BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Department of Electrical & Computer Engineering - ECPv6.15.20//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:20190310T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20191103T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20200308T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20201101T060000
END:STANDARD
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
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20210112T100000
DTEND;TZID=America/New_York:20210112T110000
DTSTAMP:20260428T221742
CREATED:20201222T022618Z
LAST-MODIFIED:20201222T022618Z
UID:4647-1610445600-1610449200@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Haoqing Li
DESCRIPTION:PhD Proposal Review: Robust Processing against Interferences in GNSS Navigation \nHaoqing Li \nLocation: Zoom Link \nAbstract: Satellite-based navigation is prevalent as positioning applications among our lives\, how-ever\, this high reliance brings potential threats when different interferences and jamming signals are considered. Jamming devices\, although illegal in many countries\, can be easily to get. Those devices can broadcast high-power jamming signals in Global Navigation Satellite System (GNSS) frequency band to destroy receiver’s performance. While jamming signals are illegal and we may get rid of it with the power of law\, other kinds of interferences will cannot even be avoided. Distance Measuring Equipment (DME) signal is applied to measure the distance between aircraft and ground station\, significant in aircraft transport but interference in GNSS processing. Besides\, the GNSS signal itself can also be a interference after reflection and refraction. Since we couldn’t simply re-move those from the source\, methods to mitigate influences of interferences is necessary for stable performance of receiver. There are three main blocks in GNSS receiver: acquisition block\, tracking block and positioning block\, where influence of interferences could be eliminate to get an accurate Position\, Velocity\, and Time (PVT) solution. In this article\, robust statistics processing is applied as one of the interference mitigation methods. This method aims to lower influence of outliers\, which is the presence of many kinds of interferences in either time domain or transformed domain. Robust statistics processing can be used in pre-correlation in both acquisition block and tracking block\, while a robust Kalman filter is designed in positioning block to get rid of interferences. Deep learning\, achieving extraordinary performance in many application domains\, also provides improvement to tracking block against multipath problem. A deep neural network is built to substitute the whole tracking loop to bring robustness to receiver.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-haoqing-li/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20210108T140000
DTEND;TZID=America/New_York:20210108T150000
DTSTAMP:20260428T221742
CREATED:20210107T213951Z
LAST-MODIFIED:20210107T213951Z
UID:4660-1610114400-1610118000@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Sungho Kang
DESCRIPTION:PhD Proposal Review: Metamaterial Absorbers for Infrared Sensing Microsystems \nSungho Kang \nLocation: Zoom \nAbstract: Infrared (IR) spectroscopic sensing has become a key technique in multidisciplinary environments such as military applications\, industrial safety control\, and smart homes\, by providing an accurate and non-disruptive analysis of the target objects. Recently the demand for high performance and compact IR spectroscopy systems has been steadily growing due to the advent of Internet of Things and the burgeoning development of miniaturized sensors. The key challenge lies in realizing high performance IR detectors that have low noise\, high IR throughput\, and spectral sensitivity in a miniaturized form factor. This challenge has been tackled in the study of micro-electromechanical sensing systems and metamaterial absorbers\, in which the ultra-high resolution sensing capability and the near-perfect IR absorption properties can be simultaneously exploited in a minimized footprint. The metal-insulator-metal (MIM) IR absorbers\, in particular\, are characterized by the near-unity absorptance with lithographically tunable peak absorption wavelength and spectral selectivity in an ultra-thin form factor\, suitable for the implementation of miniaturized spectroscopic IR microsystems. The exceptional IR absorption characteristics realized by the MIM IR absorbers and their sub-wavelength form factor allow for seamless integration with the existing IR sensing microsystem and the unprecedented IR sensing performance for the next generation IoT sensing solutions. In this proposal\, novel development of zero-power long-wavelength IR (LWIR) detector and miniaturized IR spectroscopic sensor based on the two key technologies are presented: (1) plasmonically-enhanced LWIR micromechanical photoswitch and (2) multispectral resonant IR detector array.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-sungho-kang/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201217T140000
DTEND;TZID=America/New_York:20201217T150000
DTSTAMP:20260428T221742
CREATED:20201210T000522Z
LAST-MODIFIED:20201210T000522Z
UID:4619-1608213600-1608217200@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Kaidi Xu
DESCRIPTION:PhD Proposal Review: Towards Empirical Implementation and Theoretical Analysis in Adversarial Machine Learning \nKaidi Xu \nLocation: Zoom Link \nAbstract: Deep learning or deep neural networks (DNNs) have achieved extraordinary performance in many application domains such as image classification\, object detection and recognition\, natural language processing and medical image analysis. It has been well accepted that DNNs are vulnerable to adversarial attacks\, which raises concerns of DNNs in security-critical applications and may result in disastrous consequences. Adversarial attacks are usually implemented by generating adversarial examples\, i.e.\, adding sophisticated perturbations\nonto benign examples\, such that adversarial examples are classified by the DNN as target (wrong) labels instead of the correct labels of the benign examples. The adversarial machine learning aims to study this phenomenon and leverage it to build robust machine learning systems and explain DNNs.\nIn this dissertation\, we present the mechanism of adversarial machine learning in both empirical and theoretical ways. Specifically\, we first introduce a uniform adversarial attack generation framework\, structured attack (StrAttack)\, which explores group sparsity in adversarial perturbations by sliding a mask through images aiming for extracting key spatial structures. Second\, we discuss the feasibility of adversarial attacks in the physical world and introduce a powerful framework\, Expectation over Transformation (EoT). Utilize EoT with Thin Plate Spline (TPS) transformation\, we can generate Adversarial T-shirts\, a robust physical adversarial example for evading person detectors even if it could undergo non-rigid deformation due to a moving person’s pose changes.\nThird\, we stand on the defense side and propose the first adversarial training method based on Graph Neural Network.\nFinally\, we introduce Linear relaxation based perturbation analysis (LiRPA) for neural networks\, which computes provable linear bounds of output neurons given a certain amount of input perturbation.\nLiRPA studies the adversarial example in a theoretical way and can guarantee the test accuracy of a model by given perturbation constraints.\nIn the future\, we plan to study a novel patch transformer network to truthfully model real-world physical transformations empirically. In addition\, at the formal robustness direction\, we plan to explore the complete verification\, that given sufficient time\, the verifier should give a definite “yes/no” answer for a property under verification. Our LiRPA framework combining with GPUs may accelerate this procedure.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-kaidi-xu/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201216T140000
DTEND;TZID=America/New_York:20201216T150000
DTSTAMP:20260428T221742
CREATED:20201210T000756Z
LAST-MODIFIED:20201210T000756Z
UID:4621-1608127200-1608130800@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Amirreza Farnoosh
DESCRIPTION:PhD Proposal Review: Unsupervised Learning of Low-Dimensional Dynamical Representations from Spatiotemporal Data \nAmirreza Farnoosh \nLocation: Zoom Link \nAbstract: Ever-improving sensing technologies offer a fast and accurate collection of large-scale spatiotemporal data\, recorded from multimodal sensors of heterogeneous natures\, in various application domains\, ranging from medicine and biology to robotics and traffic control. In this proposal\, we are learning the underlying representation of these data in an unsupervised manner\, tailored towards several emerging applications\, namely indoor navigation and mapping\, neuroscience hypothesis testing\, and time series segmentation and forecasting.\nAs such\, (1) we present an unsupervised framework for real-time depth and view-angle estimation from an inertially augmented video recorded from an indoor scene by employing geometric-based machine learning and deep learning models. (2) We introduce a hierarchical deep generative factor analysis framework for temporal modeling of neuroimaging datasets. Our model approximates high dimensional data by a product between time-dependent weights and spatially dependent factors which are in turn represented in terms of lower dimensional latents. This framework can be extended to perform clustering in the low dimensional temporal latent or perform factor analysis in the presence of a control signal. (3) We present a deep switching dynamical system for dynamical modeling of multidimensional time-series data. Specifically\, we employ a deep vector auto-regressive latent model switched by a chain of discrete latents to capture higher-order multimodal latent dependencies. This results in a flexible model that (i) provides a collection of potentially interpretable states abstracted from the process dynamics\, and (ii) performs short- and long-term vector time series prediction in a complex multi-relational setting.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-amirreza-farnoosh/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201216T140000
DTEND;TZID=America/New_York:20201216T150000
DTSTAMP:20260428T221742
CREATED:20201205T015159Z
LAST-MODIFIED:20201205T015159Z
UID:4614-1608127200-1608130800@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Pu Zhao
DESCRIPTION:PhD Proposal Review: Towards Robust Image Classification with Deep learning and Real-Time DNN Inference on Mobile \nPu Zhao \nLocation: Zoom Link \nAbstract: As the rapidly increasing popularity of deep learning\, deep neural networks (DNN) have become the fundamental and essential building blocks in various applications such as image classification and object detection. However\, there are two main issues which potentially limit the wide application of DNNs: 1) the robustness of DNN models raises security concerns\, and 2) the large computation and storage requirements of DNN models lead to difficulties for its wide deployment on popular yet resource-constrained devices such as mobile phones.\nTo investigate the DNN robustness\, we explore the DNN attack\, robustness evaluation and defense. More specifically\, for DNN attack\, we achieve various attack goals (e.g. adversarial examples and fault sneaking attacks) with different algorithms (e.g. alternating direction method of multipliers (ADMM) and natural gradient descent (NGD) attacks) under various conditions (white-box and black-box attacks). For robustness evaluation\, we propose a fast evaluation method to obtain the model perturbation bound such that any model perturbation within the bound does not alter the model classification outputs or incur model mis-behaviors. For the DNN defense\, we investigate the defense performance with model connection techniques and successfully mitigate the fault sneaking and backdoor attacks.\nWith a deeper understanding of the DNN robustness\, we further explore the deployment problem of DNN models on edge devices with limited resources. To satisfy the storage and computation limitation on edge devices\, we adopt model pruning to remove the redundancy in models\, thus reducing the storage and computation during inference. Besides\, as some applications have real-time requirements with high inference speed sensitivities such as object detection on autonomous cars\, we further try to implement real-time DNN inference for various DNN applications on mobile devices with pruning and compiler optimization. To summary\, we mainly investigate the DNN robustness and implement real-time DNN inference on the mobile.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-pu-zhao/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201216T100000
DTEND;TZID=America/New_York:20201216T110000
DTSTAMP:20260428T221742
CREATED:20201214T194420Z
LAST-MODIFIED:20201214T194420Z
UID:4631-1608112800-1608116400@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Hongjia Li
DESCRIPTION:PhD Proposal Review: Automation Design and DNN Acceleration Algorithms: From Software Implementation to Hardware Physical Design \nHongjia Li \nLocation: Zoom Link \nAbstract: Deep learning has been growing at a fast pace in recent years and has been expanded into many application fields\, with a wide range from image recognition\, object detection to medical applications. Meanwhile\, edging devices such as mobile devices are rapidly becoming the central computer and carrier for deep learning tasks. However\, real-time execution has been limited due to the computation/storage resource constraints on these devices.\nIn this proposal review\, I will dive into some aspects of DNN acceleration methods\, including model compression techniques and software implementation optimizations. The goal is to achieve an unprecedented\, real-time performance of large-scale neural network inference on edging devices. Additionally\, an efficient physical design automation design is introduced for Adiabatic Quantum-Flux-Parametron (AQFP) circuits\, meeting the unique features and constraints.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-hongjia-li/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201216T100000
DTEND;TZID=America/New_York:20201216T110000
DTSTAMP:20260428T221742
CREATED:20201210T000641Z
LAST-MODIFIED:20201210T000641Z
UID:4620-1608112800-1608116400@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Majid Sabbagh
DESCRIPTION:PhD Proposal Review: The perils of shared computing: A hardware security perspective \nMajid Sabbagh \nLocation: Teams Link \nAbstract: The enormous computation power of modern processors and accelerators has rendered them shared computing resources for multiple users and applications\, both in the cloud and on the edge. Despite software techniques for security such as virtualization and containers\, recently a new attack surface is emerging that pertains to the hardware vulnerabilities of shared computing resources\, posing serious threats to shared computing.\nFault attacks (FAs) and Side-Channel Attacks (SCAs) are two hardware-oriented attacks that target the system implementations. FAs aim to tamper the integrity of application execution through different fault injection methods\, to compromise the data or disrupt computation at run-time. SCAs exploit the information leakage of sensitive applications in physical parameters\, such as power consumption\, electromagnetic emanations\, and timing\, to breach the confidentiality of the application. \nIn this dissertation\, we introduce a new class of FAs against Graphics Processing Units (GPUs)\, called overdrive fault attacks. We discover the security vulnerability of GPU’s voltage-frequency scaling (VFS) mechanism\, a common feature to balance power consumption and performance. An out-of-specification configuration of GPU voltage and frequency can be set by an adversary on the host CPU\, through the software interfaces to GPU’s power management units. This setting will cause timing violations for the computation and result in silent data corruptions (SDCs). We apply the overdrive fault attacks on two common victim applications. One is cryptographic applications accelerated by GPU. We launch a differential fault analysis (DFA) attack on an AES kernel running on an AMD RX 580 GPU and successfully recover the secret key. The other victim is deep neural network (DNN) inference. In modern GPUs that support multiple kernels\, the adversary is able to track the execution of the victim DNN through shared resources and control the timing of fault injections precisely. We launch a successful attack on a convolutional neural network kernel running on an NVIDIA RTX 2080 SUPER GPU with misclassifications. We further study the characteristics of fault injections and the fault propagation through the network.\nWe evaluate a timing side-channel attack called Prime+Probe attack on Central Processing Units (CPUs) and propose a Side-Channel Attack DEtection Tool (SCADET). SCADET is a methodology and a tool that analyzes an x86 program’s memory accesses. It records and analyzes the memory accesses using dynamic binary instrumentation by running the program in a controlled environment to accurately identify the malicious access patterns corresponding to the Prime+Probe attack.\nFinally\, I propose an FPGA-based RISC-V processor prototype as an evaluation platform for various cache timing attacks and transient attacks\, and implement a taint tracking-based countermeasure against transient attacks. For the first phase\, we have ported spectre v1 and v2 and return-stack-buffer attack to the SonicBOOM RISC-V processor.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-majid-sabbagh/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201216T093000
DTEND;TZID=America/New_York:20201216T103000
DTSTAMP:20260428T221742
CREATED:20201214T194558Z
LAST-MODIFIED:20201214T194558Z
UID:4632-1608111000-1608114600@ece.northeastern.edu
SUMMARY:ECE MS Thesis Defense: Jinghan Zhang
DESCRIPTION:MS Thesis Defense: Allocating One Common Accelerator-Rich Platform for Many Streaming Applications \nJinghan Zhang \nLocation: Zoom Link \nAbstract: Many demanding streaming applications share functional and structural similarities with other applications in their respective domain\, e.g. video analytics\, software-defined radio\, and radar. This opens the opportunity for specialization (e.g. heterogeneous computing) to achieve the needed efficiency and/or performance. However\, current Design Space Exploration (DSE) focuses on an individual application in isolation (e.g. one particular vision flow)\, but not a set of similar applications. Hence\, optimizations that occur due to considering multiple applications simultaneously are missed. New DSE methodologies and tools are needed with a broader scope of application sets instead of individual applications.\nThis thesis introduces a novel Domain DSE approach focusing on streaming applications. Key contributions are: (1) a formalized method to extract the functional and structural similarities of domain applications\, (2) domain application generation to provide enough synthetic domains as study cases\, (3) a rapid platform performance estimation and comparison at two abstraction levels: Domain Score (DS) and Analytic Performance Estimation (APE) model\, (4) a methodology to evaluate a platform’s benefit for a set of applications\, and (5) two novel algorithms\, Dynamic Score Selection (DSS) and GenetIc Domain Exploration (GIDE)\, for hardware/software partitioning of a domain-specific platform to maximize the throughput across domain applications (under certain constraints).\nThis thesis demonstrates DSS’s and GIDE’s benefits using OpenVX applications and synthetic domains. The DSS and GIDE generated domain-specific platforms improve performance over application-specific platforms by 58%\, and 75% for OpenVX\, as well as by 23% and 48% for synthetic applications. GIDE’s platforms reach 99.8% (OpenVX) and 97.6% (synthetic) throughput of the domain optimal platform obtained through exhaustive search.
URL:https://ece.northeastern.edu/event/ece-ms-thesis-defense-jinghan-zhang/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201215T110000
DTEND;TZID=America/New_York:20201215T120000
DTSTAMP:20260428T221742
CREATED:20201214T194726Z
LAST-MODIFIED:20201214T194726Z
UID:4633-1608030000-1608033600@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Arjun Singh
DESCRIPTION:PhD Proposal Review: Design\, Modeling and Operation of Plasmonic Devices for Smart Communication Systems in the Terahertz Band \nArjun Singh \nLocation: Teams Link \nAbstract: The terahertz (THz) band is an attractive spectral resource for future communication systems\, for supporting very high-speed data rates and increasingly dense networks. However\, the lack of a well-developed technology that operates at these frequencies has remained a challenge for the scientific community. The very high propagation losses at THz frequencies and the decimating impact of everyday objects on THz wave propagation necessitate an up-haul of the conventional communication link\, with smart control over the radiation\, propagation\, and detection of THz signals. To overcome these obstacles\, novel plasmonic devices that exploit the attractive properties of graphene have been proposed. However\, there are several challenges\, such as low output power and high reflection losses\, that are not yet addressed. The objective of the proposed research herein is to facilitate an end-to-end communication link with graphene plasmonics as the cornerstone of the fundamental device physics. The devices designed can be utilized at both the communication endpoints\, as well as across the channel\, to effectively overcome the limited communication distance – The grand challenge of the THz band.\nTo this end\, a graphene-based plasmonic array architecture is first proposed\, explained\, and modeled. The fundamental radiating element of the array architecture\, called the plasmonic front-end\, consists of a self-sufficient plasmonic source\, a plasmonic modulator that acts as a phase controller\, and a plasmonic nano-antenna. The array designed through an integration of these front-ends is compact and provides complete beamsteering support\, with a new tailored algorithm developed for beamforming weight selection. Numerical evaluations and full-wave finite difference frequency domain (FDFD) simulations with COMSOL Multi-physics are utilized to verify array operation. The array is also demonstrated to provide a strong effective isotropic radiated power (EIRP)\, that increases exponentially with array size. To mitigate the negative effects of the channel environment\, such as unwanted blockages and high path losses for simpler devices\, a hybrid reflectarray is presented. The fundamental element is modeled as a jointly designed and integrated metal-graphene patch. Numerical and simulation results are utilized to demonstrate the attractive properties of the proposed reflectarray as compared to other proposed counterparts\, including independence from the incoming angle of the impinging wave\, dynamic phase control capability\, and a strong reflection efficiency. The unique design properties of the plasmonic array\, as well as the hybrid reflectarray\, open the option of incorporating techniques such as multi-beam beamforming design and interleaved\, independent arrays\, to boost the channel capacity.\nAs a part of the proposed work\, the impact of the design properties of these devices on the communications link will be investigated by developing the fundamental problem and considering all trade-offs. The undertaking will be significantly more robust and conclusive than those that have been performed previously\, both due to the consideration of a complete end to end link\, as well as the incorporation of the characteristics of the device design model. Finally\, preliminary fabrication results in the realization of these devices are presented\, and the roadmap ahead is outlined.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-arjun-singh/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201212T103000
DTEND;TZID=America/New_York:20201212T113000
DTSTAMP:20260428T221742
CREATED:20201207T214335Z
LAST-MODIFIED:20201207T214335Z
UID:4617-1607769000-1607772600@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Ning Liu
DESCRIPTION:PhD Dissertation Defense: Real-World Applicable Deep Learning Techniques: From Efficient Modeling to Automated Model Optimization \nNing Liu \nLocation: Zoom Link \nAbstract: Recently\, deep neural networks (DNNs) have been widely studied and achieved tremendous success in a variety of real-world applications\, such as computer vision\, medical diagnosis and machine translation. Deep reinforcement learning (DRL)\, as an emerging powerful deep learning technique\, combines DNNs with reinforcement learning into an interactive system. DRL opens up many new applications in domains such as healthcare\, robotics and smart grids. With the rapid evolution of IT infrastructures\, cloud computing has been witnessed as the prevailing computing paradigm. The underlying infrastructure of cloud computing relies on a large amount of data centers. The energy efficiency issue from “cloud” becomes more crucial and calls for more attentions.\nIn this dissertation\, to solve the real-world energy efficiency problems\, we take advantage of the deep learning and deep reinforcement learning techniques for efficient modeling of “cloud” applications. We present a DNN-based power management framework for regulation service and a novel DRL-based hierarchical framework for solving the overall resource allocation and power management problem. On the other hand\, the powerful DNNs themselves are massive\, consuming tremendous energy. Therefore\, we explore the efficiency on deep neural networks. We propose an automatic model pruning framework to reduce the storage and computation requirements and accelerate inference. Our framework outperforms the prior work on automatic model compression by up to 33× in pruning rate (120× reduction in the actual parameter count) under the same accuracy. Significant inference speedup has been observed from the proposed framework on actual measurements on smartphone.
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-ning-liu/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201211T150000
DTEND;TZID=America/New_York:20201211T160000
DTSTAMP:20260428T221742
CREATED:20201207T214155Z
LAST-MODIFIED:20201207T214227Z
UID:4615-1607698800-1607702400@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Qifan Li
DESCRIPTION:PhD Dissertation Defense: Development of Magnetodielectric Materials with Low Loss and High Snoek’s Product for Microwave Applications \nQifan Li \nLocation: Teams Link \nAbstract: Exhibiting both relative magnetic permeability and electric permittivity greater than unity\, magnetodielectric materials have been attracting great attention in both academia and industry for next-generation communication\, sensing\, and radar applications. It is always of great interest for researchers to tailor the magnetic properties of magnetodielectric materials for high permeability\, low magnetic loss and large Snoek’s product towards higher-frequency applications.\nHexagonal ferrites form an important group of magnetodielectric materials. Besides the six best known hexagonal structures\, i.e.\, M-\, W-\, X-\, Y-\, Z- and U-type hexaferrites\, some unique hexagonal structures\, named 18H hexaferrites\, were discovered in 1970s. For the first time\, the dynamic magnetic properties and their temperature dependence of polycrystalline Mg-Zn 18H hexaferrites at microwave frequencies are investigated. Owing to a remarkably low damping coefficient\, the frequency dispersion of complex permeability reveals a narrow and strong resonance. The Mg-Zn 18H hexaferrites show excellent loss tangent of 0.07 at 3 and 4 GHz. Accordingly\, narrow FMR linewidths in the range of 486-660 Oe are measured. The temperature dependence of the damping coefficient is 0.0004 /°C\, indicating a small variation of the intrinsic loss with temperature. These results are the best performance among the polycrystalline microwave ferrites reported so far for the S- and C-band applications.\nMagnetodielectric composites\, prepared by dispersing magnetic particles homogenously in an electrically insulating matrix\, are another type of magnetodielectric materials. It is crucial to predict the effective magnetic properties of the multi-phase mixture. A modified effective medium theory is proposed by extending the traditional formulas with the effects of particle-size distribution and clustering of inclusions. Its accuracy is verified by two kinds of magnetodielectric composites over wide ranges of both particle concentration and frequency.\nThe magnetic properties of microwave ferrites are strongly affected by their polycrystalline microstructure\, which is mainly controlled by the sintering process. The two-step sintering technique is systematically studied for the preparation of hexaferrites. With optimal combinations of sintering temperatures in each step\, significant reduction in magnetic loss and enhancement in Snoek’s product are achieved with uniform and fine-grained structures.\nPrecise measurement of broadband permeability and permittivity is crucial to develop advanced magnetodielectric materials. A straightforward\, explicit and noniterative method is proposed by eliminating the error from the direct measurement of sample position in the standard Nicolson-Ross-Weir method. Based on the results from two kinds of magneto-dielectric materials measured in two sets of test fixtures of different geometries\, this method is theoretically and experimentally proven to have high and position-independent accuracy over a wide frequency range.\nFinally\, a patch antenna on Mg-18H magnetodielectric substrate is designed to operate at 3.6 GHz for 5G wireless communication. Benefiting from the large refractive index of the magnetodielectric material\, the size of the patch antenna is significantly reduced. Moreover\, compared to the dielectric substrate providing the same miniaturization factor\, magnetodielectric antennas exhibit significant advantages for larger bandwidth and gain.
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-qifan-li/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201211T140000
DTEND;TZID=America/New_York:20201211T150000
DTSTAMP:20260428T221742
CREATED:20201210T000912Z
LAST-MODIFIED:20201210T000912Z
UID:4622-1607695200-1607698800@ece.northeastern.edu
SUMMARY:ECE MS Thesis Defense: Xinan Huang
DESCRIPTION:MS Thesis Defense: Exploring Effectiveness of Naive Spatio-Temporal Exploits for Depth Completion \nXinan Huang \nLocation: Zoom Link \nAbstract: With an increasing need for usable depth for autonomous navigation systems such as self-driving cars\, depth completion is becoming an increasingly studied subject. RGB data provide much-needed aid in providing good recreation of dense depth maps from sparse LiDAR output. Yet\, these data are also provided in sequential form. And thus for this thesis\, we aim to explore how effective using network layers that exploit Spatio-temporal features would be in achieving higher depth completion accuracy. We propose adding 3D convolutional layers and ConvGRU layers to a preexisting depth completion network and perform ablation studies on the effectiveness of these methods. We were able to verify that naive approaches are able to garner improvements quantitatively and qualitatively\, but training results show that additional geometric constraints would perhaps boost such exploits even further for better depth completion results.
URL:https://ece.northeastern.edu/event/ece-ms-thesis-defense-xinan-huang/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201210T090000
DTEND;TZID=America/New_York:20201210T100000
DTSTAMP:20260428T221742
CREATED:20201207T234717Z
LAST-MODIFIED:20201207T234717Z
UID:4618-1607590800-1607594400@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Chenyang Zhu
DESCRIPTION:PhD Dissertation Defense: Remote Monitoring of Multiple Ships over Instantaneous Continental-shelf Scale Region with a Large-aperture Coherent Hydrophone Array \nChenyang Zhu \nLocation: Zoom Link \nAbstract: Multiple mechanized ocean vessels\, including both surface ships and submerged vehicles\, can be simultaneously monitored over instantaneous continental-shelf scale regions >10\,000 km 2 via passive ocean acoustic waveguide remote sensing. A large-aperture densely-sampled coherent hydrophone array system is employed in the Norwegian Sea in Spring 2014 to provide directional sensing in 360 degree horizontal azimuth and to significantly enhance the signal-to-noise ratio (SNR) of ship-radiated underwater sound\, which improves ship detection ranges by roughly two orders of magnitude over that of a single hydrophone. Here\, 30 mechanized ocean vessels spanning ranges from nearby to over 150 km from the coherent hydrophone array\, are detected\, localized and classified. The vessels are comprised of 20 identified commercial ships and 10 unidentified vehicles present in 8 h/day of POAWRS observation for two days. The underwater sounds from each of these ocean vessels received by the coherent hydrophone array are dominated by narrowband signals that are either constant frequency tonals or have frequencies that waver or oscillate slightly in time. The estimated bearing-time trajectory of a sequence of detections obtained from coherent beamforming are employed to determine the horizontal location of each vessel using the Moving Array Triangulation (MAT) technique. For commercial ships present in the region\, the estimated horizontal positions obtained from passive acoustic sensing are verified by Global Positioning System (GPS) measurements of the ship locations found in historical AIS database. We provide time-frequency characterizations of the underwater sounds radiated from the commercial ships and the unidentified vessels. The time-frequency features along with the bearing-time trajectory of the detected signals are applied to simultaneously track and distinguish these vessels.\nNext\, three approaches for simultaneous ship long-range automatic detection\, acoustic signature characterization\, and bearing-time trajectory estimation have been developed and applied\, each focusing on a different aspect of a ship’s radiated underwater sound received on a large-aperture densely-sampled coherent hydrophone array. (i) Ships narrowband machinery tonal sound is analyzed via temporal coherence using Mean Magnitude-Squared Coherence (MMSC) calculations. (ii) Ships broadband cavitation noise amplitude modulated by propeller rotation is examined using Cyclic Spectral Coherence (CSC) analysis that provides estimates for propeller blade pass rotation frequency\, shaft rotation frequency\, and hence the number of propeller blades. (iii) Mean power spectral densities averaged across specific broad bandwidths are calculated to detect and compare output sound pressure levels from acoustically energetic ships. Each of these techniques are applied after coherent beamforming of the received acoustic signals on a coherent hydrophone array\, leading to significantly enhanced signal-to-noise ratios for simultaneous detection and characterization of multiple ships over continental-shelf scale regions. The approaches are illustrated by application to\nroughly two hours of acoustic recordings of a 160-element coherent hydrophone array deployed in the Norwegian Sea during an experiment in February 2014. Six ocean vessels are simultaneously detected and their acoustic signatures characterized\, located at a variety of bearings and ranges out to 200 km from the coherent hydrophone array\, with speeds ranging from 0.5 knots to 13 knots\, verified by Global Positioning System (GPS) information from Automatic Identification System (AIS) database. Hybrid usage of the three methods provide a robust approach for ship characterization in terms of machinery tonal sound signature\, propeller rotation signature\, and ship broadband energetics that can be employed for efficient ship classification. The CSC approach is demonstrated to be also useful for automatic detection and bearing-time estimation of repetitive marine mammal vocalizations present in coherent hydrophone array recordings\, providing estimates of inter-pulse-train and inter-pulse intervals from CSC spectra cyclic fundamental and first recurring peak frequencies respectively.
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-chenyang-zhu/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201208T150000
DTEND;TZID=America/New_York:20201208T160000
DTSTAMP:20260428T221742
CREATED:20201203T223238Z
LAST-MODIFIED:20201203T223238Z
UID:4606-1607439600-1607443200@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Sheng Lin
DESCRIPTION:PhD Dissertation Defense: Platform-specific Model Compression for Deep Neural Networks with Joint Methods \nSheng Lin \nLocation: Zoom Link \nAbstract: Deep learning has delivered its powerfulness in many application domains\, especially in computer vision\, natural language processing and speech recognition. As the backbone of deep learning\, deep neural networks (DNNs) consist of multiple layers of various types with hundreds to thousands of neurons. Embedded platforms are now becoming essential for deep learning deployment due to their portability\, versatility\, and energy efficiency. The large model size of DNNs\, while providing excellent accuracy\, also burdens the hardware platforms with intensive computation and storage. To consider the requirements of specific tasks\, many researchers have investigated reducing DNN model size for efficient implementation in hardware devices with reasonable accuracy prediction. However\, it lacks a systemic investigation on platform-specific DNN acceleration frameworks. \nIn this dissertation\, we present several software-hardware co-design techniques to speed up the DNN algorithm on specific platforms. At the software level\, we present joint model compression techniques for DNN model training and inference with reasonable accuracy performance. At the hardware level\, these algorithms and methods are targeting storage reduction\, low power consumption\, efficient inference\, and data security. By using joint methods to optimize different types of networks\, the targeted hardware platforms can reduce asymptotic complexity of both computation and storage\, making our approach distinguished from existing approaches. First\, we present a Fast Fourier Transform-based DNN model for inference phase on embedded platforms. Second\, we build a framework for two most commonly used model compression techniques\, low-bit linear weight quantization and its combination with different weight pruning methods. Third\, we apply quantization techniques for the always-on keyword spotting system and eliminate the energy-consuming ADC with an energy-efficient analog processing circuit. Finally\, we propose a federated learning framework to protect user’s data privacy while reducing overall communication cost during the training process.
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-sheng-lin/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201208T140000
DTEND;TZID=America/New_York:20201208T150000
DTSTAMP:20260428T221742
CREATED:20201202T012637Z
LAST-MODIFIED:20201202T012637Z
UID:4605-1607436000-1607439600@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Raffaele Guida
DESCRIPTION:PhD Dissertation Defense: Remotely Rechargeable Embedded Platforms for Next Generation IoT Systems in Critical Environments \nRaffaele Guida \nLocation: Teams Meeting \nAbstract: In the near future\, a new generation of miniaturized\, multi-function and smart wireless devices for Internet of Things (IoT) systems\, designed for real-time monitoring and with real-time reconfiguration will be deployed in critical and challenging environments\, e.g.\, underwater and inside the human body. These futuristic IoT platforms can now be realized thanks to advances in low-power electronics and wireless communications. However\, the need for long-term and reliable power supply\, together with the need to support innovative functions\, impose new powering requirements that cannot be satisfied by traditional batteries. Batteries have in fact a major impact on the size and lifetime of the device\, and often need to be replaced through complex\, expensive and non-scalable procedures. For example\, powering of Internet of Underwater Things (IoUT) devices in deep water remains one of the main challenges\, since these systems are typically powered by batteries that need to be recharged through difficult and expensive operations.\nFurthermore\, existing medical implants do not provide at once the miniaturized end-to-end sensing-computation-communication-recharging capabilities to implement Implantable Internet of Medical Things (IIoMT) applications.\nThis dissertation fills the existing research gaps by presenting innovative designs of battery-less devices remotely rechargeable through ultrasonic wireless power transfer. Specifically\, two major systems are presented\, U-Verse – the first FDA-compliant IIoMT platform packing sensing\, computation\, communication\, and recharging circuits into a penny-scale platform – and the first IoUT battery-less sensor node that can be wirelessly recharged through ultrasonic waves.\nU-Verse uses a single miniaturized transducer for data exchange and for wireless charging. To predict U-Verse’s performance\, a mathematical model of its charging efficiency is derived and experimentally validated. A matching circuit to maximize the amount of power transferred from the outside is proposed\, and the design of a full-fledged cm-scale printed circuit board (PCB) is presented. Extensive experimental evaluation indicates that U-Verse (i) is able to recharge a 330mF and 15F energy storage unit – several orders of magnitude higher than existing work – respectively under 20 and 60 minutes at a depth of 5cm; (ii) achieves stored charge duration of up to 610 and 40 hours in case of battery and supercapacitor energy storage\, respectively. Finally\, U-Verse is demonstrated through (i) a closed-loop application where a periodic sensing/actuation task sends data via ultrasounds through real porcine meat; and (ii) a real-time reconfigurable pacemaker. As for the underwater sensor node\, the architecture of an underwater platform capable of extracting electrical energy from ultrasonic waves is first introduced. Then\, the interfacing of the system with an underwater communication unit is illustrated. The design of a prototype where the storage unit is realized with a batch of supercapacitors is also discussed. Experimental results show that the harvested energy is sufficient to provide the sensor node with the power necessary to perform a sensing operation and power a modem for ultrasonic communications. Given the reduced attenuation of ultrasonic waves in water\, the proposed approach proves to cover longer distances with less transmission power than alternative solutions. Last\, the overall operating efficiency of the system is evaluated.
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-raffaele-guida/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201207T160000
DTEND;TZID=America/New_York:20201207T170000
DTSTAMP:20260428T221742
CREATED:20201130T195726Z
LAST-MODIFIED:20201130T195726Z
UID:4600-1607356800-1607360400@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Trinayan Baruah
DESCRIPTION:PhD Proposal Review: Improving the Virtual Memory Efficiency of GPUs \nTrinayan Baruah \nLocation: Zoom Link \nAbstract: GPUs have been adopted widely based their ability to exploit data-level parallelism found in modern-day applications\, ranging from high performance computing to machine learning. This widespread adoption has\, in part\, been accelerated by the development of more intuitive high-level programming languages\, efficient runtimes and drivers\, and easier mechanisms to manage data movement. Modern day GPUs and multi-GPU systems utilize virtual memory systems\, enabling programmers to access large address spaces that are beyond the physical memory limits a GPU. There mechanisms have built in mechanisms for memory translation\, sparing the programmer from having to reason about complex data-movement operations. Virtual memory support on a GPU includes both hardware and software support. At the hardware level\, Translation Lookaside Buffers (TLBs) are used to cache translations close to the compute units. At the software level\, the programming model supports a unified memory model which automates the movement of pages across multiple devices in a a system. Despite the improvements in programmability\, due to the inefficiency in existing TLB mechanisms for TLB management and page migration\, the performance of current virtual memory support on GPUs is sub-optimal.\nIn this dissertation\, we first identify the key challenges in virtual memory support for GPUs today. We then propose mechanisms to reduce the bottlenecks arising from virtual memory management at both a hardware level and at the runtime level. This allows GPUs to fully enjoy the benefits of virtual memory\, while ensuring high performance. We also develop simulation tools that enable researchers to explore new and novel virtual memory features in future single GPU and multi-GPU systems.\nTo enhance hardware support for virtual memory on a GPU\, we explore a mechanism that enables prefetching of page-table entries into the GPUs TLBs\, thereby reducing the number of TLB misses and improving performance. We also leverage the fact that many page-table entries can be shared across different GPU cores. We design a low-cost interconnect that enables sharing of page-table entries across the GPU cores. To improve the performance of unified memory on multi-GPU systems\, we propose a hardware/software mechanism that monitors accesses to each page\, and uses this information when making page-migration decisions. We also propose mechanisms to reduce the cost of TLB shootdowns on the GPU during page-migration in NUMA multi-GPU systems. \n 
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-trinayan-baruah/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201204T150000
DTEND;TZID=America/New_York:20201204T160000
DTSTAMP:20260428T221742
CREATED:20201123T205204Z
LAST-MODIFIED:20201123T205204Z
UID:4593-1607094000-1607097600@ece.northeastern.edu
SUMMARY:ECE MS Thesis Defense: Yuxuan Cai
DESCRIPTION:MS Thesis Defense: Real-Time Object Detection on Mobile Devices via Compression-Compilation Co-Design \nYuxuan Cai \nLocation: Zoom Link \nAbstract: The rapid development and wide utilization of object detection techniques have aroused attention on both accuracy and speed of object detectors. However\, the current state-of-the- art object detection works are either accuracy-oriented using a large model but leading to high latency or speed-oriented using a lightweight model but sacrificing accuracy. In this work\, we propose YOLObile framework\, a real-time object detection on mobile devices via compression compilation co-design. A novel block-punched pruning scheme is proposed for any kernel size. To improve computational efficiency on mobile devices\, a GPU-CPU collaborative scheme is adopted along with advanced compiler-assisted optimizations. Experimental results indicate that our pruning scheme achieves 14× compression rate of YOLOv4 with 49.0 mAP. Under our YOLObile framework\, we achieve 17 FPS inference speed using GPU on Samsung Galaxy S20. By incorporating our proposed GPU-CPU collaborative scheme\, the inference speed is increased to 19.1 FPS\, and outperforms the original YOLOv4 by 5× speedup.
URL:https://ece.northeastern.edu/event/ece-ms-thesis-defense-yuxuan-cai/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201202T140000
DTEND;TZID=America/New_York:20201202T150000
DTSTAMP:20260428T221742
CREATED:20201130T201126Z
LAST-MODIFIED:20201130T201212Z
UID:4601-1606917600-1606921200@ece.northeastern.edu
SUMMARY:ECE MS Thesis Defense: Kathan Vyas
DESCRIPTION:MS Thesis Defense: Data-Efficient analysis of Human Behavior by Spatio-Temporal Pose Generation and Inference \nKathan Vyas \nLocation: Zoom Link  \nPasscode: 474462 \nAbstract: Identifying human pose over time provides critical information towards understanding human behavior and their physical interaction with the environment surrounding them. In the past few decades\, the human pose estimation topic has witnessed groundbreaking research in the computer vision field thanks to the powerful deep learning models. These models are trained using several thousands of labeled sample images if not more. Such extensive data requirement posed a fundamental problem for domains (i.e. Small Data domains)\, in which data collection or labeling is expensive or limited due to privacy or security concerns such as healthcare. In this thesis\, we present a data-efficient learning pipeline to address small data problem in a healthcare-related human pose estimation application. In particular\, we infer spatio-temporal human poses to analyze typical vs. atypical behaviors in children with Autism spectrum disorder (ASD). To mitigate data limitation\, we propose two thrusts in our learning pipeline. The first thrust is a data-efficient machine learning approach\, in which a pre-trained (on adult pose images) pose estimation model with deep structure is fine-tuned on a small set of children pose videos\, provided to us by our collaborators. We implement a non-linear particle filter interpolation to deal with any missing body keypoints in the estimated poses and employ a novel PoTion (pose motion) based temporal aggregation technique to evaluate poses over time. The second thrust is a synthetic data augmentation approach\, in which we build a framework to create synthetic 3D humans with articulated bodies in order to render more pose images/videos in our application contexts. We use a novel 3D registration approach based on RANSAC and implement iterative closest point (ICP) to obtain 3D meshes from the scanned point clouds from both adult and kid mannequins\, which is then rigged and articulated in the Blender to generate our human avatars. We then infuse these avatars in various synthetic environments to create contexts similar to the target application\, which is a kid with both typical and atypical behaviors in a home-like environment.
URL:https://ece.northeastern.edu/event/ece-ms-thesis-defense-kathan-vyas/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201202T100000
DTEND;TZID=America/New_York:20201202T110000
DTSTAMP:20260428T221742
CREATED:20201119T022728Z
LAST-MODIFIED:20201119T022728Z
UID:4580-1606903200-1606906800@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Bilgehan Donmez
DESCRIPTION:PhD Dissertation Defense: Topology Error Detection in Power System State Estimation \nBilgehan Donmez \nLocation: Teams Link \nAbstract: Growth of renewable energy\, changes in weather patterns\, and increases in cyber- and physical-attacks are examples of recent challenges in power system operation. To keep up with these rapid transformations\, it is imperative to improve the tools used in modern-day control centers.\nAs the centerpiece of system operations\, improvements in state estimation (SE) accuracy would result in better situational awareness for system operators. The state estimate can often be compromised when there are errors in the assumed topology of the network. Therefore\, topology error detection plays a key role in SE. In the first part of this dissertation\, topology errors in the external systems\, which are the neighboring control areas\, are investigated. When a subset of measurements coming from an external area is lost\, some parts of the system can become unobservable. Since SE cannot be carried out for the unobservable portion of the system\, the topology of the external system cannot be tracked in its usual way. This dissertation offers a computationally efficient external line outage detection algorithm that uses only the internal bus phase angles\, any available phasor measurement units (PMUs)\, and the pre-contingency system topology of the system. Coupled with a post-verification step\, this method is shown to be effective in detecting external line outages.\nThe second part of the dissertation focuses on topology errors in the internal system. The conventional SE implementations use the simplified bus-branch (BB) electrical network provided by the topology processor (TP). When the status of circuit breakers are not reported correctly to the TP\, the electrical equivalent it creates will be inaccurate. Therefore\, topology errors usually result in SE convergence problems or yield significantly biased estimates. To properly detect these types of errors\, rather than using the typical BB representation\, the network model is expanded to include circuit breakers and other switching devices in substations. SE is then reformulated to work with this detailed node-breaker (NB) model.\nAlthough the expansion of the model introduces operational and computational challenges\, several strategies are employed to counter these issues. The proposed innovations include the formulations of two separate equality-constrained SE algorithms\, the development of optimal meter placement algorithms\, and utilization of parallel processing. As demonstrated through the simulations conducted\, the methods developed in this dissertation are practical enough for adaptation to real-world systems.
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-bilgehan-donmez/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201202T100000
DTEND;TZID=America/New_York:20201202T110000
DTSTAMP:20260428T221742
CREATED:20201117T013509Z
LAST-MODIFIED:20201117T013509Z
UID:4575-1606903200-1606906800@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Leili Hayati
DESCRIPTION:PhD Dissertation Defense: Ceramic Magnetic Wires at Wireless Communication Frequencies \nLeili Hayati \nLocation: Online \nAbstract: Ferrite magnetic devices play an important role in modern wireless telecommunication systems. They generally require permanent magnets in order to magnetically polarize the ferrite material component used in these devices. The permanent magnets are bulky and take up most of the size and weight of a magnetic circuit. The aim of this research is to do away with permanent magnet bias circuits as utilized in circulators and ferrite planar devices\, especially in wireless communication systems operating below 2 GHz. Recently\, ferromagnetic nanowires (NWs) have been embedded into porous templates\, are used to design various microwave magnetic and electronics devices. The main advantage of magnetic NWs is that in zero magnetic field\, the microwave absorption frequency can be easily tuned over a large range of frequencies. Clearly\, the metallic nature of the magnetic NWs contributed to the high loss. It is expected that insulating magnetic NWs will improve the insertion loss sufficiently to produce viable ferrite devices at wireless communication frequencies below 2GHz and at higher frequencies. There are no pure insulating magnetic materials. However\, there are ferrites that are nearly insulating and are ferrimagnetic. Their saturation magnetization is much lower than the metallic ferromagnetic counterpart. This is a desirable property for magnetic device operating below 2 GHz. Of all the ferrite materials yttrium iron garnet (YIG) exhibits the lowest FMR linewidth ever measured and low saturation magnetization. In this work\, an array of high-purity YIG NWs embedded in a porous silicon membrane\, were synthesized using sol-gel method and the magnetic properties of the pure YIG Nanoparticles and the composite substrate were characterized by utilizing vibrating sample magnetometer (VSM) technique. From the ferromagnetic resonance (FMR) spectra\, it has been found that the measurements are characterized by a uniaxial magnetic anisotropy energy due to the high aspect ratio of the NWs. Based on the magnetic parameters of the composite substrate and characterizing YIG NWs\, a coplanar waveguide was designed by HFSS software. By applying a small external magnetic field and changing the internal magnetic H field by ±8%\, the phase of S21 parameter shifts up to 30̊ degrees near 1.7GHz.
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-leili-hayati/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201130T130000
DTEND;TZID=America/New_York:20201130T140000
DTSTAMP:20260428T221742
CREATED:20201123T204938Z
LAST-MODIFIED:20201123T204938Z
UID:4592-1606741200-1606744800@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Berkan Kadioglu
DESCRIPTION:PhD Proposal Review: Sample Complexity of Pairwise Ranking Regression \nBerkan Kadioglu \nLocation: Zoom \nAbstract: We consider a rank regression setting\, in which a dataset of $N$ samples with features in $\mathbb{R}^d$ is ranked by an oracle via $M$ pairwise comparisons.\nSpecifically\, there exists a latent total ordering of the samples; when presented with a pair of samples\, a noisy oracle identifies the one ranked higher w.r.t. the underlying total ordering. A learner observes a dataset of such comparisons\, and wishes to regress sample ranks from their features.\nWe show that to learn the model parameters with $\epsilon > 0$ accuracy\, it suffices to conduct $M \in \Omega(dN\log^3 N/\epsilon^2)$ comparisons uniformly at random when $N$ is $\Omega(d/\epsilon^2)$. \n 
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-berkan-kadioglu/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201130T093000
DTEND;TZID=America/New_York:20201130T103000
DTSTAMP:20260428T221742
CREATED:20201121T024753Z
LAST-MODIFIED:20201123T205506Z
UID:4591-1606728600-1606732200@ece.northeastern.edu
SUMMARY:ECE MS Thesis Defense: Sila Deniz Calisgan
DESCRIPTION:MS Thesis Defense: MEMS Infrared Resonant Detectors With Near-Zero Power Readout For Miniaturized Low Power Systems \nSila Deniz Calisgan \nLocation: Online \nAbstract: The demand for low-cost and low-power microsystems for spectrally-selective IR sensing has been rising with the proliferation of Internet of Things (IoT) for applications such as security surveillance and natural disaster monitoring. As a result\, there is a need for low-power\, high sensitivity IR sensors with minimum deployment and maintenance cost that can detect trace levels of chemicals. This thesis reports on the first experimental demonstrations of passive integrated microsystems based on transmission spectroscopy using narrowband uncooled microelectromechanical resonant infrared (IR) detectors. Moreover\, the MEMS-CMOS integrated microsystem can turn itself ON to quantify the intensity of infrared radiation when an above-threshold IR signature is present\, but otherwise remain dormant with near-zero standby power consumption. The proposed sensor system combines the unique advantage of two recently developed technologies\, namely\, the zero-power nature of micromechanical photoswitches (MPs) and the high resolution of aluminum nitride (AlN) MEMS resonant infrared detectors\, to achieve an unprecedented IR sensing capability. Thanks to the spectral selectivity enabled by the plasmonically enhanced thermo-mechanical transduction in MEMS structures\, the proposed sensor system is capable of discriminating the spectral content of incoming IR radiation for the identification of events of interest. The prototype presented here is automatically powered up by the MP when the incoming IR radiation exceeds 440 nW showing a high IR detection resolution in active state and a near-zero power consumption (~3 nW) in standby. The ultrathin plasmonic absorber with narrow bandwidth (FWHM<17% ) and near-perfect IR absorption (η>92%) coupled with the high IR detection capability ( NEP~ 463 pW/√Hz) of the AlN resonator was exploited for a filter-free spectroscopic chemical sensor based on uncooled AlN resonant IR detectors with a minimum concentration detection limit of <0.01% (Benzonitrile in Hexane).
URL:https://ece.northeastern.edu/event/ece-ms-thesis-defense-sila-deniz-calisgan/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201125T120000
DTEND;TZID=America/New_York:20201125T130000
DTSTAMP:20260428T221742
CREATED:20201112T213551Z
LAST-MODIFIED:20201112T213551Z
UID:4572-1606305600-1606309200@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Aykut Onol
DESCRIPTION:PhD Dissertation Defense: Planning of Contact-Interaction Trajectories Using Numerical Optimization \nAykut Onol \nLocation: Zoom Link \nAbstract: Dynamic multi-contact behaviors\, such as locomotion and item manipulation\, remain to be a challenge for today’s robotic systems. This is primarily due to the discontinuous and non-smooth dynamics introduced by contacts. For mobile manipulators (e.g.\, humanoids) to become useful for dangerous\, dirty\, and dull tasks\, such as those in disaster response\, they need to be capable of interacting with their cluttered\, constrained\, and changing environments. It is therefore essential to develop methods that would enable robots to plan and execute contact-rich motions in dynamic surroundings.\nIn this dissertation research\, we investigate the planning of contact-interaction trajectories and utilize numerical optimal control techniques to solve this problem in a generalizable and computationally-tractable way. We develop a contact-implicit trajectory optimization framework for the automatic discovery of dynamic contact-rich behaviors given only a high-level goal\, i.e.\, the desired configuration of the environment. A variable smooth contact model is introduced to improve the convergence of gradient-based optimization without compromising the physical fidelity of resulting motions. This is achieved by employing smooth virtual forces that act as a decoupled relaxation of the rigid-body contact model. Second\, we develop a sequential convex optimization procedure that provides reliable convergence characteristics while solving this non-convex problem. Third\, a penalty loop approach is proposed to generalize this method to a wide range of robotic applications.\nIn addition to these\, we develop a novel Coulomb friction model and an on-the-fly contact constraint activation method using state-triggered constraints\, STCs. STCs are a more modular alternative to complementarity constraints which are widely used to model discrete aspects in contact-related problems. Our extensive simulation experiments demonstrate that STCs hold immense promise to efficiently model a broad range of discrete elements in the planning and control of contact-interaction trajectories. As a result\, this dissertation presents methods that enable the planning of dynamic contact-rich behaviors for different robots and tasks without requiring any parameter tuning or tailored initial guess.
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-aykut-onol/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201124T140000
DTEND;TZID=America/New_York:20201124T150000
DTSTAMP:20260428T221742
CREATED:20201103T210959Z
LAST-MODIFIED:20201103T210959Z
UID:4552-1606226400-1606230000@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Joseph Robinson
DESCRIPTION:PhD Dissertation Defense: Automatic Face Understanding: Recognizing Families in Photos \nJoseph Robinson \nLocation: Zoom Link \nAbstract: Visual kinship recognition has an abundance of practical uses. For this\, we built the largest database for kinship recognition\, FIW. Built entirely in-house with no cost using a semi-automatic labeling scheme. Specifically\, we first aligned faces detected in family photos with names in the corresponding text metadata to mine the label proposals with high confidence. The remaining data were labeled using a novel clustering algorithm that used label proposals as side information to guide more accurate clusters. Great savings in time and human input was had. Statistically\, FIW shows enormous gains over its predecessors. We have several benchmarks in kinship verification\, family classification\, tri-subject verification\, and large-scale search & retrieval. We also trained CNNs on FIW and deployed the model on the renowned KinWild I and II to gain state-of-the-art (SOTA). Most recently\, we further augmented FIW with multimedia (MM) for 200 of its 1\,000 families- a labeled collection we dubbed FIW-MM. Now\, video dynamics\, audio\, and text captions can be used in the decision making of kinship recognition systems. \nFIW continues to pave the way for this research track: (1) advanced SOTA (e.g.\, marginalized denoising auto-encoder based on metric learning that preserves intrinsic structures of kin-data and encapsulates discriminating information in learned features); (2) introduced generative models to predict a child’s appearance from a parent pair (i.e.\, proposed an adversarial autoencoder conditioned on age and gender to map between facial appearance and these higher-level features for control of age and gender); (3) designed evaluations with benchmarks to support challenges\, workshops\, and tutorials at top tier conferences (e.g.\, CVPR\, MM\, FG\, ICME)\, and a premiere Kaggle Competition. We expect FIW will significantly impact research and reality. \nAdditionally\, we tackled the classic problem of facial landmark localization in images. This is a task that has been in focus for decades\, and many solutions have been proposed. However\, there are revamped interests in pushing facial landmark detection technologies to handle more challenging data with deep networks now prevailing throughout machine learning. A majority of these networks have objectives based on L1 or L2 norms\, which inherit several disadvantages. First of all\, the locations of landmarks are determined from generated heatmaps (i.e.\, confidence maps) from which predicted landmark locations (i.e.\, the means) get penalized without accounting for the spread: a high scatter corresponds to low confidence and vice-versa. To address this\, we introduced a LaplaceKL objective that penalizes for low confidence. Another issue is a dependency on labeled data\, which is expensive to collect and susceptible to error. We addressed both issues by proposing an adversarial training framework that leverages unlabeled data to improve model performance. Our method claims SOTA on renowned benchmarks. Furthermore\, our model is robust with a reduced size: 1/8 the number of channels (i.e.\, 0.0398 MB) is comparable to state-of-that-art in real-time on a CPU. Thus\, our method is of high practical value to real-life applications. \nFinally\, we built the Balanced Faces in the Wild (BFW) dataset to serve as a proxy to measure bias across ethnicity and gender subgroups\, allowing us to characterize FR performances per subgroup. We show performances are non-optimal when a single score threshold is used to determine whether sample pairs are genuine or imposter. Furthermore\, actual performance ratings vary greatly from the reported across subgroups. Thus\, claims of specific error rates only hold for populations matching that of the validation data. We mitigate the imbalanced performances using a novel domain adaptation learning scheme on the facial encodings extracted using SOTA deep nets. Not only does this technique balance performance\, but it also boosts the overall performance. A benefit of the proposed is to preserve identity information in facial features while removing demographic knowledge in the lower dimensional features. The removal of demographic knowledge prevents future potential biases from being injected into decision making. Additionally\, privacy concerns are satisfied by this removal. We explore why this works qualitatively with hard samples. We also show quantitatively that subgroup classifiers can no longer learn from the encodings mapped by the proposed. \n 
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-joseph-robinson/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201113T140000
DTEND;TZID=America/New_York:20201113T150000
DTSTAMP:20260428T221742
CREATED:20201110T024923Z
LAST-MODIFIED:20201110T024923Z
UID:4570-1605276000-1605279600@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Yumin Liu
DESCRIPTION:PhD Proposal Review: Learning from Spatio-Temporal Data with Applications in Climate Science \nYumin Liu \nLocation: Zoom Link \nAbstract:Climate change is one of the major challenges to human beings and many other species in our time. In the recent decade\, the number of disasters related to climate change such as wildfires\, storms\, floods and droughts are increasing\, and the casualty and economic losses caused by them are larger compared to those of decades ago. This calls for better and efficient ways to predict climate change in order to better prepare and reduce losses. Predicting climate change involves using historical observational data and model simulated data\, both of which usually involve multiple locations and timestamps and are spatio-temporal. With the rapid development and progress of machine learning\, these methods have achieved several impactful contributions in many domains; we would like to translate its impact to climate science.\nIn this thesis we addressseveral problems in climate science. This challenging complex domain enable us to develop\, innovate\, adapt\, and advance machine learning in the following ways. 1) We develop a multi-task learning method to estimate relationships between tasks and learn the basis tasks in different locations especially for nearby locations which may share similar climate patterns. This method assumes that the weights of an observed task is a linear combination of several latent basis tasks and that the task relationships can be learnt by imposing a regularized precision matrix. 2) We propose a nonparameteric mixture of sparse linear regression models to cluster and identify important climate models for prediction. This model incorporates Dirichlet Process (DP) to automatically determine the number of clusters\, imposes Markov Random Field (MRF) constraints to guarantee spatio-temporal smoothness\, and selects a subset of global climate models (GCMs) that are useful for prediction within each spatio-temporal cluster with a spike-and-slab prior. We derive an effective Gibbs sampling method for this model. 3) We adapt image super resolution method to climate downscaling — increasing spatial resolution for climate variables for local impact analysis. The proposed method is called YNet which is a novel deep convolutional neural network (CNN) with skip connections and fusion capabilities to perform downscaling for climate variables on multiple GCMs directly rather than on reanalysis data.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-yumin-liu/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201112T130000
DTEND;TZID=America/New_York:20201112T140000
DTSTAMP:20260428T221742
CREATED:20201103T200924Z
LAST-MODIFIED:20201103T200924Z
UID:4546-1605186000-1605189600@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Yaoshen Yuan
DESCRIPTION:PhD Proposal Review: Enhancements for Monte Carlo Light Modeling Method and Its Applications in Near-infrared-based Brain Techniques \nYaoshen Yuan \nLocation: Microsoft Teams Link \nAbstract: Studying light propagation in biological tissues is critical for developing biophotonics techniques and its applications. Monte Carlo (MC) method\, a stochastic solver for radiative transfer equation\, has been recognized as the gold standard for modeling light propagation in turbid media. However\, due to the stochastic nature of MC method\, millions even billions of photons are usually required to achieve accurate results using MC method\, leading to a long computational time even with the acceleration using graphical processing units (GPU).\nFurthermore\, due to the rapid advances in multi-scale optical imaging techniques such as optical coherence tomography (OCT) and multiphoton microscopy (MPM)\, there is an increasing need to model light propagation in extremely complex tissues such as vessel networks. The mesh-based Monte Carlo (MMC) is usually superior than the voxel-based MC method for such modeling since unlike grid-like voxels\, tetrahedral meshes can represent arbitrary shapes with curved boundaries. However\, the mesh density can be excessively high when the tissue structure is extremely complex\, resulting in high computational costs and memory demand. The goal of this proposal is to focus on solving the challenges mentioned above. \nTo tackle the first challenge\, we came up with a filtering approach with GPU acceleration to improve the signal-to-noise ratio (SNR) of the results while keeping the simulated photons low. The adaptive non-local means (ANLM) filter is selected to suppress the stochastic noise in MC results because 1) the filtering process on each voxel is mutually independent\, making it possible for parallel computing; 2) it has high performance in denoising and a strong capacity in edge-preserving. For the second problem\, a novel method\, implicit mesh-based Monte Carlo (iMMC)\, was proposed to significantly reduce the mesh density. The iMMC utilizes the edge\, node and face of the tetrahedral mesh to model tissue structures with shapes of cylinder\, sphere and thin layer. The typical applications for edge\, node and face-based iMMC are vessel networks\, porous media and membranes\, respectively. Lastly\, we applied MC simulations and aforementioned filter on segmented brain models derived from MRI neurodevelopmental atlas to estimate the light dosage for transcranial photobiomodulation (t-PBM)\, a technique for treating major depressive disorder using near infrared\, across lifespan. Furthermore\, a new approach that can improve the penetration depth in optical brain imaging as well as PBM is proposed. In this approach\, the possibility of placing light sources in head cavities is investigated using MC simulations. The preliminary results demonstrate a better performance in deep brain monitoring compared to the standard transcranial approach using 10-20 EEG positioning system. \n 
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-yaoshen-yuan/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201109T150000
DTEND;TZID=America/New_York:20201109T160000
DTSTAMP:20260428T221742
CREATED:20201105T220658Z
LAST-MODIFIED:20201105T220658Z
UID:4557-1604934000-1604937600@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Carlos Bocanegra
DESCRIPTION:PhD Proposal Review: A Systems Approach to Spectrum Sharing and Multi-antenna Operation for Emerging Networks \nCarlos Bocanegra \nLocation: TBD \nAbstract: The demands on wireless connectivity and sensing are ever-increasing\, fueled by both emerging applications and an exponential growth in the number of connected devices. Availability of new spectrum in the sub-6GHz bands is limited\, which motivates research on innovative ways to utilize the current available spectrum and explore the use of additional spectrum beyond the 6GHz threshold.\nThis thesis explores three promising techniques with focus on the Physical (L1) and Link (L2) layers. Approach 1 concerns spectrum sharing in the sub-6GHz band\, where wireless standards are granted opportunistic access within unlicensed spectrum to increase their usable bandwidth. Approach 2 concerns the design of massive multi-antenna systems\, through which devices can benefit from beamforming gains at transmission and diversity gains at reception. Approach 3 concerns the use of very high frequency bands (VHFB)\, or so called millimeter-wave bands. Each of these approaches\, however\, has its own set of challenges\, such as fairness in channel access\, interference management\, and optimal beamforming user-specified quality-of-service\, respectively.\nFor spectrum sharing as described in Approach 1\, this thesis presents E-Fi\, an interference-evasion mechanism that allows Wi-Fi devices to survive opportunistic in-band LTE transmissions. The main contribution is to achieve this without any cooperation between these two\, using Almost Blank Subframes (ABS). E-Fi ensures fair channel access while reusing existing Wi-Fi standards\, i.e.\, Wi-Fi Direct\, and thus incurring minimal deployment costs.\nFor Approach 2\, this thesis introduces two multi-antenna frameworks\, a decentralized one for cellular- and a centralized one for IoT-oriented applications\, respectively. For the former\, it presents NetBeam\, a reconfigurable system of distributed 3D beamformers (3DBF). While NetBeam uses 3DBF to tackle multi-user interference in 3D multi-user deployments\, it enforces Machine Learning and efficient antenna selection strategies to deliver the individual required SINR levels to users. As a centralized multi-antenna system\, it presents RFGo\, a privacy-preserving self-checkout system using passive Radio Frequency ID (RFID) tags. RFGo achieves fast tag discovery using a custom-built RFID reader\, which simultaneously decodes a tag’s response from multiple carrier-level synchronized antennas. RFGo achieves reliable tag detection by means of a neural network that accurately discriminates products within the checkout area from those laying outside of it.\nIn the proposed work that covers Approach 3\, this thesis describes an outline of an algorithmic framework for millimeter-wave communications that efficiently allocates antenna elements from Base Stations (BS) to users for hybrid beamforming\, while considering their individual traffic demands. We propose to trade-off flexible array geometries (that allows to limit interference to specific regions) versus the irregularity that results in the sidelobes.\nIn summary\, this thesis tackles complex challenges in the future 5G and beyond wireless networks through a combination theory\, algorithm design and experimental implementation.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-carlos-bocanegra/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201105T120000
DTEND;TZID=America/New_York:20201105T130000
DTSTAMP:20260428T221742
CREATED:20201023T211519Z
LAST-MODIFIED:20201023T211519Z
UID:4533-1604577600-1604581200@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Aziz Kocanaogullari
DESCRIPTION:PhD Dissertation Defense: Active Recursive Bayesian Classification (Querying and Stopping) for Event Related Potential Driven Brain Computer Interface Systems \nAziz Kocanaogullari \nLocation: Remote (contact akocanaogullari@ece.neu.edu) \nAbstract: Recursive Bayesian classification (RBC) requires optimal latent variable estimation in the presence of noisy observation to achieve real-time sequential decision making. Active RBC introduced in this dissertation attempts to effectively select queries that lead to more informative observations to rapidly reduce uncertainty until a confident decision is made. Accordingly\, active RBC includes the following fundamental components:(S)A stopping criterion based on the posterior probability to stop evidence collection;(Q)a querying step to decide how to collect further evidence from relevant sources to benefit speed and accuracy objectives of RBC;(C)a classification objective based on the posterior distribution and loss values attributed to each true label and decision option pair to determine the optimal decision once the stopping criterion has been satisfied.\nThis dissertation specifically focuses on optimizing querying (Q) and stopping (S) for RBC. Conventional stopping criterion design methodologies lack insight of the RBC geometry and evolution of the posterior probability vector. Additionally\, conventional active querying methods stagger due to misleading prior information. In this case\, the system uses time inefficiently to overcome the provided belief by querying most likely candidates a number of iterations. Furthermore\, in contrast to inference and querying being coactive\, typically the optimality objectives are designed separately.\nAn electroencephalography (EEG)-based brain computer interface (BCI) system specifically de-signed for typing is used as a testbed for active RBC. BCI systems provide a communication pathway between the user and the environment both in medical and non-medical domains. EEG signals are widely used with promising performance to estimate user intent in BCI systems. BCI typing systems are epitomes of RBC driven systems as repeated evidence collection is mandated due to highly variable EEG signals given a particular user intent (latent variable hidden in the brain). However\, in many cases\, EEG-based communication staggers and lacks accuracy and speed due to inefficient RBC.\nTo increase the performance of RBC\, motivated by information theoretic approaches to coding and active learning this dissertation contributes to the literature in three folds: (i) A complete analysis of stopping criterion and geometrical description of the RBC problem is provided. Motivated by the posterior motion a stopping criterion design is proposed. Moreover\, an early stopping scheme with one step ahead prediction is shown to make a decision with marginal accuracy deficit. (ii)Influenced by the posterior motion\, a new query selection objective is proposed. This querying mechanism is shown to result in rapid and accurate inference in scenarios in which the recursive inference starts with a misleading (or adversarial) prior probability distribution for the latent variable of interest (e.g. user attempting to type a letter/word that is unlikely according to the language model). (iii) Querying and stopping approaches are taken together into consideration and an experimental study specifically on BCI typing is presented. Additionally\, the dissertation shows it is possible to reformulate RBC with Rényi entropy measures solidifying the connection between stopping and querying objective design. All contributions are verified using a BCI typing system “BCIPy” with simulations and human-in-the-loop experiments.
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-aziz-kocanaogullari/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201102T090000
DTEND;TZID=America/New_York:20201102T100000
DTSTAMP:20260428T221742
CREATED:20201026T214743Z
LAST-MODIFIED:20201026T214743Z
UID:4537-1604307600-1604311200@ece.northeastern.edu
SUMMARY:ECE PhD Proposal Review: Ya Guo
DESCRIPTION:PhD Proposal Review: Power Optimization and Management of PV Grid-Connected Microgrid in Energy Market \nYa Guo \nLocation: Zoom Link \nAbstract: Microgrids can integrate renewable energy resources (RES)\, such as photovoltaics (PV) and wind energy generation\, with the main power grid to provide reliable\, secure and affordable energy. Fortunately\, the electricity markets have evolved to facilitate RES participation. One major challenge lies in how to manage power and energy flow within grid-connected microgrid system\, to optimize financial gains while maintaining high reliability. This becomes challenging since electricity trading policies and tariffs vary by utility companies from area to area. Furthermore\, RES are mostly intermittent sources. Adding additional energy storage systems (ESS) into microgrids becomes a vital solution to mitigating the energy production intermittency\, as well as providing energy backup in emergency. Battery ESS (BESS) are deployed on a large scale in grid-connected installations worldwide. Optimal operation of the energy storage system also becomes important for microgrid end-users to ensure that they will at least recover BESS operating cost. Moreover\, there always exist uncertainties in RES power generation\, load power consumption\, and even dynamic electricity pricing. It is vital to deal with the forecasted errors in real-time. Developing proper uncertainty characterization can better facilitate the whole system power management to limit the negative influences of these uncertainties.\nIn this research\, dynamic programming (DP) algorithm is proposed to forecast the global optimal solution to power flow dispatch of PV grid-connected microgrid. Various electricity pricing structures\, including fixed pricing\, time-of-use (TOU) pricing and real-time pricing (RTP) are explored for customers in different areas. The battery nonlinear charging/discharging degradation model is also exploited for system power optimization. The objective is to achieve the minimum microgrid system operation cost\, in other words\, the maximum economic benefits for end-users. Besides\, this research proposes power control methods to implement forecasted optimal power schedule\, as well as dealing with errors among forecast and real-time PV\, load and RTP. Rule-based (RB) algorithm is also studied as a baseline for comparison. Moreover\, uncertainty characterization for PV\, load and dynamic pricing will be developed using Monte Carlo Simulation (MCS)\, and stochastic optimization approach will be explored in cooperation with these uncertainties.
URL:https://ece.northeastern.edu/event/ece-phd-proposal-review-ya-guo/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20201030T110000
DTEND;TZID=America/New_York:20201030T120000
DTSTAMP:20260428T221742
CREATED:20201024T021519Z
LAST-MODIFIED:20201024T021519Z
UID:4534-1604055600-1604059200@ece.northeastern.edu
SUMMARY:ECE PhD Dissertation Defense: Ran Liu
DESCRIPTION:PhD Dissertation Defense: Optimal Proactive Services with Uncertain Predictions \nRan Liu \nLocation: Zoom Link \nAbstract: With the evolution of technologies such as machine learning and data science\, proactive services with the aid of predictive information have been recognized as a promising method to exploit network bandwidth\, storage\, and computation resources to achieve improved user experiences\, especially delay performance.\nSpecifically\, services can be processed proactively when the system is lightly loaded\, with the results stored to meet user demand in the future.\nOur primary goal in the thesis is to investigate the fundamental performance improvement that can be achieved from proactive services under uncertain predictions. We aim to analyze the queueing behavior of proactive systems under certain proactive strategies and characterize the improvement in terms of the limiting fraction of proactive work and the limiting average delay. \nIn the first work\, we analytically investigate the problem of how to efficiently utilize uncertain predictive information to design proactive caching strategies with provably good access-delay characteristics.\nFirst\, we derive an upper bound for the average amount of proactive service per request that the system can support.\nThen we analyze the behavior of a family of threshold-based proactive strategies with a Markov chain\, which shows that the average amount of proactive service per request can be maximized by properly selecting the threshold.\nFinally\, we propose the UNIFORM strategy\, which is the threshold-based strategy with the optimal threshold\, and show that it outperforms the commonly used Earliest-Deadline-First (EDF) type proactive strategies in terms of delay.\nWe perform extensive numerical experiments to demonstrate the influence of thresholds on delay performance under the threshold-based strategies\, and specifically\, compare the EDF strategy and the UNIFORM strategy to verify our results. \nIn the second work\, we study a more generalized proactive service problem with a more generalized service model and derive explicit solutions on the limiting average fraction of proactive work and the limiting average delay in closed-form expressions.\nIn this work\, we analytically investigate how to optimally take advantage of under-utilized network resources for proactive services with the aid of uncertain predictive information.\nSpecifically\, we first derive an upper bound on the fraction of services that can be completed proactively by a single-server system.\nThen we analyze a family of fixed-probability (FIXP) proactive strategies in two proactive systems\, namely the Genie-Aided system and the Realistic Proactive system.\nWe analyze the asymptotic behaviors of the FIXP strategies by modeling a Markov process and the corresponding embedded Markov Chain.\nWe obtain optimal FIXP strategies in both systems and prove that the optimal FIXP strategies maximize the limiting fraction of proactive service among all proactive strategies and minimize average delay among FIXP strategies.\nWe perform extensive numerical experiments to demonstrate the influence of the parameter of FIXP on the performance of the limiting fraction of proactive service and the limiting average delay in both proactive systems and verify our theoretical results in multiple scenarios.
URL:https://ece.northeastern.edu/event/ece-phd-dissertation-defense-ran-liu-2/
END:VEVENT
END:VCALENDAR