BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Department of Electrical & Computer Engineering - ECPv6.16.2//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Department of Electrical & Computer Engineering
X-ORIGINAL-URL:https://ece.northeastern.edu
X-WR-CALDESC:Events for Department of Electrical & Computer Engineering
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20210314T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20211107T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20220313T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20221106T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20230312T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20231105T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20220804T140000
DTEND;TZID=America/New_York:20220804T150000
DTSTAMP:20260623T092627
CREATED:20221103T182334Z
LAST-MODIFIED:20221103T182334Z
UID:5892-1659621600-1659625200@ece.northeastern.edu
SUMMARY:Tong Jian's PhD Dissertation Defense
DESCRIPTION:“Robust Sparsified Deep Learning” \nAbstract: \nThis dissertation studies robustness issues around DNN deployments on resource constrained systems\, under both environmental and adversarial input adaptation. We propose a means of compressing a Radio Frequency deep neural network architecture through weight pruning\, and provide a systems-level analysis of the implementation of such a pruned architecture at resource-constrained edge devices. In particular\, we jointly train and sparsify neural networks tailored to edge hardware implementations. \nNext\, we propose a new learn-prune-share (LPS) algorithm for achieving robustness to environment adaptation in the field of lifelong learning. Our method maintains a parsimonious neural network model and achieves exact no forgetting by splitting the network into task-specific partitions via a weight pruning strategy optimized by the Alternating Direction Methods of Multipliers (ADMM). Moreover\, a novel selective knowledge sharing scheme is integrated seamlessly into the ADMM optimization framework to address knowledge reuse.\nFurthermore\, we investigate the Hilbert-Schmidt Information Bottleneck as regularizer (HBaR) as a means to enhance adversarial robustness. We show that the Hilbert-Schmidt Information bottleneck enhances robustness to adversarial attacks both theoretically and experimentally. In particular\, we prove that the HSIC bottleneck regularizer reduces the sensitivity of the classifier to adversarial examples. \nFinally\, we propose a novel framework Pruning-without-Adversarial-training (PwoA) for the purpose of achieving adversarial robustness on resource-constrained systems. PwoA can efficiently prune a previously trained robust neural network while maintaining adversarial robustness\, without further generating adversarial examples. We leverage concurrent self-distillation and pruning to preserve knowledge in the original model as well as regularizing the pruned model via the HBaR. \nCommittee: \nProf. Stratis Ioannidis (Advisor) \nProf. Jennifer Dy\nProf. Kaushik Chowdhury \nProf. Yanzhi Wang
URL:https://ece.northeastern.edu/event/tong-jians-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20220805T093000
DTEND;TZID=America/New_York:20220805T103000
DTSTAMP:20260623T092627
CREATED:20221103T182421Z
LAST-MODIFIED:20221103T182421Z
UID:5894-1659691800-1659695400@ece.northeastern.edu
SUMMARY:Mahdiar Sadeghi's PhD Dissertation Defense
DESCRIPTION:“Model-based decision making in life sciences” \nAbstract: \nMathematical models are key tools in rational decision-making processes. A “good” model is expected to reproduce experimental observations\, which enables predictions outside the previous experimental settings. The accuracy of predictions depends on the assumptions used to model the system. The objective of this study is to explore possible approaches to deploy models in order to generate new hypotheses in life sciences. A few biological systems relevant to protein translation\, chemotherapy\, immunotherapy\, and epidemics are considered. Models are analyzed numerically/analytically to optimize a new decision/control. In protein translation processes\, it is discovered that no switching policy is better than constant rates to maximize ribosome flow. In a particular experimental setting of chemotherapy\, a new dosing plan for chemotherapy is identified and predicted to result in maximum shrinkage of the tumor volume. In immunotherapy\, key features of binding kinetics of T-cell engagers in pre-clinical experiments are discussed. Moreover\, epidemic models under social distancing guidelines are studied. Considering a single-interval social distancing based on the start time and the duration of the social distancing shows a surprising linear relationship. Some of the results presented in this dissertation are shown to be valid in multiple applications. \n  \nCommittee: \nProf. Eduardo Sontag (Advisor)Dr. Irina KarevaProf. Carey RappaportProf. Bahram ShafaiProf. Mark Niedre
URL:https://ece.northeastern.edu/event/mahdiar-sadeghis-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20220815T140000
DTEND;TZID=America/New_York:20220815T150000
DTSTAMP:20260623T092627
CREATED:20221103T182152Z
LAST-MODIFIED:20221103T182152Z
UID:5888-1660572000-1660575600@ece.northeastern.edu
SUMMARY:Nikita Mirchandani's PhD Dissertation Defense
DESCRIPTION:“Ultra-Low Power and Robust Analog Computing Circuits and System Design Framework for Machine Learning Applications” \nAbstract: \nAs the scaling of CMOS transistors has almost halted\, performance gains of digital systems have also started to stagnate. There is a renewed interest in alternate computing techniques such as in-memory computing\, hybrid computing\, approximate computing\, and analog computing. In particular\, analog computing has reemerged as a promising alternative to save power and improve performance specifically for machine-learning (ML) applications. Power and chip area efficiency make analog computing highly appealing for implementing deep learning algorithms on-chip\, computing circuits for the internet-of-things (IoT) devices\, and implantable and wearable biomedical devices. However\, compared to digital computing\, analog computing methods have not nearly been utilized to their fullest potential due to longstanding challenges related to reliability\, programmability\, power consumption\, and high susceptibility to variations. \nThe subject of this dissertation research is to develop robust ultra-low power analog hardware suitable for machine learning applications. First\, a robust analog design methodology is presented to address issues of variability in analog circuits. A constant transconductance design technique using switched capacitor circuits is presented. The design approach is then applied to build circuits for ML applications. An analog vector matrix multiplier (VMM) is designed to be used in the convolutional layer in an ML analog computing vision hardware platform. Computing circuits are tested as part of an image classification DNN algorithm on the MNIST dataset and can achieve a classification accuracy of 96.1%.\nThe design approach is also used to design an analog computing system architecture for detection of seizures using EEG signals. A conventional EEG monitoring system includes an analog front-end (AFE)\, ADC\, digital filtering stage\, EEG feature extraction engine\, and SVM classification. Such systems suffer from high power and chip area requirements. The corresponding analog architecture is composed of AFE amplifiers to provide gain for the incoming signal. The AFE is followed by an analog filtering stage\, where spectral power from each of the bands is used as a feature for seizure classification. The output of each filter is applied to a corresponding feature extraction circuit to continuously monitor the onset of a seizure in an ultra-lower power mode with sub-threshold analog processing. The system level architecture is first modeled to obtain classification accuracy of seizures. Simulation times for the design of such complex analog systems can be prohibitively long\, particularly when the impacts of nonidealities such as noise\, nonlinearity\, and device mismatches have to be considered at the system level. The simulation time is reduced by building accurate models of the analog blocks for faster simulations. The analog models help to define the required specifications for each block in order to achieve a specified system-level classification accuracy.\nInfrastructure circuits like oscillators and voltage regulators for the proposed SoC are presented. A 254 nW 21 kHz on-chip RC oscillator with 21.5 ppm/oC temperature stability is presented to provide stable clock source for the proposed SoC. Finally\, novel lightweight hardware security primitives are described to equip individual IoT device with side-channel resistant crypto-implementations\, and unique ID or key \ngeneration. \nCommittee: \nProf. Aatmesh Shrivastava (Advisor) \nProf. Marvin Onabajo \nProf. Yong-Bin Kim
URL:https://ece.northeastern.edu/event/nikita-mirchandanis-phd-dissertation-defense/
LOCATION:432 ISEC\, 360 Huntington Ave\, Boston\, MA\, 02115\, United States
GEO:42.3396156;-71.0886534
X-APPLE-STRUCTURED-LOCATION;VALUE=URI;X-ADDRESS=432 ISEC 360 Huntington Ave Boston MA 02115 United States;X-APPLE-RADIUS=500;X-TITLE=360 Huntington Ave:geo:-71.0886534,42.3396156
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20220817T110000
DTEND;TZID=America/New_York:20220817T120000
DTSTAMP:20260623T092627
CREATED:20221103T182252Z
LAST-MODIFIED:20221103T182252Z
UID:5890-1660734000-1660737600@ece.northeastern.edu
SUMMARY:Mithun Diddi's PhD Dissertation Defense
DESCRIPTION:“Multiple UAVs for Synchronous – Shared Tasks and Long-term Autonomy” \nAbstract: \nThis thesis focuses on the use of multiple unmanned aerial vehicles(UAVs) in a distributed framework from a systems perspective to synchronously perform shared tasks such as aerial beamforming and coordinated mapping and to enhance the reliability of performing periodic (mapping) tasks at remote locations for long-term autonomous(LTA) missions. We present an autonomy stack for multiple\, heterogeneous UAVs with a simulation framework. We implemented the end-to-end pipelines for perception and communication applications involving multiple UAVs. \nRepeated deployments in harsh-weather\, real-world locations are challenging and are limited by the need for human operators. These infrastructure-poor\, remote locations pose unique challenges to long-term autonomous missions. In these locations\, harvesting power onboard using solar panels may be a viable alternative for recharging batteries.\nIn the first part of the thesis\, we focus on hardware architecture for UAVs to enable reliable LTA missions with minimal human intervention. We developed a Size\, Weight\, and Power(SWaP) constrained Smart charging stack to minimize hotel loads seen during the recharging process and enable efficient charging of batteries. This leads to the design of a standalone\, solar rechargeable quadcopter.\nReal-world applications such as reconstructing a dynamic scene from multiple viewpoints and distributed aerial beamforming require multiple robots(agents) to coordinate and synchronously act to accomplish shared tasks. These tasks require spatially distant\, multiple UAVs to have time\, phase\, and frequency synchronization. We demonstrate a Synchronous UAV(S-UAV) architecture for wireless synchronization based on GPS disciplined oscillators and the associated software framework needed for temporal registration of data across multiple UAVs.\nWe have built four S-UAVs and demonstrate the ability to 3D reconstruct a dynamic scene from overlapping viewpoints. Dynamic baseline camera arrays formed using multiple S-UAVs are used to synchronously capture a dynamic environment with people moving around. A single-time instance of synchronously captured images of the scene is used to 3D reconstruct the dynamic environment while preserving static scene assumptions of Structure from Motion(SFM). \nIn the second part of the thesis\, we focus on multi UAV autonomy framework for real-world applications of UAVs in perception\, wireless communications\, and reliable LTA missions. We present ‘Simplenav\,’ a navigation stack for heterogeneous\, multiple UAVs\, and ‘OctoRosSim\,’ a computationally lightweight multi-UAV simulation framework for validating the multi-UAV planning and autonomy pipeline. We demonstrate this framework with novel applications of end-to-end autonomy pipelines developed for a coordinated swarm of UAVs. \nCommittee: \nProf. Hanumant Singh (Advisor) \nProf. Kaushik Chowdhury \nProf. Taskin Padir
URL:https://ece.northeastern.edu/event/mithun-diddis-phd-dissertation-defense/
LOCATION:432 ISEC\, 360 Huntington Ave\, Boston\, MA\, 02115\, United States
GEO:42.3396156;-71.0886534
X-APPLE-STRUCTURED-LOCATION;VALUE=URI;X-ADDRESS=432 ISEC 360 Huntington Ave Boston MA 02115 United States;X-APPLE-RADIUS=500;X-TITLE=360 Huntington Ave:geo:-71.0886534,42.3396156
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20220817T120000
DTEND;TZID=America/New_York:20220817T130000
DTSTAMP:20260623T092627
CREATED:20221103T182520Z
LAST-MODIFIED:20221103T182520Z
UID:5896-1660737600-1660741200@ece.northeastern.edu
SUMMARY:Mengshu Sun's PhD Dissertation Defense
DESCRIPTION:“Deep Learning Acceleration on Edge Devices with Algorithm/Hardware Co-Design” \nAbstract: \nAs deep learning has succeeded in a broad range of applications in recent years\, there is an increasing trend towards deploying deep neural networks (DNNs) on edge devices such as FPGAs and mobile phones. However\, there exists a significant gap between the extraordinary accuracy of state-of-the-art DNNs and efficient implementations on edge devices\, due to their limited resources for DNNs with high computation and memory intensity. With the target of simultaneously accelerating the inference and maintaining the accuracy of DNNs\, efficient implementations are investigated of deep learning on low-power and resource-constrained devices\, by presenting algorithm/hardware co-design frameworks that incorporate hardware-friendly DNN compression algorithms with hardware design optimizations.\nFirst\, the DNN compression algorithms are explored\, leveraging quantization and weight pruning techniques. As for quantization\, intra-layer mixed precision/scheme weight quantization is proposed to boost utilization of heterogeneous FPGA resources and therefore improving the FPGA throughput\, by assigning multiple precisions and/or multiple schemes at the filter level within each layer and maintaining the same ratio of filters across all the layers for each type of quantization assignment. As for weight pruning\, novel structured and fined-grained sparsity schemes are proposed and obtained with the reweighted regularization pruning algorithm\, and then incorporated into acceleration frameworks on FPGAs to make the acceleration rate of sparse models approach the pruning rate of the number of operations.\nSecond\, the hardware implementations are studied\, proposing an automatic DNN acceleration framework to generate DNN accelerators to satisfy a target frame rate (FPS). Unlike previous approaches that start from model compression and then optimizing the FPS for hardware implementations\, this automatic framework will provide an estimation of the FPS with the FPGA resource utilization analysis and performance analysis modules\, and the bit-width is reduced until the target FPS is met and the mixing ratio for quantization precisions/schemes is automatically determined to guide the quantization process and the accelerator implementation on hardware. A resource utilization model is developed to overcome the difficulty in estimating the LUT consumption\, and a novel computing engine for DNNs is designed with various optimization techniques in support of DNN compression to improve the computation parallelism and resource utilization efficiency. \nCommittee: \nProf. Xue Lin (Advisor)\nProf. Miriam Leeser\nProf. Xiaolin Xu
URL:https://ece.northeastern.edu/event/mengshu-suns-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20220822T150000
DTEND;TZID=America/New_York:20220822T160000
DTSTAMP:20260623T092627
CREATED:20221103T182044Z
LAST-MODIFIED:20221103T182044Z
UID:5886-1661180400-1661184000@ece.northeastern.edu
SUMMARY:Hamed Mohebbi Kalkhoran's PhD Dissertation Defense
DESCRIPTION:“Machine learning approaches for classification of myriad underwater acoustic events over continental-shelf scale regions with passive ocean acoustic waveguide remote sensing” \nAbstract: \nUnderwater acoustic data contain a myriad of sound sources. Among underwater acoustic events\, marine mammal vocalization classification is one of the most challenging problems due to their transient broadband calls\, high variation in the calls of a specie\, and high similarity between the calls of some species. Here\, we developed machine learning approaches for classifying marine mammal vocalizations for real-time applications. We utilize acoustic data from a 160-element coherent hydrophone array and employ the passive ocean acoustic waveguide remote sensing technique to enable sensing and detections over instantaneous wide areas more than 100 km in diameter from the array. A variety of computational accelerating approaches\, combining hardware and software\, that make the methods desirable for real-time applications are also developed. \nThe humpback whale vocalizations can be divided into two classes\, song and non-song calls. Here we use wavelet signal denoising and coherent array processing to enhance the signal-to-noise ratio. To build features vector for every time sequence of the beamformed signals\, we employ Bag of Words approach to time-frequency features. Finally\, we apply Support Vector Machine (SVM)\, Neural Networks\, and Naive Bayes to classify the acoustic data and compare their performances. Best results are obtained using Mel Frequency Cepstrum Coefficient (MFCC) features and SVM which leads to 94% accuracy and 72.73% F1-score for humpback whale song versus non-song vocalization classification. \nTo classify a large variety of whale species vocalizations\, we extracted time-frequency features from Power Spectrogram Density (PSD) of the beamformed signals. Then we used these features to train three classifiers\, which are SVM\, Neural Networks\, and Random forest to classify six whale species: Fin\, Sei\, Blue\, Minke\, Humpback\, and general Odontocetes. We also trained a set of Convolutional Neural Networks (CNN) to detect and classify each of these six whale vocalization categories directly using Per-Channel Energy Normalization (PCEN) spectrograms. Best results were obtained with Random forest classifier\, which achieved 95% accuracy\, and 85% F1 score. To detect transient sound sources\, first we applied PCEN on the PSD of the beamformed signals. We applied thresholding on the PCEN data followed by morphological image opening to find potential sound sources and reduce noisy detections. Then we applied connected component analysis to obtain the final detected sounds for each bearing. To estimate the Direction of Arrival (DoA) of detected sounds\, we applied non-maximum suppression (NMS)\, which is widely used in object detection applications in computer vision\, on the detected sounds. We used mean power of each detected sound as the scores for NMS. To speed up the data processing\, we investigated a variety of accelerating approaches\, such as analyzing the effect of floating point precision\, applying parallel processing\, and implementing fast algorithms to run on GPU. During an experiment in the U.S. Northeast coast on board the US research vessel RV Endeavor in September 2021\, we utilized the software and hardware advances developed here to record underwater acoustic data using Northeastern University in-house fabricated large aperture 160- element coherent hydrophone array with sampling frequency of 100 kHz per element. \nCommittee: \nProf. Purnima Ratilal (Advisor) \nProf. Themistoklis Sapsis \nProf. Devesh Tiwari
URL:https://ece.northeastern.edu/event/hamed-mohebbi-kalkhorans-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20220825T140000
DTEND;TZID=America/New_York:20220825T150000
DTSTAMP:20260623T092627
CREATED:20221103T181955Z
LAST-MODIFIED:20221103T181955Z
UID:5884-1661436000-1661439600@ece.northeastern.edu
SUMMARY:Tarik Kelestemur's PhD Dissertation Defense
DESCRIPTION:Location: ISEC 532 \n“Combining Classical and Learning-based Methods for Visual and Tactile Manipulation” \nAbstract: \nRobots that operate in dynamic and ever-changing environments need to make sense of their surroundings and act in them safely and efficiently. This requires the integration of multiple sensory modalities such as visual and tactile. Humans can naturally fuse different feedbacks from the environment or substitute them with one another to perform everyday tasks. For example\, to use a computer mouse\, we first locate the mouse using vision and then use touch feedback from our fingers to precisely localize the buttons. Ideally\, we would like robots to have human-level perception and control of the environment to achieve various tasks. This dissertation address two significant problems toward this overarching goal. \nThe first problem we consider in this dissertation is figuring out how to use tactile information in conjunction with visual feedback. Robotic manipulators that interact with objects and environments are often equipped with visual sensors such as RGB and depth cameras. They estimate the state of their environment using these sensors and act upon the estimated state. While a large body of previous work has shown that we can achieve impressive results only with visual sensors\, more precise and delicate tasks require touch information which gives direct feedback from the environment. To this end\, we propose methods for efficiently combining the tactile and visual information to leverage the advantages of these modalities.\nThe second problem we investigate is how to build visual and tactile manipulation methods that can generalize over the different novel environments and objects. The rise of deep learning has enabled robots to solve challenging perception and control problems using visual and tactile observations while generalizing to novel objects and environments. However\, a common issue among deep learning-based methods is that these methods usually work only within the distribution of the training data and do not perform well when they are presented with unseen examples. Furthermore\, they cannot distinguish whether they are dealing with in or out-of-distribution data. We propose to address this issue by combining well-established and principled algorithmic priors with the generalization capabilities of deep learning. \nIn the first part of this dissertation\, we investigate the problem of pose estimation of the robotic grippers with respect to the environment and objects. The proposed framework introduces a learnable Bayes filter that can estimate the position of a gripper in a single image of the environment. We learn the observation and motion models of the Bayes filter using modern neural network architectures and use recursive belief updates for tracking the position of the gripper over time. Later\, the belief estimation is used as an input to policies where the aim is to solve manipulation tasks using tactile feedback. In the second part\, we look at the problem of estimating shapes from partial observations. We propose a method called DeepGPIS that combines a powerful deep learning-based implicit shape representation with a non-parametric inference approach model for implicit surfaces (GPIS) which allows us to generate complete shapes of novel objects and estimate their predictive uncertainties. \nCommittee: \nProf. Taskin Padir (Advisor) \nProf. Robert Platt (Advisor) \nProf. David Rosen (Advisor)
URL:https://ece.northeastern.edu/event/tarik-kelestemurs-phd-dissertation-defense/
END:VEVENT
END:VCALENDAR