Hooman Vaseli
Title: ProtoASNet: Dynamic Prototypes for Inherently Interpretable and Uncertainty-Aware Aortic Stenosis Classification in Echocardiography
Abstract: The work is about a novel classification model for ultrasound video data, that is inherently interpretable by design, thus more trustworthy, and can flag a case if the input data is noisy and can lead to uncertain prediction.
Bio: [Seeking Internship] Hooman Vaseli is a PhD candidate at ECE department, in Robotics and Control Lab (RCL), researching deep learning in computer vision and natural language processing for cardiac disease diagnosis. His focus is mostly towards developing models that are interpretable by design. He has industry research experience as an intern at Hitachi Energy. He got his Bachelor’s from UBC ECE as well, in electrical engineering with biomedical specialization. He also enjoys athletic activities, like skiing in winters, and soccer or cycling in summer.
Andrew Gunter
Title: Reformulating FPGA Routability Prediction with Machine Learning
Abstract: This work is applying Machine Learning (ML) to alleviate time wasted in the Field-Programmable Gate Array (FPGA) circuit design compilation flow. We apply ML techniques to predict whether design compilation will be successful and, if so, how long compilation will take. This allows FPGA Engineers to quickly terminate doomed compilation runs instead of wasting time by finishing compilation and receiving a failed design implementation. We have successfully applied our techniques to the routing phase of design compilation and are working to extend these ideas to the full compilation flow.
Bio: Andrew Gunter is a UBC ECE PhD student researching applications of machine learning to electronic design automation. Andrew’s research has won best paper awards at FPL 2018 and FCCM 2023. He is currently working part-time at a local Vancouver startup, Singulos Research, along with ECE faculty members Dr. Brad Quinton and Dr. Scott Chin where together they are researching applications of real-time computer vision on mobile hardware for augmented reality.
Andrew Musgrave
Title: Load Coincidence Factors for Robust Optimization of Distributed Energy Resources
Abstract: Addressing the need to optimize utilization of distributed energy resources (DERs) in the face of network constraints and considerable uncertainty in uncontrollable generation and demand, we present a robust optimal power flow problem to establish DER setpoints which remain feasible for any realization of renewable generation and conventional loads. Leveraging load coincidence factors to define a polytopic uncertainty set, we derive closed-form expressions for worst-case voltages enabling efficient solution of the robust optimal power flow problem.
Bio: Andrew Musgrave is a M.A.Sc student in the department of Electrical and Computer Engineering at UBC, returning to studies following two years of industry experience as a hardware designer. He completed his undergraduate degree in Engineering Physics (Minor in Mathematics) at Carleton University. He has broad interests in power systems and clean energy and is researching optimization of distributed energy resources and electricity markets. Andrew is seeking opportunities related to power systems following expected graduation in April 2024.
Kithmin Wickremasinghe
Title: A New On-Chip Silicon Photonic Microfluidic Thermal Flow-rate Sensors (MTFS) and a Compact Automated Microfluidic Control Systems
Abstract: In this work, a novel silicon photonic microfluidic flor rate sensor (MTFS) is presented, which can be integrated on-chip to measure the fluid flow rate within the microfluidic channels carrying low-flow liquids in a non-intrusive manner. The MTFS comprises of a microheater and two micro-ring resonators (MRRs), which act as temperature sensors that are located equidistantly upstream and downstream from the microheater. The sensor chip was fabricated using the SiEPIC Program EBeam PDK for the UBC ZEP process. Further, a standalone compact fluid delivery system is designed to accommodate multiple reagent delivery for diagnostic immunoassays. This system features 2 channels with 8 reservoirs per channel that can be simultaneously controlled and automated to deliver a sequence of reagents required for an immunoassay. The entire system has been designed compactly with a focus on modularity.
Bio: Kithmin Wickremasinghe is a third year Masters Student, currently enrolled in the MASc in Electrical and Computer Engineering (ECE) and a Research Assistant in the System-on-a-chip (SoC) Group and BioSensors Group, co-supervised by Prof. Sudip Shekhar and Prof. Lukas Chrostowski. His research interests include On-chip Silicon Photonic Sensors and Point-of-Care Diagnostics. He is looking for employment for August 2024.
Md Nazmul Hasan
Title: Antennas and Smart Surfaces for Next Generation Wireless Communication
Abstract: Full duplex(FD) is a major breakthrough that allows doubling the throughput and spectral efficiency of the link but it comes with major challenge of self-interference (SI). SI is the impact of the transmitted signal on its own receiver in the full-duplex system. Another challenge in mmwave is to achieve reconfigurability of antennas with low loss. Varactor and PIN diodes perform poorly due to high loss in mmwave. Therefore, new techniques are required in mmwave to obtain reconfigurable antennas. My research work involves designing FD antenna with techniques that can minimize SI and designing mmwave antenna systems with low loss tunable dielectric liquid crystal.
Bio: Md Nazmul Hasan is a PhD candidate in ECE department at UBC Vancouver. He is a current recipient of Four Year Fellowship (FYF). His research interests include microwave and mm-wave antennas and smart surfaces.
Sean Lam
Title: Dynamic Electro-Optic Analog Memory for Neuromorphic Photonic Computing
Abstract: In neuromorphic photonic computing, photonic devices require analog control, meaning digital-to-analog converters (DAC) and analog-to-digital converters (ADC) are needed to interface with these devices to perform inference and training. However, data movement from memory through DACs and ADCs in traditional von Neumann computing systems consumes energy and limits computational bandwidth. Therefore, analog memory collocated with photonic computing devices is proposed to reduce the need for DACs and ADCs and reduce data movement to improve compute efficiency. This paper demonstrates a monolithically integrated neuromorphic photonic circuit with collocated capacitive analog memory and compares various analog memory technologies for neuromorphic photonic computing on the MNIST dataset.
Bio: Sean Lam is a MASc Electrical Engineering Student specializing in silicon photonics and is co-supervised by Lukas Chrostowski and Sudip Shekar. Sean graduated from a BASc in Electrical and Computer Engineering from UBC in 2021 and started his MASc immediately after. He has been designing silicon photonic devices and circuits in his MASc, taping out in several foundries including Advanced Microelectronics Foundry, Global Foundries, and Applied Nanotools. Sean’s research is targeted toward silicon photonic device design automation, and he has worked on designing photonic ring modulators, adiabatic tapers, and interlayer transition devices.
Zahra Jamalouei
Title: Integrated frequency swept lasers (FSLs)
Abstract: Integrated frequency swept lasers (FSLs) play a crucial role in LiDAR (Light Detection And Ranging) technology. FSLs that are currently available are off-chip, making them bulky, sensitive to vibration, and expensive. The goal of this research is to build an on-chip FSL which is considerably cheaper, has high accuracy which results in longer visible depth and precise imaging and long life-time as the entire system is on-chip, significantly reducing points of failure.
Bio: Zahra Jamalouei is currently pursuing her second master’s degree at ECE UBC, where she is a member of Sudip’s group. Her research focuses on on-chip electro-optic systems. She earned her first master’s degree in digital systems at the University of Tehran in Iran.
Joel Nider
Title: Offloading the file system from the CPU
Abstract: File systems help us organize data, but require the CPU to access the files. What do we do when other devices (e.g. GPUs) want to use files? What do we do when we have a system that does not have a CPU? It is time to revisit our design.
Bio: Joel works on operating systems and related technologies. He has worked on projects in processing-in-memory, virtualization, file systems and drivers. Before coming to UBC, he work at IBM Research in Haifa, Israel. At UBC, he is contemplating what is required to build an operating system from autonomous devices that does not require a CPU.
Mateo Rendon
Title: Design of aging and wear-out sensors on a novel 12nm FinFET technology
Abstract: Wear-out and aging in transistors have become a pressing issue in modern technologies. Performance and area are no longer the only constraints, since now concepts such as right-to-repair, biomedical devices, and Internet of Things have brought forward the importance of a reliable and long-lasting operation. Several mechanisms threaten transistor integrity and the Ivanov SoC lab has dedicated efforts to developing and improving novel sensors designed for these specific aging processes. This research poster will present the main sensor topologies designed to read HCI, BTI, TDDB, EM, and SILC effects on a 12nm FinFET platform. The ultimate goal of the research project is to create an core IP capable of outputting information relevant to the temporal aging and wear-out of a chip.
Bio: Mateo Rendon is a second year MASc student at the Ivanov SoC Lab. His main research interests are the design of analog and digital circuits targeted at the study of transistor reliability. His accomplishments include the tapeout of a 12nm FinFET GlobalFoundries chip in which he was the main chip designer. Mateo is seeking employment for his graduation in 2024.
Pegah Tekieh
Title: Exploring the Attributes of Parametric Circuits
Abstract: The emergence of parametric excitation in the electrical domain traces its roots back to the late 19th century. Initially recognized for its application in low-noise amplification, parametric circuits can expand their scope to include mm-Wave frequency division. This presentation delves into the characteristics of parametric circuits, shedding light on their distinctive advantages and drawbacks when compared with alternative amplifiers and frequency dividers.
Bio: Pegah received her B.Eng in Electrical Engineering from the Isfahan University of Technology in 2020. She is currently working towards her M.A.Sc degree in Electrical Engineering at the University of British Columbia, with a primary research focus on investigating the attributes of parametric circuits for communication applications.
Mohammad Hossein Olyaiy
Title: Shaheen: Combining Homomorphic Encryption with Network Sparsity for Efficient Private Inference
Abstract: Trained deep learning models constitute valuable intellectual property, and many entities employ the Inference-as-a-Service (IaaS) model to serve clients through public cloud infrastructure without disclosing their models. However, this requires the user to send data to the provider — a significant challenge for private and legally protected data. Cryptographic solutions like homomorphic encryption (HE) allow clients to retain data privacy by allowing service providers to evaluate models on encrypted inputs. However, they dramatically increase computation, memory, and communication requirements. We introduce optimization at the system and algorithm level to make privacy preserving inference in the cloud more efficient.
Bio: Mohammad is a PhD candidate in ECE. His recent work focuses on efficient and cost-effective privacy preserving machine learning systems. Prior to this, he worked on efficient execution of ML workloads on hardware accelerators.
Nima Nasiri
Title: FeMux: A Flexible Serverless Lifetime Management Framework
Abstract: Our poster will be on a recent submission titled “FeMux: A Flexible Serverless Lifetime Management Framework”. Prior work on serverless resource management scales predictively or reactively scales resources based on incoming function invocations. In designing their systems, two areas of improvement are in (i) using the same metrics for all components of their system (e.g., cold starts and wasted resources), and (ii) tailoring their predictive mechanisms based on the characteristics of individual functions. We propose using a Representative Unified Metric (RUM) to tune and evaluate all system components based on the metrics of interest, which also enables the system to be adaptive to several sets of metrics simultaneously. We implement FeMux, a serverless lifetime management system which uses an ensemble of forecasters to predict incoming invocations, and is implemented and evaluated using RUM.
Bio: Nima Nasiri and Nalin Munshi are upper year M.Apsc students with an interest in resource management for serverless systems.
Nikhil Pratap Ghanathe
Title: T-RECX: Tiny-Resource Efficient Convolutional neural networks with early-eXit
Abstract: T-RECX is an early-exit network architecture optimized for tinyML models running on ultra-low-power devices. T-RECX achieves an average of 30% reduction in execution time by skipping several layers of computations for easy-to-classify inputs.
Bio: Nikhil P Ghanathe is a 5th year PhD student, working with Prof Steve Wilton. He completed his Masters from University of Florida in 2016 and went on to work at CERN and Microsoft Research before his PhD. His current interests involve optimizing/deploying/monitoring tiny ML algorithms on ultra-low-power devices.
Arshia Moghimi, Praveen Gupta
Title: Serverless need not be Security-less
Abstract: Serverless computing offers a lot of benefits: no servers to manage, only paying for your use, auto scaling, etc. As a result, it is growing in popularity everyday. An often neglected aspect of serverless applications is their security. The highly distributed nature of serverless exacerbates the issue. We propose a low overhead serverless security framework to help developers write secure applications and comply with data and processing regulations.
Bio: Arshia and Praveen are second-year Masters’ students with an interest in serverless applications and how to secure them.
Victoria Wu (with Andrea Fung)
Title: A Transformer-based Approach to Pretraining and Representation Learning of Echocardiography Reports
Abstract: Echocardiography reports contain rich patient information, and can often be organized into tabular measurements and findings. Current state-of-the-art approaches for tabular data involve traditional machine learning methods such as random forest or XGBoost; however, these methods do not learn a representation of the data, and instead perform supervised learning tasks. In this research, we explore using transformers to learn a representation of the tabular echo report, which can be used in a variety of downstream tasks.
Bio: Victoria Wu is a second year PhD student in Electrical and Computer Engineering, with a research focus on multi-modal deep learning methods for echocardiography. She is currently working on using LLMs and vision transformer models to improve echocardiography workflow. Her previous background is as a software engineer at Microsoft, and is currently seeking internship experiences.
Andrea Fung is a second year PhD student in the School of Biomedical Computing. With a background in life sciences and experimental medicine, her research focuses on applications of deep learning methods such as LLMs and vision transformers in clinical settings.
Bo Chen
Title: Frequency Dynamics-aware Real-time Marginal Pricing of Electricity Under Uncertainty
Bio: Bo Chen is currently working toward the Ph.D. degree. His research interests include power system operation and electricity markets.
Farhad Abbasi
Title: Interleaved CLLC Converters with Dual Phase-Shift Modulation
Abstract: Component mismatch in interleaved CLLC converters causes uneven current sharing and thermal distribution, where phase shift modulation is presented to balance currents in interleaved cells.
Bio: Farhad Abbasi was born in Tabriz, Iran. He received the B.Sc. degree in electrical engineering from Amirkabir University of Technology, Tehran, Iran, in 2013, and the M.Sc. degree in electrical engineering from University of Tabriz, Tabriz, Iran, in 2016. He is currently working toward the Ph.D. degree in electrical engineering with the University of British Columbia, Vancouver, BC, Canada.
He has been also a Research Scholar with Delta-Q Technologies, Burnaby, BC, Canada, since 2020. His current research interests include high-power bidirectional DC–DC resonant converters for battery chargers and renewable energy applications. He is seeking for internship and employment opportunities.