. Using Heisenberg neural machine on spin networks of a
variety of structures. In the extreme case where measurements are taken from only one spin,
the achieved tomography fidelity values can reach about 90%. The developed machine
learning framework is applicable to any time-dependent systems whose quantum dynamical
evolution is governed by the Heisenberg equation of motion.
ContributorsHan, Chendi (Author) / Lai, Ying-Cheng (Thesis advisor) / Yu, Hongbin (Committee member) / Dasarathy, Gautam (Committee member) / Seo, Jae-Sun (Committee member) / Arizona State University (Publisher)
Created2022