BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Northeastern University College of Engineering - ECPv6.15.20//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Northeastern University College of Engineering
X-ORIGINAL-URL:https://coe.northeastern.edu
X-WR-CALDESC:Events for Northeastern University College of 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: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
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20240310T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20241103T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231207T140000
DTEND;TZID=America/New_York:20231207T150000
DTSTAMP:20260520T023125
CREATED:20231204T190346Z
LAST-MODIFIED:20231204T190346Z
UID:40688-1701957600-1701961200@coe.northeastern.edu
SUMMARY:Yisi Liu MS Thesis Defense
DESCRIPTION:Title: Experimental research on the Nonlinear Magnetoelectric Effect of the VLF ME Antennas \nCommittee Members:\nProf. Nian Sun (Advisor)\nProf. Yongmin Liu\nProf. Xufeng Zhang \nAbstract:\nMagnetoelectric (ME) coupling effects in ferromagnetic and piezoelectric composites involve the control of electric polarization (P) by applying a magnetic field (H) (direct ME effect)\, or the manipulation of magnetization (M) through an electric field (E) (converse ME effect) . These effects are facilitated by the mechanical deformation in the ferroic phases resulting from the combination of magnetostriction and piezoelectricity. In single-phase materials\, the breakthrough in achieving large ME coefficients has further advanced the development of ME materials and devices. Consequently\, numerous multifunctional ME devices\, such as mechanical antennas\, magnetic sensors\, tunable inductors\, and filters\, have been developed. This thesis has provides a summary and categorization of these devices based on their physical mechanism and type of ME effects. The inclusion of mechanical ME antennas based on piezoelectric/magnetostrictive heterostructures with acoustic actuation reflects the significant interest in this topic. Notably\, a maximum communication distance of 120 m for a very low frequency (VLF) communication system has been achieved using a pair of mechanical ME antennas. Subsequently\, we will focus on introducing and reviewing the materials and devices related to the ME effect\, as well as the application of ME mechanical antennas in very low frequency (VLF) communication systems. \nIn addition to that\, we developed a transmitter with a Metglas/PZT/Metglas structure antenna. Our study focuses on investigating the transmission effects of this antenna when employing direct antenna modulation techniques to enhance data transmission. Through our research\, we have introduced a novel modulation method by modulating the antenna. We observed that this modulation method produces a more stable and stronger signal. \n 
URL:https://coe.northeastern.edu/event/yisi-liu-ms-thesis-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231207T100000
DTEND;TZID=America/New_York:20231207T110000
DTSTAMP:20260520T023125
CREATED:20231204T185514Z
LAST-MODIFIED:20231204T185514Z
UID:40692-1701943200-1701946800@coe.northeastern.edu
SUMMARY:Mauro Belgiovine PhD Proposal Review
DESCRIPTION:Title: Wireless Intelligence: A Comprehensive Exploration of AI-Driven Solutions in Channel Estimation\, Beam Refinement\, and Protocol Classification for Next Generation Networks \nCommittee Members:\nProf. Kaushik Chowdhury (advisor)\nProf. Stratis Ioannidis\nDr. Chris Dick \nAbstract:\nhis thesis explores the transformative impact of artificial intelligence (AI) on wireless systems through model-driven simulations and real-world datasets\, with a focus on enhancing both local and cellular wireless networks through the deployment of highly customized deep learning solutions that target specific bottlenecks affecting traditional signal processing based communication. \nThe research delves into three key areas that address critical challenges in the current wireless landscape. The first focal point of the investigation involves channel estimation using deep learning techniques to denoise pilots and expedite the accurate estimation of Channel State Information (CSI). By leveraging deep learning methodologies\, the proposed solution aims to enhance the reliability and computation for MIMO and massive MIMO channel estimation\, thereby contributing to improved communication efficiency and reduced errors. The second major topic encompasses the application of reinforcement learning for 5G New Radio (NR) millimeter-wave (mmWave) beam refinement. The study aims to develop a Deep Reinforcement Learning algorithm capable of adjusting beamsteering angles\, starting from a coarse beam scanning procedure and further refining them for higher transmission efficiency. This innovation is expected to substantially decrease traffic overhead while simultaneously enhancing beam steering precision\, thus optimizing the performance of mmWave communication. The third and final area of focus introduces a transformer-based WiFi multi-protocol classifier\, strategically deployed on a DeepWave Air-T edge device\, which is equipped with Module on Chip (MoC) low power CPU-GPU and programmable Software Defined Radio (SDR). This classifier outperforms existing modulation classification models and legacy methods under lower SNR conditions\, leveraging TensorRT’s model compression capabilities to efficiently process extended sequences of raw IQ samples\, ensuring high performance at a low computational cost. The proposed solution addresses the growing demand for efficient and adaptable wireless communication systems\, paving the way for advancements in edge-based processing and intelligent protocol classification. \nThis work seeks to contribute significantly to the ongoing AI revolution in wireless systems by addressing crucial issues in channel estimation\, beam refinement\, and protocol classification. The outcomes of this research hold the potential to redefine the landscape of wireless communication\, offering enhanced performance\, reduced overhead\, and increased adaptability in both local and cellular networks.
URL:https://coe.northeastern.edu/event/mauro-belgiovine-phd-proposal-review/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231206T150000
DTEND;TZID=America/New_York:20231206T170000
DTSTAMP:20260520T023125
CREATED:20231204T185947Z
LAST-MODIFIED:20231204T185947Z
UID:40690-1701874800-1701882000@coe.northeastern.edu
SUMMARY:Suyash Pradhan MS Thesis Defense
DESCRIPTION:Title: COPILOT: Cooperative Perception using Lidar for Handoffs between Road Side Units \nCommittee Members:\nProf. Kaushik Chowdhury (Advisor)\nProf. Stratis Ioannidis\nProf. Jennifer Dy \nAbstract:\nThis thesis presents COPILOT\, a ML-based approach that allows vehicles requiring ubiquitous high bandwidth connectivity to identify the most suitable road side units (RSUs) through proactive handoffs. By cooperatively exchanging the data obtained from local 3D Lidar point clouds within adjacent vehicles and with coarse knowledge of their relative positions\, COPILOT identifies transient blockages to all candidate RSUs along the path under study. Such cooperative perception is critical for choosing RSUs with highly directional links required for mmWave bands\, which majorly degrade in the absence of LOS. COPILOT proposes three modules that operate in an inter-connected manner: (i) As an alternative to sending raw Lidar point clouds\, it extracts and transmits low-dimensional intermediate features to lower the overhead of inter-vehicle messaging; (ii) It utilizes an attention-mechanism to place greater emphasis on data collected from specific vehicles\, as opposed to nearest neighbor and distance-based selection schemes\, and (iii) it experimentally validates the outcomes using an outdoor testbed composed of an autonomous car and Talon AD7200 60GHz routers emulating the RSUs\, accompanied by the public release of the datasets. Results reveal COPILOT yields upto 69.8% and 20.42% improvement in latency and throughput compared to traditional reactive handoffs for mmWave networks\, respectively
URL:https://coe.northeastern.edu/event/suyash-pradhan-ms-thesis-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231204T103000
DTEND;TZID=America/New_York:20231204T113000
DTSTAMP:20260520T023125
CREATED:20231127T163905Z
LAST-MODIFIED:20231127T163905Z
UID:40523-1701685800-1701689400@coe.northeastern.edu
SUMMARY:Cheng Gongye PhD Dissertation Defense
DESCRIPTION:Title:\nHardware Security Vulnerabilities in Deep Neural Networks and Mitigations \nDate:\n12/4/2023 \nTime:\n10:30:00 AM \nCommittee Members:\nProf. Yunsi Fei (Advisor)\nProf. Aidong Ding\nProf. Xue Lin\nProf. Xiaolin Xu \nAbstract:\nIn the past decade\, Deep Neural Networks (DNNs) have become pivotal in numerous fields\, including security-sensitive autonomous driving and privacy-critical medical diagnosis. This Ph.D. dissertation delves into the hardware security of DNNs\, discovering their vulnerabilities to fault and side-channel attacks and exploring novel countermeasures essential for their safe deployment in critical applications. \nFault attacks disrupt computation or inject faults into parameters\, compromising the integrity of targeted applications. This dissertation demonstrates a power-glitching fault injection attack on FPGA-based DNN accelerators\, common in cloud environments\, which exploits vulnerabilities in the shared power distribution network and results in model misclassification. In response to these threats\, we introduce a novel\, lightweight defense mechanism to protect DNN parameters from adversarial bit-flip attacks. The proposed framework incorporates a dynamic channel-shuffling obfuscation scheme coupled with a logits-based model integrity monitor. The approach effectively safeguards various DNN models against bit-flip attacks\, without necessitating retraining or structural changes to the models. Furthermore\, our research expands the scope of fault analysis beyond just the parameters of DNN models. We thoroughly examine the entire implementation of commercial products\, defying the prevailing assumption that quantized DNNs are inherently resistant to bit-flips. \nSide-channel attacks exploit information leakage of system implementations\, such as power consumption and electromagnetic emanations\, to reveal system secrets and therefore compromise confidentiality. This dissertation makes significant contributions to side-channel assisted model extraction of DNNs. We present a floating-point timing side-channel attack on x86 CPUs that reverse-engineers DNN model parameters in software implementations. For hardware accelerators\, we target the state-of-the-art AMD-Xilinx deep-learning processor unit (DPU)\, a reconfigurable engine dedicated to convolutional neural networks (CNNs) and representing the most complex commercial FPGA accelerator with encrypted IPs. Our work demonstrates that electromagnetic analysis can be leveraged to recover the data flow and scheduling of the DNN accelerators\, facilitating follow-on architecture and parameter extraction attacks. To mitigate EM side-channel model extraction attacks\, we introduce a novel defense mechanism that devises a random importance-aware activation mask on input pixels to disrupt the operation alignment on EM traces\, with minimal performance and efficiency impacts. \nOverall\, this dissertation significantly deepens the understanding of hardware security of DNN models. It makes important contributions in discovering novel and critical vulnerabilities of DNN inference pertaining to system implementations\, and proposing effective and practical solutions for securing DNNs in mission-critical environments. The research work marks a substantial step forward in the development of resilient and secure AI systems.
URL:https://coe.northeastern.edu/event/cheng-gongye-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231129T163000
DTEND;TZID=America/New_York:20231129T170000
DTSTAMP:20260520T023125
CREATED:20231127T164136Z
LAST-MODIFIED:20231127T164136Z
UID:40525-1701275400-1701277200@coe.northeastern.edu
SUMMARY:Aria Masoomi PhD Proposal Review
DESCRIPTION:Title:\nMaking Deep Neural Network Transparent \nDate:\n11/29/2023 \nTime:\n4:30:00 PM \nCommittee Members:-\nProf. Jennifer Dy\nProf. Eduardo Sontag\nProf. Mario Sznaier\nProf. Peter Castaldi \nAbstract:\nAs machine learning algorithms are deployed ubiquitously to a variety of domains\, it is imperative to make these often black-box models transparent. The ability to interpret and comprehend the reasoning behind machine learning models plays a pivotal role in increasing user trust. It not only offers insights into how a model functions but also opens avenues for model enhancements. \nThis research delves into the realm of interpretability\, focusing on the dichotomy between ‘intrinsic’ and ‘post hoc’ interpretability. Intrinsic interpretability involves constraining the complexity of the machine learning model itself\, resulting in models inherently interpretable due to their simplicity\, such as decision trees or sparse linear regression. On the other hand\, post hoc interpretability employs techniques that assess the model’s behavior after training\, offering insights into the model’s outcomes. Examples of post hoc techniques include permutation feature importance and the Shapley value method for feature importance. \nThe core contribution of this Thesis proposal lies in the development of novel methods to enhance both intrinsic and post hoc interpretability. These methods aim to advance the field by offering new perspectives on understanding machine learning models\, thereby contributing to the ongoing discourse on model transparency and user trust.
URL:https://coe.northeastern.edu/event/aria-masoomi-phd-proposal-review-2/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231129T150000
DTEND;TZID=America/New_York:20231129T160000
DTSTAMP:20260520T023125
CREATED:20231127T163756Z
LAST-MODIFIED:20231127T163756Z
UID:40521-1701270000-1701273600@coe.northeastern.edu
SUMMARY:Aria Masoomi PhD Proposal Review
DESCRIPTION:Title:\nMaking Deep Neural Network Transparent \nDate:\n11/29/2023 \nTime:\n3:00:00 pm \nCommittee Members:\nProf. Jennifer Dy (Advisor)\nProf. Mario Sznaier\nProf. Eduardo Sontag\nProf. Peter Castaldi \nAbstract:\nAs machine learning algorithms are deployed ubiquitously to a variety of domains\, it is imperative to make these often black-box models transparent.\nThe ability to interpret and comprehend the reasoning behind machine learning models plays a pivotal role in increasing  user trust. It not only offers insights into how a model functions but also opens avenues for model enhancements. \nThis research delves into the realm of interpretability\, focusing on the dichotomy between ‘intrinsic’ and ‘post hoc’ interpretability. Intrinsic interpretability involves constraining the complexity of the machine learning model itself\, resulting in models inherently interpretable due to their simplicity\, such as decision trees or sparse linear regression. On the other hand\, post hoc interpretability employs techniques that assess the model’s behavior after training\, offering insights into the model’s outcomes. Examples of post hoc techniques include permutation feature importance and the Shapley value method for feature importance. \nThe core contribution of this Thesis proposal lies in the development of novel methods to enhance both intrinsic and post hoc interpretability. These methods aim to advance the field by offering new perspectives on understanding machine learning models\, thereby contributing to the ongoing discourse on model transparency and user trust.
URL:https://coe.northeastern.edu/event/aria-masoomi-phd-proposal-review/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231127T080000
DTEND;TZID=America/New_York:20231127T170000
DTSTAMP:20260520T023125
CREATED:20231127T163640Z
LAST-MODIFIED:20231127T163640Z
UID:40519-1701072000-1701104400@coe.northeastern.edu
SUMMARY:Bruno Souto Maior Muniz Morais PhD Dissertation Defense
DESCRIPTION:Title:\nEnabling Domain Platform Design for Streaming Applications: A Holistic Approach \nCommittee Members:\nGunar Schirner (Advisor)\nProf. David Kaeli\nProf. Hamed Tabkhi (UNCC) \nTime:\n10:00:00 AM \nLocation: ISEC 601 \nAbstract:\nIn recent years\, more demanding streaming applications make striking a balance between high compute performance and efficiency paramount in platforms designs for edge computing. In addition\, designing a platform that is optimized for a single application is costly due to non-recurring engineering (NRE) costs. In contrast\, multiple applications can be grouped in domains\, e.g. computer vision\, software-defined radio. Leveraging shared characteristics of similar applications within a domain\, e.g. structural composition/computation patterns\, a single domain platform that caters to these similarities and accelerates applications can be generated\, thus benefiting multiple applications at once and dramatically improving NRE and time-to-market (TTM). \nThis dissertation introduces methodologies atvarious abstraction levels to enable streamlined domain platform design for streaming applications. Thrust 1 introduces high level DSE methods based on integer linear programming (ILP)\, Tile-based Synchronization Aware ILP (TSAR-ILP). Initially\, single-application platform allocations are considered using TSAR-ILP. While TSAR-ILP only focuses on applications in isolation\, its formulation lays the foundations for DmTSAR-ILP\, a method that performs domain DSE with multiple applications\, obtaining an optimal unified platform allocation that and achieving an increase of 22.5% in throughput\, while being 70x faster when compared to previous methods (MG-DmDSE). However\, DmTSAR-ILP aims to aggregate all applications fairly. This presents a challenge when the designer wishes to focus on a subset of applications. To enable ultimate flexibility in a product-oriented setting\, modeled after a market analysis process\, this dissertation introduces ProdDSE. ProdDSE enables application prioritization while also introducing concurrent application modeling and a multi-objective optimization (area\, performance) approach. This enables up to a 3.4x boost in performance depending on use case\, while also providing gains in DSE runtime (4.3x faster). \nThrust 2 introduces Sedona\, a domain-specific language (DSL) and exploration enviroment that captures parametric dataflow application descriptions with language features dedicated to streaming applications. A design identified by Thrust 1 can be further refined using the tools in Thrust 2\, by capturing the connectivity of a design using Sedona. Then\, automatic wiring is performed for target outputs such as timing-aware simulations or RTL-level code\, enabling structural manipulation at a high-level description without the burden of low-level manual integration. \nFinally\, to better guide the high-level decisions performed in Thrust 1 and further exploration/integration in Thrust 2\, Thrust 3 considers the implications of HWACC topology choices in an HWACC-rich SoC. The ACTAR flow is introduced to explore different topologies in a RISC-V based SoC and the side-effects of topology and memory sizing choices on the system-wide performance and synchronization burdens due computation offloading to HWACCs. This produces valuable and actionable insights for designers to make informed choices on system-level compositions depending on application communication and computation demands.
URL:https://coe.northeastern.edu/event/bruno-souto-maior-muniz-morais-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231117T083000
DTEND;TZID=America/New_York:20231117T093000
DTSTAMP:20260520T023125
CREATED:20231020T143903Z
LAST-MODIFIED:20231020T143903Z
UID:39994-1700209800-1700213400@coe.northeastern.edu
SUMMARY:Mahshid Asri PhD Dissertation Defense
DESCRIPTION:Title:\nDevelopment of Anomaly Detection and Characterization Algorithms Using Wideband Radar Image Processing for Security Applications \nDate:\n11/17/2023 \nTime:\n8:30:00 AM \nLocation: 302 Stearns \nCommittee Members:\nProf. Carey Rappaport (Advisor)\nProf. Charles DiMarzio\nProf. Edwin Marengo \nAbstract:\nDetection and characterization of suspicious body-worn objects is necessary for safe and effective personnel screening. In airports\, developing a precise system that can distinguish threats and explosives from objects like money belt can reduce the pat-down significantly while maintaining effective security. This dissertation proposes two main algorithms which are developed for different millimeter-wave radar systems. The first project is a material characterization algorithm designed for a 30 GHz wideband multi bi-static radar system used for passenger screening in airports. The proposed algorithm can automatically distinguish lossless materials from lossy ones and calculate their thickness and permittivities. Starting from the radar reconstructed image showing a cross-section of the body\, we extract the nominal body contour using Fourier series\, separate body and object responses\, categorize the object as lossy or lossless based on the depression and protrusion of the body contour\, and finally predict possible values for the object’s permittivity and thickness. Our resulting classification is good\, implying fewer nuisance alarms at check points. We have also trained a deep learning model for pixel-wise localization of body worn anomalies. The second project is a metal detection algorithm developed to monitor pedestrians walking along a sidewalk for large\, concealed metallic objects. Finite Difference Frequency Domain and SAR algorithms are used to simulate the images produced by this 6 GHz wideband radar system. A deep learning model has then been used to predict a pixel level mask for the body and anomaly based on the inputted radar image.
URL:https://coe.northeastern.edu/event/mahshid-asri-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231024T130000
DTEND;TZID=America/New_York:20231024T140000
DTSTAMP:20260520T023125
CREATED:20231020T144256Z
LAST-MODIFIED:20231020T144857Z
UID:39990-1698152400-1698156000@coe.northeastern.edu
SUMMARY:Baolin Li PhD Proposal Review
DESCRIPTION:Title:\nMaking Machine Learning on HPC Systems Cost-Effective and Carbon-Friendly \nDate:\n10/24/2023 \nTime:\n1:00:00 PM \nCommittee Members:\nProf. Devesh Tiwari (Advisor)\nProf. Ningfang Mi\nDr. Vijay Gadepally \nAbstract:\nThe end users want the machine learning (ML) training/inference services to be lightning-fast. However\, the cost\, incurred by service providers\, to support these lighting-fast ML services is often prohibitively high. Large-scale HPC and data centers are struggling to keep their cost low as they provide faster ML services\, while the excessive demand for these services is already negatively impacting our environment due to the large carbon footprint of ML services. Therefore\, this dissertation focuses on better understanding the complex trade-off among performance\, cost\, and environmental footprint of ML models. \nThis dissertation asks three simple questions: (1) Is slower hardware always worse? (2) Is more expensive hardware always better? (3) Should we always strive to design and train ML models with the highest possible accuracy? As this dissertation reveals\, the answers to these questions are more complex than what the conventional wisdom suggests. In fact\, simplistic answers — based on first-order intuitions — can lead to missed opportunities in terms of performance efficiency\, cost-effectiveness\, and carbon footprint. \nIn this dissertation\, we build multiple novel frameworks to demonstrate that mixing slower-and-cheaper hardware with faster-and-expensive hardware can unlock much higher performance- and cost-effectiveness than using only faster-and-expensive hardware configurations. But\, unlocking this potential requires a careful design — a design that carefully exploits the diversity in ML inference workload characteristics and adapts to varying ML inference request loads. Next\, this dissertation demonstrates that while the highest-possible-accuracy ML models are desirable\, using such models can have a severe negative environmental impact. To mitigate this challenge\, this dissertation builds an experimental framework to reduce the carbon footprint of ML inference services. The key idea\, behind this framework\, is to mix the lower-quality ML models with higher-quality ML models intelligently and share the hardware resources during inference query execution to reduce the excessive carbon footprint of high-quality ML model inference\, esp. during the periods when a data center’s energy source has high carbon intensity. The extensive experimental evaluation confirms that significant carbon emission reductions can be achieved with transient\, very minimal\, and configurable loss in accuracy. \nAs we make rapid advances in the era of large-language models (LLMs) and foundation models\, the novel methods and open-source tools presented in this dissertation will enable us to build ML services faster but cheaper and in an environmentally sustainable manner
URL:https://coe.northeastern.edu/event/baolin-li-phd-proposal-review/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230919T100000
DTEND;TZID=America/New_York:20230919T113000
DTSTAMP:20260520T023125
CREATED:20230901T134555Z
LAST-MODIFIED:20230901T134555Z
UID:38048-1695117600-1695123000@coe.northeastern.edu
SUMMARY:Zhengnan Li PhD Proposal
DESCRIPTION:Title:\nMulti-user Communications \nDate:\n9/19/2023 \nTime:\n10:00:00 AM \nLocation:\nISEC 432 \nCommittee Members:\nProf. Milica Stojanovic (Advisor)\nDr. Xiaowen Wang (Apple)\nProf. Tommaso Melodia \nAbstract:\nExtensive research has been conducted with respect to underwater acoustic communications and networking\, owing to their profound importance in various applications\, such as fish farming and the oil-and-gas industry. Acoustic networks involve scenarios where multiple users necessitate transmitting data to a central base station. Techniques such as time-division or code-division multiple access are commonly employed in such networks. However\, these techniques entail a trade-off: each user’s transmission rate remains confined to a fraction of the overall available resources. Consequently\, the endeavor to accommodate an increased number of users within the usable resources results in a diminished data rate for each user. Conversely\, if the goal is to maintain a consistent per-user data rate\, the acquisition of additional bandwidth becomes imperative. Yet\, the feasibility of this approach is challenged by the inherent limitations of available bandwidth in underwater acoustic systems. \nIn this proposal\, I will present preliminary findings that involve the utilization of code and space division multiple access systems in conjunction with orthogonal frequency division multiplexing (OFDM)\, which is a key enabler for the current and future generations of wireless systems. Additionally\, the proposal delves into the prospect of establishing an underwater acoustic channel repository—an effort designed to emulate underwater acoustic channels\, thereby alleviating the necessity for extensive real-world underwater experimentation. This proposal also includes several recent long-distance underwater experiments carried out in Japan\, examining various ideas regarding the frequency offset compensation problem in OFDM systems. Furthermore\, beyond the outcomes in underwater acoustic communication\, this proposal encompasses a series of explorations involving micro electromechanical systems (MEMS) and the utilization of terahertz frequencies.
URL:https://coe.northeastern.edu/event/zhengnan-li-phd-proposal/
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:20230918T083000
DTEND;TZID=America/New_York:20230918T150000
DTSTAMP:20260520T023125
CREATED:20230915T172337Z
LAST-MODIFIED:20230915T172337Z
UID:38782-1695025800-1695049200@coe.northeastern.edu
SUMMARY:Technical Exchange Program between NU and DFG Delegation
DESCRIPTION:8:30-8:50 am: Lab tour of the NU Cleanroom\, labs\, etc. (Meet at the entrance of Egan Building\, 120 Forsyth Street\, Boston\, MA 02115\, Led by Nian Sun)\n8:50-9am Move to McLeod Suites\, Curry Student Center\, Northeastern University\n9:00-9:15: Opening remarks from NU and DFG (VPRD Kim Holloway\, Senator Professor Marion Merklein)\n9:15-10:30: Presentations from NU (9 speakers with 8 minutes each)\n\nRavinder Dahiya\, Electrical and Computer Engineering\nMatteo Rinaldi\, Electrical and Computer Engineering\nYongmin Liu\, Electrical and Computer Engineering\, Mechanical and Industrial Engineering\nXufeng Zhang\, Electrical and Computer Engineering\nBen Davaji\, Electrical and Computer Engineering\nSwastik Kar\, Physics Department\nPaul Stevenson\, Physics Department\nRuobing Bai\, Mechanical and Industrial Engineering\nHongli Zhu\, Mechanical and Industrial Engineering\n\n\n10:30-10:45: Coffee break\n10:45-11:45: Presentations from NU and DFG Delegation (7 speakers with 8 minutes each)\n\nSrinivas Tadigadapa\, Electrical and Computer Engineering\nSrirupa Chakraborty\, Chemical Engineering\nNian Sun\, Electrical and Computer Engineering\nArne Berger\, University of Applied Sciences Anhalt\nRobert Böhm\, Leipzig University of Applied Sciences\nAlexander Prange\, University of Applied Sciences Niederrhein\nJessica Friess\, University of Applied Sciences Niederrhein\n\n\n12-1:30pm: Lunch break (working lunch at McLeod Suites)\n1:30-2:30pm: Presentations from DFG Delegation (7 speakers with 8 minutes each)\n\nMargit Geißler Bonn-Rhein-Seig University of Applied Sciences\nJens Helbig\, Nürnberg Tech\nBenjamin Neding\, University of Applied Sciences Lübeck\nJörn V. Wochnowski University of Applied Sciences Lübeck\nRomana Piat\, University of Applied Sciences Darmstadt\nHolger Saage\, Univ. of Applied Sciences Landshut\, Competence Center for Lightweight Design\nJulian Tornow\, Hochschule Ruhr-West University of Applied Sciences\n\n\n2:30-3pm: DFG / NU discussions on potential collaborations\, programs\, etc.
URL:https://coe.northeastern.edu/event/technical-exchange-program-between-nu-and-dfg-delegation/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230901T130000
DTEND;TZID=America/New_York:20230901T140000
DTSTAMP:20260520T023125
CREATED:20230816T150411Z
LAST-MODIFIED:20230816T150432Z
UID:37858-1693573200-1693576800@coe.northeastern.edu
SUMMARY:Mostafa Abedi PhD Proposal Review
DESCRIPTION:Title: Power-Efficient and Security-Enhancing Techniques for Ultra-low Power IoT  Devices \nCommittee Members: \nProf. Aatmesh Shrivastava (Advisor)\nProf. Marvin Onabajo\nProf. Nian X. Sun \nAbstract:\nInternet-of-things (IoT) devices often rely on ambient energy sources such as photovoltaic (PV) cells and thermoelectric generators (TEGs) for their operation. Minimizing power loss through ambient energy harvesting optimization can significantly extend the battery life or support battery-free sensor nodes in IoT devices. A maximum power point tracking (MPPT) circuit is often used for impedance matching to maximize energy transfer efficiency. This research proposes an ultra-low power\, high-tracking efficiency MPPT circuit based on Hill-Climbing (HC) algorithm suitable for micro-power DC harvesters. The proposed system employs a modified version of the hill-climbing algorithm. In case of input power changes and consequent deviation of the harvester from the MPP\, an integrated Power Change Detector (PCD) is proposed to reactivate the MPPT circuit. The PCD detects changes in input power and activates the MPPT circuit\, enabling automatic activation and resulting in substantial power savings. Furthermore\, due to the proposed power estimation technique\, the MPPT is not dependent on the internal structure of the energy source\, and its tracking efficiency is unrelated to the conversion ratio of the converter. This approach enables us to achieve a peak tracking efficiency of over 99.9\%. To adjust the input power of the harvester to track the maximum power point\, we propose a new\, efficient Pulse Width Modulation (PWM) circuit. This circuit exhibits a wide duty cycle range\, low power consumption\, linearity\, and robustness against variations. \nThis research also focuses on increasing the security of IoT devices. In the past\, chip fabrication was mostly done internally by semiconductor firms. Now\, it is more collaborative\, pulling in designs from various sources and having a few factories produce them. This new way of working means that companies that only handle design might face more challenges like the threat of hardware Trojans (HT) being added either during the design phase or production. With that in mind\, we introduce a different circuit design approach. We aim to find these Trojans\, particularly the newer analog Trojans. The idea is to boost the security of IoT devices by detecting these issues early. In addition\, to improve the security of IoT systems\, we propose an ultra-low power energy monitoring system (EMS) to detect and mitigate denial-of-sleep (DoSL) attacks. In this project\, we explore a new method of defense against DoSL attacks by monitoring energy consumption. We will implement a low-power system to monitor the lifetime of the IoT node by continuously evaluating the harvested\, stored\, and consumed energy in the node.
URL:https://coe.northeastern.edu/event/mostafa-abedi-phd-proposal-review/
LOCATION:532 ISEC\, 360 Huntington Ave\, Boston\, MA\, 02115\, United States
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230825T100000
DTEND;TZID=America/New_York:20230825T110000
DTSTAMP:20260520T023125
CREATED:20230816T145957Z
LAST-MODIFIED:20230816T145957Z
UID:37862-1692957600-1692961200@coe.northeastern.edu
SUMMARY:Julian Gutierrez PhD Proposal Review
DESCRIPTION:Title: Towards Real-Time Safe Flight Paths for Urban Air Mobility \nCommittee Members:\nProf. David Kaeli\, (Advisor)\nProf. Pau Closas\nDr. Evan Dill (NASA)\nDr. Natasha Neogi (NASA) \nAbstract:\nThe emergence and development of advanced technologies and vehicle types have created a growing demand for introducing new forms of flight operations. These new and increasingly complex operational paradigms\, such as Advanced and Urban Air Mobility (AAM/UAM)\, present regulatory authorities and the aviation community with the challenge of finding methods to integrate these emerging operations without significant additional risk to pedestrians and infrastructure. Predictive and autonomous risk mitigation capabilities become critical to meet this challenge. However\, urban environments experience effects that are computationally expensive to model\, limiting conventional aviation concepts\, policy\, and risk prediction tools from being effectively translated into this space. With the emergence of High-Performance Computing (HPC) ecosystems in the last two decades\, we can use these software and hardware capabilities to help bridge the gap between real-time predictive responses and modeling accuracy. \nIn this dissertation we first present a simulation framework to estimate the quality of Global Navigation Satellite System (GNSS) performance for autonomous aircraft in urban environments. We propose a new algorithm designed for HPC to accelerate modeling the characteristic effects of dense urban canyons on GNSS\, allowing the extension of established GNSS integrity techniques into urban navigation. Additionally\, we provide a thorough validation of the simulator\, which proves high-accuracy modeling when compared to sensors in the real world. Second\, we use this simulation framework as the input into a new 4D path-planning algorithm based on an adaptation of the Bellman-Ford algorithm. HPC techniques are employed to accelerate the algorithm to produce flight paths that minimize exposure to GNSS risks. We evaluate the computational cost of satellite availability fluctuations by prioritizing events when satellite availability changes as triggers for these updates.
URL:https://coe.northeastern.edu/event/julian-gutierrez-phd-proposal-review/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230818T143000
DTEND;TZID=America/New_York:20230818T153000
DTSTAMP:20260520T023125
CREATED:20230817T142938Z
LAST-MODIFIED:20230817T142938Z
UID:37895-1692369000-1692372600@coe.northeastern.edu
SUMMARY:Xu Yizhe MS Thesis Defense
DESCRIPTION:Title:\nIntegration of Polyimide Flexible PCB Wings in Northeastern’s Aerobat \nLocation:\nRoom: ISEC 532\, Teams link \nCommittee Members:\nProf. Alireza Ramezani(Advisor)\nProf. Rifat Sipahi \nAbstract:\nThe principal aim of this Master’s thesis is to propel the optimization of the membrane wing structure of the Northeastern Aerobat through origami techniques and enhancing its capacity for secure hovering within confined spaces. Bio-inspired drones offer distinctive capabilities that pave the way for innovative applications\, encompassing wildlife monitoring\, precision agriculture\, search and rescue operations\, as well as the augmentation of residential safety. The evolved noise-reduction mechanisms of birds and insects prove advantageous for drones utilized in tasks like surveillance and wildlife observation\, ensuring operation devoid of disturbances. Traditional flying drones equipped with rotary or fixed wings encounter notable constraints when navigating narrow pathways. While rotary and fixed-wing systems are conventionally harnessed for surveillance and reconnaissance\, the integration of onboard sensor suites within micro aerial vehicles (MAVs) has garnered interest in vigilantly monitoring hazardous scenarios in residential settings. Notwithstanding the agility and commendable fault tolerance exhibited by systems such as quadrotors in demanding conditions\, their inflexible body structures impede collision tolerance\, necessitating operational spaces free of collisions. Recent years have witnessed an upsurge in integrating soft and pliable materials into the design of such systems; however\, the pursuit of aerodynamic efficiency curtails the utilization of excessively flexible materials for rotor blades or propellers. This thesis introduces a guard design incorporating feedback-driven stabilizers\, enabling stable hovering flights within Northeastern’s Robotics-Inspired Study and Experimentation (RISE) cage.
URL:https://coe.northeastern.edu/event/xu-yizhe-ms-thesis-defense/
LOCATION:532 ISEC\, 360 Huntington Ave\, Boston\, MA\, 02115\, United States
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230818T140000
DTEND;TZID=America/New_York:20230818T150000
DTSTAMP:20260520T023125
CREATED:20230817T143207Z
LAST-MODIFIED:20230817T143207Z
UID:37889-1692367200-1692370800@coe.northeastern.edu
SUMMARY:Haoling Li MS Thesis Defense
DESCRIPTION:Title: Ultracompact and Conformal Magnetodielectric Antennas \nCommittee Members:\nProf. Nian-Xiang Sun (Advisor)\nProf. Xufeng Zhang\nProf. Marvin Onabajo \nAbstract:\nNovel approaches are needed for improving antenna performance\, enhancing efficiency\, and reducing the size\, profile\, number\, and signature of antennas. Efficient conformal antennas are increasingly replacing traditional antennas across platforms such as ships\, aircraft\, and human interfaces. Magnetodielectric antennas made with high-hesistivity magnetic materials are getting more and more attention. Defined as the maximum magnetic conductivity\, hesitivity is directly related to the radiation efficiency of magnetodielectric antennas\, with a higher hesitivity corresponding to higher attainable efficiency. In this study\, new ultra-compact conformal magnetodielectric antennas are demonstrated\, employing commercially available ferrite ceramic substrates. Through rigorous simulation and fabrication\, a comprehensive comparison of our magnetodielectric antennas with reference monopole antennas demonstrated superior efficiency\, enhanced gain\, bandwidth\, and a substantial reduction in antenna size compared to monopole antennas. State-of-the-art hesitivity as high as 6×10^6 Ω/m has been reported in CoZrNb alloy films\, with an expectation of further 10× improvement in thin carrier substrates. This study forecasts the potential development of new magnetic materials with higher hesitivity\, leading to further advancements in magnetodielectric antennas with enhanced radiation efficiency and ground plane immunity.
URL:https://coe.northeastern.edu/event/haoling-li-ms-thesis-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230818T100000
DTEND;TZID=America/New_York:20230818T110000
DTSTAMP:20260520T023125
CREATED:20230817T143431Z
LAST-MODIFIED:20230817T143431Z
UID:37897-1692352800-1692356400@coe.northeastern.edu
SUMMARY:Jaehyeon Ryu PhD Dissertation Defense
DESCRIPTION:Title: Materials Strategies for Scaling Soft Neuroelectrode Arrays \nLocation: Snell 012/Teams \nCommittee Members:\nProf. Hui Fang (Advisor)\nProf. Yongmin Liu\nProf. Ryan Koppes \nAbstract:\nThe evolution of electronics to seamlessly interface with biological tissue hinges on addressing multifaceted material constraints spanning electrochemical\, electrical\, and mechanical domains. Conventional bioelectronic interfaces\, while endowed with established electrochemical functionality\, remain hampered by rigidity that contradicts the pliability of surrounding tissue. While conductive materials exhibiting tissue-like softness and stretchability have been realized\, their potential for electrochemical probing of tissue is impeded by strain-induced performance degradation and an ill-suited integration with the irregular tissue interface. Nevertheless\, a significant challenge in ultrasoft bioelectronics pertains to scalability for achieving cellular resolution\, primarily due to mechanical disparities between conventional microelectronic materials and soft elastomer substrates. In this thesis\, by using a novel approach involving a multifunctional nanomesh\, composed of distinct purposefully designed layers including polymer for mechanical buffering\, metal for electrical conduction\, and low impedance coating for electrochemical interfacing in the same nanomeshed structure\, the resultant microelectrodes\, scalable down to 20μm at cellular resolution\, exhibit comparable performance to rigid devices alongside a stretchablity of approximately 50%\, with potential for future enhancement through in-plane structural optimizations. In addition\, we introduce a high-density neuroelectronic array featuring 256 filamentary neuroelectrodes on a flexible substrate. These electrodes are integrated with a single-transistor multiplexing acquisition circuit\, effectively reducing noise and footprint while potentially extending device lifetime. Remarkably\, the array’s rollable contact pad design allows for minimally invasive delivery through a syringe. Experimental validation demonstrates the array’s capability to record neural signals with high tone decoding accuracy. Utilizing high-density arrays of these microelectrode arrays\, this unique frame works holds significant promise for advancing the field of neural interfacing\, enabling a wide range of application from fundamental neuroscience studies to various biomedical applications.
URL:https://coe.northeastern.edu/event/jaehyeon-ryu-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230817T130000
DTEND;TZID=America/New_York:20230817T140000
DTSTAMP:20260520T023125
CREATED:20230816T150620Z
LAST-MODIFIED:20230816T150620Z
UID:37856-1692277200-1692280800@coe.northeastern.edu
SUMMARY:Zhiyong Zhang PhD Proposal
DESCRIPTION:Title: Towards Indoor Mapping and Navigation with Perceptual Aliasing using Visual Semantic SLAM \nCommittee Members:\nProf. Hanumant Singh (Advisor)\nProf. Huaizu Jiang\nProf. David Rosen \nAbstract:\nModern SLAM (Simultaneous Localization And Mapping) techniques allow us to create accurate 3D maps of the environment primarily using visual sensors in GPS-denied regions. In this context\, numerous deep learning-based approaches have emerged\, enabling the extraction of rich semantic information from images\, including shapes\, objects\, and text. \nLeveraging these technologies\, our aim is to construct comprehensive 3D maps of indoor environments\, which could be utilized by robots for path planning and navigation. Additionally\, the solution can be integrated with a large language model\, enabling the robot to interact intuitively with people. \nThis research comprises four main components: Semantic Feature Extraction and Tracking with SLAM: Given that the same semantic features can appear in multiple frames\, some of which may not be conducive to feature detection and recognition (such as blurry images or distant views)\, we are developing a pipeline to ensure the optimal detection and recognition of semantic features within the most suitable frame. The pipeline also involves tracking the same feature across frames while maintaining its 3D location in the global map. \nResolving Perceptual Aliasing: Many indoor places can exhibit high visual similarity\, which confuses the robot when powered up with a prior map in its memory. Semantic features can be used to localize the robot in the map\, determining its specific floor or room. This capability can also aid SLAM in performing loop closure with high-level information. \nCross-Floor Constraints for SLAM Optimization: Most buildings contain a symmetric layout across floors\, which can be exploited to establish constraints between them. For instance\, vertically aligned rooms like 425 and 525\, as well as elevators\, offer opportunities for vertical constraint. Such constraints can enhance SLAM optimization\, resulting in improved map accuracy. \nIndoor Path Planning and Navigation: Once we have a comprehensive 3D map of the indoor environment\, path planning becomes an intuitive way to utilize this map. With semantic features integrated into the map\, the robot can associate 3D point clouds with high-level information\, such as door numbers or office names. Large language models are available to provide a more human-like way to interact with the robot. For example\, a command like “Navigate to Professor Hanumant’s office and locate the book ‘The Hitchhiker’s Guide to the Galaxy'” can be executed by the robot.
URL:https://coe.northeastern.edu/event/zhiyong-zhang-phd-proposal/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230817T090000
DTEND;TZID=America/New_York:20230817T110000
DTSTAMP:20260520T023125
CREATED:20230802T192116Z
LAST-MODIFIED:20230802T192116Z
UID:37698-1692262800-1692270000@coe.northeastern.edu
SUMMARY:Jagatpreet Nir PhD Proposal
DESCRIPTION:Title: Low Contrast Visual Sensing and Inertial Navigation in GPS Denied Environments \nCommittee Members:\nProf. Hanumant Singh\nProf. Martin Ludvigsen\nProf. Pau Closas\nProf. Michael Everrett \nAbstract:\nVisual inertial navigation has shown remarkable performance in publicly available datasets\, assuming certain ideal conditions such as textured scenes\, uniform illumination\, and static environments. However\, real-world scenarios often violate these assumptions\, resulting in significant visual degradation. Consequently\, the classical visual navigation pipelines fail and produce erroneous results\, rendering these systems ineffective for demanding field robotic missions. \nThis research aims to enhance the robustness of visual-inertial systems in visually degraded situations\, taking a comprehensive approach from both systems and algorithm perspectives. The work encompasses two primary objectives. Firstly\, it focuses on refining the characterization of MEMS-based inertial sensors and their error propagation in position\, while proposing improved dead-reckoning algorithms. Secondly\, it explores the performance limits of visual navigation under moderate to extreme visual degradation and investigates novel algorithms that leverage deep learning methods to bolster the visual navigation engine. To validate the efficacy of these advancements\, new datasets comprising drone and underwater robot scenarios are utilized\, demonstrating the applicability of this work in field robotic applications. \nBy addressing the limitations of existing visual-inertial navigation systems and developing robust algorithms\, this research aims to significantly enhance the reliability and performance of such systems in visually degraded environments\, thus expanding their potential for real-world applications in demanding field robotic missions.
URL:https://coe.northeastern.edu/event/jagatpreet-nir-phd-proposal/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230815T153000
DTEND;TZID=America/New_York:20230815T163000
DTSTAMP:20260520T023125
CREATED:20230816T150241Z
LAST-MODIFIED:20230816T150241Z
UID:37860-1692113400-1692117000@coe.northeastern.edu
SUMMARY:Sumegha Singhania MS Thesis Defense
DESCRIPTION:Title: Exploring Log of RGB Space as a Better Input for Computer Vision Tasks \nCommittee Members:\nProf. Bruce Maxwell (Advisor)\nProf. Hanumant Singh\nProf. David Rosen\nProf. Mahdi Imani \nAbstract:\nThere are specific\, physics-based rules that govern the interaction of light and matter. Though studied extensively in the greater computer vision community\, these rules are largely broken by common image processing techniques like JPEG compression and sRGB conversion. While the reliability and usability of color and intensity found in RAW images might better train networks to successfully complete vision-based tasks\, these smaller\, more heavily-processed formats have become the standard input for training sets. As a result\, many of the images used to train neural networks do not retain the inherent structure that would enable neural networks to learn more general rules that exist in the natural world. \nWe hypothesize that using linear RGB or log RGB images\, which preserve the physics of reflection\, can simplify the learning process for certain vision tasks\, enhance overall robustness and performance\, and provide invariance to visual variations that exist in real-world vision applications. Our research demonstrates that employing linear and log RGB images to train deep networks for the task of object detection improves their performance when using the same network architecture and the same set of training images. Additionally\, we also show that the networks trained on linear and log RGB show greater resilience to variations in intensity and color balance. Specifically\, the network trained on linear and log RGB inputs shows invariance to intensity and color balance variations that were not encountered during training\, while the network trained on the same images in sRGB JPEG format experiences significant performance degradation. To understand the reasons behind this disparity\, we analyze and visualize low-level features in log RGB\, linear RGB\, and JPEG data. Our findings reveal that the log space preserves certain relevant features across variations in intensity and color balance.
URL:https://coe.northeastern.edu/event/sumegha-singhania-ms-thesis-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230811T140000
DTEND;TZID=America/New_York:20230811T150000
DTSTAMP:20260520T023125
CREATED:20230731T152837Z
LAST-MODIFIED:20230731T152837Z
UID:37660-1691762400-1691766000@coe.northeastern.edu
SUMMARY:Jinkun Zhang PhD Proposal Review
DESCRIPTION:Location: ISEC305 \nTitle: Low-latency forwarding\, caching and computation placement in data-centric networks \nCommittee Members:\nProf. Edmund Yeh (Advisor)\nProf. Stratis Ioannidis\nProf. Kaushik Chowdhury \nAbstract:\nWith the exponential growth of data- and computation-intensive network applications\, such as real-time augmented reality/virtual reality rendering and large-scale language model training\, traditional cloud computing frameworks exhibit inherent limitations. These limitations include significant round-trip delays caused by backhaul network capacity bottlenecks and exorbitant costs associated with centralized computing power\, e.g.\, training GPT-4 requires over 16\,000 A100 GPUs.\nTo address these challenges\, dispersed computing has emerged as a promising next-generation networking paradigm. By enabling geographically distributed nodes with heterogeneous computation capabilities to collaborate\, dispersed computing overcomes the bottlenecks of traditional cloud computing and facilitates in-network computation tasks\, including the training of large models. \nFurthermore\, in data-centric networking\, communication and computation are resolved around data names instead of host addresses.\nThe deployment of network caches\, by enabling data reuse\, offers substantial benefits for data-centric networks.\nFor instance\, consider a scenario where multiple machine learning applications seek to train different models simultaneously. These application could (partially) share data samples and intermediate results\, and carefully designed data-reusing mechanisms become necessary. Optimal caching of data or intermediate results can significantly reduce the overall training cost\, compared to each application independently gathering and transmitting data. \nTo efficiently manage computation and storage resources in heterogeneous data-centric networks\, several frameworks have been proposed with different design objectives\, such as optimizing throughput or incorporating multicast flows. However\, previous approaches have failed to minimize average user delay despite the latency sensitivity of numerous real-world applications. \nThis proposal aims to address this gap by introducing a low-latency framework that jointly optimizes packet forwarding\, storage deployment\, and computation placement. The proposed framework effectively supports data-intensive and latency-sensitive computation applications in data-centric networks with heterogeneous communication\, storage\, and computation capabilities. \nSpecifically\, to minimize user latency in congestible networks\, we model delays caused by link transmissions and CPU computations using\ntraffic-dependent nonlinear functions. We formulate the joint forwarding\, caching\, and computation problem as an NP-hard mixed-integer non-submodular optimization\, for which no constant-factor approximation algorithms are currently known. To make progress\, we approach the joint problem by dividing it into two subproblems: the joint forwarding/computation problem and the joint forwarding/caching problem. Despite the non-convexity of the former subproblem\, we provide a set of sufficient optimality conditions that lead to a distributed algorithm with polynomial-time convergence to the global optimum. For the latter subproblem\, we demonstrate its NP-hardness and non-submodularity\, even after continuous relaxation. We show that the objective function is a sum of a convex function and a geodesic convex function\, and propose a set of conditions that provide a finite bound from the optimum. To the best of our knowledge\, our method represents the first analytical progress in addressing the joint caching and forwarding problem with arbitrary topology and non-linear costs. Furthermore\, our theoretical bound leads to a constant-factor approximation under additional assumptions. \nAs future work\, we propose to develop a novel in-network large model training framework\, building upon the aforementioned method.\nDue to the substantial model size and extensive data samples required for training\, centralized model storing and training are nearly infeasible for small and intermediate service providers.\nConsequently\, we will adopt horizontal model partitioning and distribute different model layers across the network nodes through caching.\nData samples or batches are input into the network and undergo the forward-backward procedure for training. Our objective is to jointly optimize data forwarding and model/computation placement\, thereby minimizing the total cost of transmission\, computation\, and storage. \nFurthermore\, we introduce several network resource allocation optimization problems related to data-centric networking\, thereby expanding the scope of our proposal.
URL:https://coe.northeastern.edu/event/jinkun-zhang-phd-proposal-review/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230809T150000
DTEND;TZID=America/New_York:20230809T163000
DTSTAMP:20260520T023125
CREATED:20230731T152624Z
LAST-MODIFIED:20230807T134836Z
UID:37662-1691593200-1691598600@coe.northeastern.edu
SUMMARY:Yuanyuan Li PhD Dissertation Defense
DESCRIPTION:Title: Sub-modularity in Cache Networks \nCommittee Members:\nProf. Stratis Ioannidis\nProf. Lili Su\nProf. Edmund Yeh \nAbstract:\nAs information-based demand surges\, distributed network services\, e.g.\, cache networks\, play an important role to mitigate network traffic. Cache networks are a natural abstraction for many applications\, including information-centric networks\, content delivery networks\, cloud computing\, and edge/wireless IoT. How to allocate resources (routing\, placing items in caches\, flow control\, etc.) in cache networks is a crucial problem\, as resources (storage space\, and bandwidths) are usually limited. Resource allocation in networks has been traditionally approached through classic convex optimization. However\, simple problems becomes combinotorial in cache networks\, which leads to NP-hardness. Enlightened by several works studying cache networks\, we identify a useful property\, submodularity\, which is the key to approximation algorithms solving those NP hard resource allocation problem in cache networks. \nLeveraging submodularity\, we study a cache network\, in which intermediate nodes equipped with caches can serve content requests\, from different angles. \nFirst\, we model this network as a universally stable queuing system\, in which packets carrying identical responses are consolidated before being forwarded downstream. We refer to resulting queues as $\info$ or counting queues\, as consolidated packets carry a counter indicating the packet’s multiplicity. Cache networks comprising such queues are hard to analyze; we propose two approximations: one via $\mminf$ queues\, and one based on $\info$ queues under the assumption of Poisson arrivals. We show that\, in both cases\, the problem of jointly determining (a) content placements and (b) service rates admits a poly-time\, $1-1/e$ approximation algorithm. We also show that our analysis\, with respect to both algorithms and associated guarantees\, extends to (a) counting queues over items\, rather than responses\, as well as to (b) queuing at nodes and edges\, as opposed to just edges. \nSecond\, we refer to the cost reduction enabled by caching as the caching gain\, and the product of the caching gain of a content request and its request rate as \emph{caching gain rate}. We aim to study \emph{fair} content allocation strategies through a utility-driven framework\, where each request achieves a utility of its caching gain rate\, and consider a family of $\alpha$-fair utility functions to capture different degrees of fairness. The resulting problem is an NP-hard problem with a non-decreasing submodular objective function. Submodularity allows us to devise a deterministic allocation strategy with an optimality guarantee factor arbitrarily close to $1-1/e$.  When $0 < \alpha \leq 1$\, we further propose a randomized strategy that attains an improved optimality guarantee\,  $(1-1/e)^{1-\alpha}$\, in expectation. \nThird\, we study a cache network\, and model the problem of jointly optimizing caching and routing decisions with link capacity constraints over an arbitrary network topology. This problem can be formulated as a continuous diminishing-returns(DR) submodular maximization problem under multiple continuous DR-supermodular constraints\, and is NP-hard. We propose a poly-time alternating primal-dual  heuristic algorithm\, in which primal steps produce solutions within $1-\frac{1}{e}$ approximation factor from the optimal. Through extensive experiments\, we demonstrate that our proposed algorithm significantly outperforms competitors. \nForth\, we study a cache network under arbitrary adversarial request arrivals. We propose a distributed online policy based on the online tabular greedy algorithm. Our distributed policy achieves sublinear $(1-\frac{1}{e})$-regret\, also in the case when update costs cannot be neglected. \nFinally\, we propose an {\em experimental design network} paradigm\, wherein learner nodes train possibly different Bayesian linear regression models via consuming data streams generated by data source nodes over a network. We formulate this problem as a social welfare optimization problem in which the global objective is defined as the sum of experimental design objectives of individual learners\, and the decision variables are the data transmission strategies subject to network constraints. We first show that\, assuming Poisson data streams\, the global objective is a continuous DR-submodular function. We then propose a Frank-Wolfe type algorithm that outputs a solution within a $1-1/e$ factor from the optimal. Our algorithm contains a novel gradient estimation component which is carefully designed based on Poisson tail bounds and sampling.
URL:https://coe.northeastern.edu/event/yuanyuan-li-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230809T110000
DTEND;TZID=America/New_York:20230809T120000
DTSTAMP:20260520T023125
CREATED:20230802T192340Z
LAST-MODIFIED:20230802T192340Z
UID:37696-1691578800-1691582400@coe.northeastern.edu
SUMMARY:Nasim Soltani PhD Proposal
DESCRIPTION:Title: Deep Learning for the Physical Layer: From Signal Classification to Decoding \nLocation: ISEC 532 \nCommittee Members:\nProf. Kaushik Chowdhury (Advisor)\nProf. Stratis Ioannidis\nProf. Robert Nowak \nAbstract:\nThe growth in wireless spectrum usage has created new physical layer applications and intensified the importance of the existing ones. Physical layer applications ranging from device authentication to signal decoding and interpretation are traditionally handled by deterministic signal processing algorithms. Such algorithms\, while effective\, often require long sequences of data for decision making\, or need approximations of the environmental conditions\, such as noise models\, which may not be always correct in practical conditions. For these reasons\, traditional algorithms are not suitable for making quick decisions on the high rate wireless data with higher noise and interference that is a result of crowded spectrum. To this end\, deep learning-based methods have been explored extensively by the researchers to substitute for the traditional signal processing algorithms for the physical layer. This thesis explores novel methods in this area in the following parts: \nPart I – Signal classification: In this part\, we look at two distinct problems of waveform classification and Radio Frequency (RF) fingerprinting. In the first problem\, we study two use cases of modulation classification on edge devices\, followed by waveform classification and spectrum localization in the Citizen Broadband Radio Service (CBRS) band. In the second problem\, we look at RF fingerprinting that is classifying received signals in terms of subtle impairments that each transmitter leaves in its emitted waveform\, due to its hardware manufacturing imperfections. We propose methods to overcome the wireless channel effect for RF fingerprinting in both stationary transmitters on a large scale dataset (i.e.\, 5k WiFi devices)\, and identical hovering Unmanned Aerial Vehicles (UAVs) that transmit proprietary signals. \nPart II – Signal decoding: In this part\, we introduce our design of a modular machine learning (ML)-aided Orthogonal Frequency Division Multiplexing (OFDM) receiver that improves the bit error rate (BER) of the traditional receiver. We show how a neural network-based demapper block can be used for secure data transmission. Furthermore\, we show how an ML-aided receiver can provide the possibility of reducing communication overhead by obviating the need for the first field of preamble in WiFi signals. We show that reducing the preamble length contributes to higher throughput in WiFi networks\, without BER degradation. \nPart III – As the proposed work\, we will explore the use of active learning for smart sampling of training sets in wireless communications tasks. Active learning reduces the labeling overhead that is often performed using the compute-intensive traditional signal processing algorithms\, by intelligently selecting the most informative training samples to be labeled instead of labeling the whole set. We will also design an ML-life cycle control scheme to monitor and update the performance of an ML-aided 5G receiver\, when deployed in the field with varying environmental conditions.
URL:https://coe.northeastern.edu/event/nasim-soltani-phd-proposal/
LOCATION:532 ISEC\, 360 Huntington Ave\, Boston\, MA\, 02115\, United States
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230808T103000
DTEND;TZID=America/New_York:20230808T113000
DTSTAMP:20260520T023125
CREATED:20230802T192515Z
LAST-MODIFIED:20230802T192515Z
UID:37694-1691490600-1691494200@coe.northeastern.edu
SUMMARY:Dinesh Murugan MS Thesis Defense
DESCRIPTION:Title: Advances in Modelling\, Control\, and Perception for Soft Robotics and Autonomous Vehicle Systems \nLocation on Campus: Snell Room \nCommittee Members:\nAdvisor: Prof. Milad Siami\nProf. Bahram Shafai\nProf. Rozhin Hajian – University of Massachusetts\, Lowell \nAbstract:\nIn this research project\, we investigate the distributed consensus and vehicle platoons control problem. We first investigate the performance deterioration of commensurate fractional-order consensus networks under exogenous stochastic disturbances. We formulate fractional-order differential equations for the network dynamics using Caputo derivatives and the Laplace transform\, and employ the H_2 norm of the dynamical system as a performance measure. By developing a graph-theoretic methodology\, we relate the structural specifications of the underlying graphs to the performance measure and explicitly quantify fundamental limits on the best achievable levels of performance in fractional-order consensus networks. We also establish new connections between the sparsity of the network and the performance measure\, characterizing fundamental tradeoffs that reveal the interplay between the two. Finally\, we provide numerical illustrations to verify our theoretical results\, which could help in the design of robust fractional-order control systems in the presence of disturbances. \nAdditionally\, the study examines the real-time application of the theoretical advancements on Quanser’s Qcars\, a scaled model vehicle used for academic purposes. The findings are highly relevant to the design and implementation of large-scale consensus networks and autonomous vehicle platoons\, as they emphasize the importance of balancing network density and update cycle speed for optimal performance. \nTo extend the research’s findings to viscoelastic based networks\, the interaction between agents is modeled as a fractional-order system.
URL:https://coe.northeastern.edu/event/dinesh-murugan-ms-thesis-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230808T100000
DTEND;TZID=America/New_York:20230808T110000
DTSTAMP:20260520T023125
CREATED:20230721T142808Z
LAST-MODIFIED:20230721T142808Z
UID:37565-1691488800-1691492400@coe.northeastern.edu
SUMMARY:Yukui Luo's PhD Dissertation Defense
DESCRIPTION:Title:\nSecuring FPGA as a Shared Cloud-Computing Resource: Threats and Mitigations \nCommittee Members:\nProf. Xiaolin Xu (Advisor)\nProf. Yunsi Fei\nProf. Xue Lin \nAbstract:\nWith the widespread adoption of cloud computing\, the demand for programmable hardware acceleration devices\, such as field-programmable gate arrays (FPGA)\, has increased. These devices benefit the growth of efficient hardware accelerators\, making cloud computing possible for a wide range of research and commercial projects\, including genetic engineering\, intensive online secure trading\, the Artificial Intelligence (AI) interface\, etc. To further improve the performance of FPGA-enabled cloud computing\, one promising technology is to virtualize the hardware resources of an FPGA device\, which allows multiple users to share the same FPGA. This solution can provide on-demand FPGA instances\, significantly improving the hardware utilization and energy efficiency of the cloud FPGA. However\, due to the hardware reconfigurability of FPGA\, current virtualization technologies used for multi-tenant CPU and GPU instances are incompatible with multi-tenant FPGA. \nWe aim to enhance the security of multi-tenant FPGA by defining the threat model and evaluating security concerns from the perspectives of confidentiality\, data integrity\, and availability. As part of this goal\, we constructed multi-tenant FPGA prototypes and demonstrated potential attacks. These attacks serve as preliminary steps toward developing a secure multi-tenant FPGA virtualization system. This system involves hardware and software co-design\, which extends the multi-tenant isolation from software to hardware\, ultimately resulting in a secure FPGA shared cloud computing service.
URL:https://coe.northeastern.edu/event/yukui-luos-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230803T100000
DTEND;TZID=America/New_York:20230803T110000
DTSTAMP:20260520T023125
CREATED:20230508T153859Z
LAST-MODIFIED:20230508T153859Z
UID:36933-1691056800-1691060400@coe.northeastern.edu
SUMMARY:Yu Yin's PhD Dissertation Defense
DESCRIPTION:“Synthetic Data Generator: Understanding Human Face & Body via Image Synthesis” \nCommittee Members:\nProf. Yun Fu (Advisor)\nProf. Sarah Ostadabbas\nProf. Ming Shao \nAbstract:\nThe community has long enjoyed the benefits of synthesizing data\, as it provides a reliable and controllable source for training machine learning models while reducing the need for data collection from the real world. Human face and body synthesis are especially appealing to research communities\, where model fairness and ethical deployment are critical concerns. However\, generating digit humans that are convincing\, realistic-looking\, identity-preserving\, and high-quality are still challenging in 2D and 3D image synthesis. \nThis dissertation investigates the potential for understanding human behavior by recreating it\, and can be broadly divided into three sections. (1) In Section one\, we explore the 2D image generation models and their interaction with face applications (i.e.\, landmark localization and face recognition tasks). Specifically\, super-resolution (SR) and landmark localization of tiny faces are highly correlated tasks. To this end\, we propose joint frameworks that enable face alignment and SR to benefit from one another\, hence enhancing the performance of both tasks. Moreover\, we demonstrate that face frontalization provides an effective and efficient way for face data augmentation and further improves face recognition performance in extreme pose scenarios. (2) In Section two\, we explore the 3D parametric generation models and how they support human body pose and shape estimation. Advancing technology to monitor our bodies and behavior while sleeping and resting is essential for healthcare. However\, keen challenges arise from our tendency to rest under blankets. To mitigate the negative effects of blanket occlusion\, we use an attention-based restoration module to explicitly reduce the uncertainty of occluded parts by generating uncovered modalities\, which further update the current estimation via a cyclic fashion. (3) In Section three\, we explore the 3D Nerf-based Generative models in generating high-quality images with consistent 3D geometry. We propose a universal method to surgically fine-tune these NeRF-GAN models in order to achieve high-fidelity animation of real subjects only by a single image.
URL:https://coe.northeastern.edu/event/yu-yins-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230801T100000
DTEND;TZID=America/New_York:20230801T113000
DTSTAMP:20260520T023125
CREATED:20230721T143101Z
LAST-MODIFIED:20230721T143101Z
UID:37563-1690884000-1690889400@coe.northeastern.edu
SUMMARY:Huan Wang's PhD Proposal Review
DESCRIPTION:Title: \nTowards Efficient Deep Learning in Computer Vision via Sparsity and Distillation \nCommittee Members: \nProf. Yun Fu (Advisor) \nProf. Octavia Camps \nProf. Zhiqiang Tao \nAbstract: \nAI\, empowered by deep learning\, has been profoundly transforming the world. However\, the excessive size of these models remains a central obstacle that limits their broader utility. Modern neural networks commonly consist of millions of parameters\, with foundation models extending to billions. The rapid expansion in model size introduces many challenges including training cost\, sluggish inference speed\, excessive energy consumption\, and negative environmental implications such as increased CO2 emissions. \nAddressing these challenges necessitates the adoption of efficient deep learning. This thesis focuses on two overarching approaches\, network sparsity and knowledge distillation\, to enhance the efficiency of deep learning models in the context of computer vision. Network sparsity focuses on eliminating redundant parameters in a model while preserving the performance. Knowledge distillation aims to enhance the performance of the target model\, referred to as the “student\,” by leveraging guidance from a stronger model\, known as the “teacher”. This approach leads to performance improvements in the target model without reducing its size. In the proposal\, I will start with the background and major challenges of leveraging these techniques towards efficient deep learning. Then\, I shall present the potential solutions in various tasks (e.g.\, image classification\, image super-resolution\, neural rendering\, and text-to-image generation)\, with preliminary results to justify the efficacy of the proposed approaches. Finally\, a comprehensive outlook of the future work will conclude this proposal.
URL:https://coe.northeastern.edu/event/huan-wangs-phd-proposal-review/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230725T130000
DTEND;TZID=America/New_York:20230725T140000
DTSTAMP:20260520T023125
CREATED:20230721T142252Z
LAST-MODIFIED:20230721T142322Z
UID:37567-1690290000-1690293600@coe.northeastern.edu
SUMMARY:Batool Salehihikouei Phd Proposal Review
DESCRIPTION:Title:\nLeveraging Deep Learning on Multimodal Sensor Data for Wireless Communication: From mmWave Beamforming to Digital Twins \nCommittee Members:\nProf. Kaushik Chowdhury (Advisor)\nProf. Hanumant Singh\nProf. Josep Jornet\nDr. Mark Eisen \nAbstract:\nWith the widespread Internet of Things (IoT) devices\, a wide variety of sensors are now present in different environments. For example\, self-driving vehicles and automated warehouses depend on sensor information for navigation and management of the robots\, respectively. In this dissertation\, we present a paradigm\, where these sensors are re-purposed to assist network management in wireless communication\, especially when classic approaches fall short to provide the required quality of service (QoS). This thesis presents data-driven and AI-based methods\, where the multimodal sensor information is used for beamforming at the mmWave band\, and envisions a systematic framework for joint optimization of the navigation and network management in factory floor environments. In particular\, the contributions in this dissertation are as follows. First\, we present deep learning fusion algorithms\, where the inputs from a multitude of sensor modalities such as GPS (Global Positioning System)\, camera\, and LiDAR (Light Detection and Ranging) are combined towards predicting the optimum beam at the mmWave band. We prove that fusing the multimodal sensor data improves the prediction accuracy compared to using single modalities. Second\, we study the trade-off between the accuracy and cost of different learning strategies for multimodal beamforming. In this regard\, we make a case for using federated learning for beamforming at the mmWave band and demonstrate that it is the most successful learning strategy\, with respect to the communication overhead. Finally\, we take measures to further optimize the computation and communication overhead\, by incorporating a pruning strategy tailored to the disturbed nature of the federated learning systems. In the proposed research work\, we suggest using digital twins to overcome the challenges of scarcity of data and close-world assumption in deep learning algorithms. A digital twin is a replica of a real world entity\, which is typically used for studying the impact of any configuration settings in a safe\, digital environment. In this dissertation\, we propose using digital twins for generating training data for multimodal beamforming\, in unseen scenarios. Moreover\, we study a robotic industrial setting\, where the path planning policy is continuously updated by monitoring the dynamics of the real world\, constructing the digital twin\, and updating the policy.
URL:https://coe.northeastern.edu/event/batool-salehihikouei-phd-proposal-review/
LOCATION:532 ISEC\, 360 Huntington Ave\, Boston\, MA\, 02115\, United States
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230721T140000
DTEND;TZID=America/New_York:20230721T153000
DTSTAMP:20260520T023125
CREATED:20230718T135222Z
LAST-MODIFIED:20230718T135222Z
UID:37521-1689948000-1689953400@coe.northeastern.edu
SUMMARY:Daniel Uvaydov's PhD Dissertation Defense
DESCRIPTION:Title: Real-Time Spectrum Sensing for Inference and Control \nCommittee Members: \nProf. Tommaso Melodia (Advisor) \nProf. Kaushik Choudhury \nProf. Francesco Restuccia \nAbstract: \nThrough growing cellular innovations\, the usage and congestion of the wireless spectrum is increasing at incredible speeds. High demand and limited supply pose a resource issue known as the “spectrum crunch”. With the high diversity of users sharing a large portion of the spectrum to request and receive diverse services\, spectrum coordination becomes very difficult. Large scale device synchronization for spectrum coordination requires high overhead and more wireless transmissions further reducing spectrum resources. However\, by monitoring the spectrum\, otherwise known as spectrum sensing\, we can develop mechanisms where users can opportunistically take action based on the current state of the spectrum\, without need for direct coordination between devices. Spectrum sensing can enable the next generation of wireless applications ranging from opportunistic spectrum access to cognitive radio networks. The key unaddressed challenges of spectrum sensing are that (i) it requires very extensive and diverse datasets; (ii) it has to be performed with extremely low latency over varying bandwidths and must guarantee strict real-time processing constraints; (iii) its underlying algorithms need to be extremely accurate\, and flexible enough to work with different wireless bands and protocols to find application in real-world settings. This dissertation focuses on addressing these challenges in multiple wireless applications by utilizing Deep Learning (DL) techniques as the main vehicle of spectrum sensing for both inference and control. Algorithmic spectrum sensing has generally been model-based which limits its performance in diverse settings and environments\, for this reason we explore data-driven spectrum sensing algorithms. Mainly\, this work takes a holistic approach to address spectrum sensing problems from multiple directions with the overarching goal of developing the core building blocks for the next generation of intelligent\, AI-driven\, efficient spectrum sharing systems. By leveraging mechanisms such as data augmentation\, channel attention\, voting\, and segmentation we are able to push beyond the capabilities of existing DL techniques and create generalizable spectrum sensing algorithms. Furthermore we deploy different spectrum sensing solutions in real testbeds for over the air evaluations and applicable proof-of-concepts. The contributions of this work includes (i) multiple datasets and implementations for DL enabled spectrum sensing with applications in radio frequency and underwater; (ii) a method for tackling the core issue of dataset generation in supervised learning algorithms for spectrum sensing via a novel data augmentation technique; (iii) a study into one of the first ever semi-unsupervised approaches for wideband multi-class spectrum sensing.
URL:https://coe.northeastern.edu/event/daniel-uvaydovs-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:20230720T130000
DTEND;TZID=America/New_York:20230720T140000
DTSTAMP:20260520T023125
CREATED:20230711T140015Z
LAST-MODIFIED:20230711T140015Z
UID:37433-1689858000-1689861600@coe.northeastern.edu
SUMMARY:Qing Jin's PhD Dissertation Defense
DESCRIPTION:Title:Decoupling Efficiency-Performance Optimization for Modern Neural Networks \nDate: \n7/20/2023 \nCommittee Members: \nYanzhi Wang (Advisor); Prof. David Kaeli; Prof. Sunil Mittal; Prof. Jennifer Dy \nAbstract: \nDeep learning has achieved remarkable success in a variety of modern applications\, but this success is often accompanied by inefficiency in terms of storage and inference speed\, which can hinder their practical use on resource-constrained hardware. Developing highly efficient neural networks that maintain high prediction accuracy is crucial and challenging. This dissertation explores the potential for simultaneously achieving high efficiency and high prediction accuracy in neural networks\, and can be broadly divided into three sections. (1) In Section One\, we explore the implementation of highly efficient generative adversarial networks (GANs) capable of generating high-quality images within a predefined computational budget. The key challenge lies in identifying the optimal architecture for the generative model while simultaneously preserving the quality of the generated images from the compressed model\, despite its reduced computational cost. To achieve this\, we propose a novel neural architecture search (NAS) algorithm and a new knowledge distillation technique. (2) In Section Two\, we explore the challenge of quantizing discriminative models without relying on high-precision multiplications. To address this issue\, we present an innovative approach to determine the optimal fixed-point formats for both weights and activations based on their statistical properties. Our results demonstrate that high accuracy in quantized neural networks can be achieved without the need for high-precision multiplications. (3) In Section Three\, we delve into the challenge of training neural networks for innovative computing platforms\, specifically processing-in-memory (PIM) systems. Through a detailed mathematical derivation of the backward propagation algorithm\, we facilitate the training of quantized models on these platforms. Additionally\, through a thorough theoretical analysis of training dynamics\, we ensure convergence and propose a systematic solution for quantizing neural networks on PIM systems.
URL:https://coe.northeastern.edu/event/qing-jins-phd-dissertation-defense/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20230629T170000
DTEND;TZID=America/New_York:20230629T173000
DTSTAMP:20260520T023125
CREATED:20230626T173009Z
LAST-MODIFIED:20230626T173036Z
UID:37275-1688058000-1688059800@coe.northeastern.edu
SUMMARY:Zifeng Wang's PhD Dissertation Defense
DESCRIPTION:Title: Effective and Efficient Continual Learning \nCommittee Members:\nProf. Jennifer Dy (Advisor)\nProf. Stratis Ioannidis\nProf. Yanzhi Wang \nAbstract:\nContinual Learning (CL) aims to develop models that mimic the human ability to learn continually without forgetting knowledge acquired earlier. While traditional machine learning methods focus on learning with a certain dataset (task)\, CL methods adapt a single model to learn a sequence of tasks continually. \nIn this thesis\, we target developing effective and efficient CL methods under different challenging and resource-limited settings. Specifically\, we (1) leverage the idea of sparsity to achieve cost-effective CL\, (2) propose a novel prompting-based paradigm for parameter-efficient CL\, and (3) utilize task-invariant and task-specific knowledge to enhance existing CL methods in a general way. \nWe first introduce our sparsity-based CL methods. The first method\, Learn-Prune-Share (LPS)\, splits the network into task-specific partitions\, leading to no forgetting\, while maintaining memory efficiency. Moreover\, LPS integrates a novel selective knowledge sharing scheme\, enabling adaptive knowledge sharing in an end-to-end fashion. Taking a step further\, we present Sparse Continual Learning (SparCL)\, a novel framework that leverages sparsity to enable cost-effective continual learning on edge devices. SparCL achieves both training acceleration and accuracy preservation through the synergy of three aspects: weight sparsity\, data efficiency\, and gradient sparsity. \nSecondly\, we present a new paradigm\, prompting-based CL\, that aims to train a more succinct memory system that is both data and memory efficient. We first propose a method that learns to dynamically prompt (L2P) a pre-trained model to learn tasks sequentially under different task transitions\, where prompts are small learnable parameters maintained in a memory space. We then improve L2P by proposing DualPrompt\, which decouples prompts into complementary “General” and “Expert” prompts to learn task-invariant and task-specific instructions\, respectively. \nFinally\, we propose DualHSIC\, a simple and effective CL method that generalizes the idea of leveraging task-invariant and task-specific knowledge. DualHSIC consists of two complementary components that stem from the so-called Hilbert Schmidt independence criterion (HSIC): HSIC-Bottleneck for Rehearsal (HBR) lessens the inter-task interference and HSIC Alignment (HA) promotes task-invariant knowledge sharing. \nComprehensive experimental results demonstrate the effectiveness and efficiency of our methods over the state-of-the-art methods on multiple CL benchmarks.
URL:https://coe.northeastern.edu/event/zifeng-wangs-phd-dissertation-defense/
END:VEVENT
END:VCALENDAR