EE Seminar :Noise Agnostic Outlier Detection on Galaxy Spectra

סמינר זה יחשב כסמינר שמיעה לתלמידי תואר שני ושלישי

31 ביולי 2024, 15:30 
חדר 011, בניין כיתות-חשמל  
  EE Seminar :Noise Agnostic Outlier Detection on Galaxy Spectra

Electrical Engineering Systems Seminar

 

Speaker: Almog Hershko

M.Sc. student under the supervision of Prof. Dovi Poznanski

 

Wednesday, 31st July 2024, at 15:30

Room 011, Kitot Building, Faculty of Engineering

 

Noise Agnostic Outlier Detection on Galaxy Spectra

 

Abstract

The field of astronomy, like many other scientific fields, is deep inside the age of big data. Thanks to technological advancements in sensors and computers, numerous large astronomical datasets already exist containing billions of observations. One such dataset is the Sloan Digital Sky Survey (SDSS), an ongoing sky survey of more than 20 years that includes (among other data) several million galaxy spectra.

Naturally, data-driven algorithms play a key role in helping scientists extract new insights from these datasets. Modern learning algorithms can handle vast amounts of data, extract trends and direct scientists to uncover the underlying physics driving them.

Alternatively, a learning algorithm can detect outliers that stand out from the rest of the data due to some unique or rare physical phenomena. Outlier detectors are unsupervised learning algorithms, that require only data and no labels and produce a ranking of the data according to some learned outlier score. Such outlier detectors can point the attention of researchers to unique objects that potentially hold key to new insights.

A unique feature of astronomical datasets is the fact that the vast majority of the data is noisy. First, because the sources are intrinsically faint, and we cannot change that. Second, because most of the volume of the Universe is far from earth, and objects grow fainter the further away from us they are. Consequently, a survey down to some sensitivity limit will usually detect most of its sources near the largest distance it can reach. This means that machine learning tools for astronomical datasets should be more robust to noise than in other domains.

This thesis builds upon an existing outlier detector for galaxy spectra that is based on unsupervised random forest (URF). URF trained on SDSS spectra has been shown previously to produce meaningful outliers but is prone to false alarm due to low signalto-noise ratio (SNR). The proposed algorithm in this work tries to preserve the good performance for high SNR data while training it to be noise agnostic in addition, by combining RF distillation with denoising in the training process.

 

השתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בדף הנוכחות שיועבר באולם במהלך הסמינר

 

 

EE Seminar :SSemi-supervised channel equalization using variational autoencoders

סמינר זה יחשב כסמינר שמיעה לתלמידי תואר שני ושלישי

29 ביולי 2024, 15:30 
חדר 011, בניין כיתות-חשמל  
 EE Seminar :SSemi-supervised channel equalization using variational autoencoders

Electrical Engineering Systems Seminar

 

Speaker: Eli Bery

M.Sc. student under the supervision of Prof. David Burshtein

 

Monday, 29th July 2024, at 15:30

Room 011, Kitot Building, Faculty of Engineering

 

Semi-supervised channel equalization using variational autoencoders

 

Abstract

This research presents methods for semi-supervised learning (SSL) from few pilot signals over nonlinear channels, using variational autoencoders (VAEs). These channels, unknown to the receiver, may have finite memory (intersymbol interference), making traditional supervised learning approaches suboptimal.

Our SSL approach leverages both labeled pilot symbols and unlabeled payload symbols, significantly reducing the number of pilot symbols required for reliable channel inference compared to standard supervised learning methods. The research demonstrates that SSL with VAEs achieves superior performance, yielding a lower error rate and greater efficiency in symbol decoding. For sufficiently many payload symbols, the VAE also has a lower error rate compared to meta-learning that uses the pilot data of the present as well as previous transmission blocks. This advancement in deep learning for communications over unknown nonlinear channels highlights the potential of VAEs in optimizing decoding processes with minimal pilot data.

 

השתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בדף הנוכחות שיועבר באולם במהלך הסמינר

 

EE Seminar :Semi-Blind Separation of Complex-Valued Gaussian Sources

סמינר זה יחשב כסמינר שמיעה לתלמידי תואר שני ושלישי

29 ביולי 2024, 15:00 
חדר 011, בניין כיתות-חשמל  
EE Seminar :Semi-Blind Separation of Complex-Valued Gaussian Sources

Electrical Engineering Systems Seminar

 

Speaker: Roy Agmon

M.Sc. student under the supervision of Prof. Arie Yeredor

Monday, 29th July 2024, at 15:00

Room 011, Kitot Building, Faculty of Engineering

 

Semi-Blind Separation of Complex-Valued Gaussian Sources

 

Abstract

Blind Source Separation (BSS) is a well-known problem in signal processing which aims to recover unobserved statistically independent sources (signals) based on observations of their mixtures. The term ``blind" reflects the facts that the sources are not observed and that there is no prior information available about the mixture or about the distribution of each source. If prior information regarding the sources' joint distribution is available, the problem is termed ``semi blind". Independent Component Analysis (ICA) is a common technique to separate independent sources in a single set. If there are multiple sets such that each set contains independent sources but the sources are correlated across sets, the problem is termed Joint BSS (JBSS), or semi-blind JBSS, and the common separation technique is Independent Vector Analysis (IVA). As part of this work we show that using IVA we can exploit the correlation between sets in order to separate sources that are not necessarily separable in a single set using ICA. The quality of separation is quantified by the interference-to-signal ratio (ISR), which measures the residual energy of a source in the reconstruction of another source.

This work addresses the semi-blind JBSS of a particular type of sources, which are complex-valued and Gaussian distributed. Our interest in complex-valued sources requires us to address the two types of complex-valued distributions, “circular” and “non-circular”, and to examine their statistical properties, mainly their covariance and pseudo-covariance matrices. The prior knowledge regarding the sources' distributions, which are Gaussian with known (zero)-mean, covariance, and pseudo-covariance matrices, gives rise to a Maximum Likelihood Estimation (MLE) - based separation approach, which exploits the prior information regarding the sources' joint distribution.

We present the mathematical derivation of the MLE – based separation approach, including performance bounds analysis and a comparison of the resulting ISR to its lower bound, the induced Cramér Rao Lower Bound (iCRLB). We also demonstrate by our simulation results the ability to use IVA in order to separate sources that are not separable in a single set.

 

השתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בדף הנוכחות שיועבר באולם במהלך הסמינר

 

EE Seminar :Development of Deep Learning-Based Methods for Molecular Magnetic Resonance Imaging

סמינר זה יחשב כסמינר שמיעה לתלמידי תואר שני ושלישי

31 ביולי 2024, 15:00 
חדר 011, בניין כיתות-חשמל  
EE Seminar :Development of Deep Learning-Based Methods for Molecular Magnetic Resonance Imaging

 

Electrical Engineering Systems Seminar

 

 Speaker: Dinor Nagar - M.Sc. student under the supervision of Dr. Or Perlman

Wednesday, 31st July 2024, at 15:00

Room 011, Kitot Building, Faculty of Engineering

 

Development of Deep Learning-Based Methods for Molecular Magnetic Resonance Imaging

 

השתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בדף הנוכחות שיועבר באולם במהלך הסמינר

 

 

 

 

 

 

 

 

 

EE Seminar :NextStop: An improved tracker for panoptic LiDAR segmentation data

סמינר זה יחשב כסמינר שמיעה לתלמידי תואר שני ושלישי

28 ביולי 2024, 15:00 
חדר 011, בניין כיתות-חשמל  
EE Seminar :NextStop: An improved tracker for panoptic LiDAR segmentation data

Electrical Engineering Systems Seminar

 

Speaker: Nirit Alkalay

M.Sc. student under the supervision of Prof. Ben-Zion Bobrovsky

 

Sunday, 28th July 2024, at 15:00

Room 011, Kitot Building, Faculty of Engineering

 

NextStop: An improved tracker for panoptic LiDAR segmentation data

 

Abstract

4D panoptic LiDAR segmentation is essential for scene understanding in autonomous driving and robotics, combining semantic and instance segmentation of LiDAR point clouds with temporal consistency. Current approaches, like 4D-PLS and 4D-STOP, use a tracking-by-detection methodology, employing deep learning networks to perform semantic and instance segmentation on each frame. To maintain temporal consistency, large-size instances detected in the current frame are compared and associated with instances within a temporal window that includes the current and preceding frames.

The above mentioned association method rely on short-term instance detection within the temporal window, lack motion estimation capabilities, and exclude small-sized  from matching, leading to frequent identity switches and reduced tracking performance.

To address these limitations, we introduce the NextStop1 tracker. Our tracker combines Kalman filter-based motion estimation, data association, and lifespan management modules, along with a tracklet state concept for prioritization. Additionally, we leverage accumulated tracking data to correct temporal inconsistencies in semantic segmentation results.

We evaluated our tracking method using the LiDAR Segmentation and Tracking Quality (LSTQ) metric, proposed by Aygun et al., on the validation set of SemanticKITTI.

NextStop showed improvements in this metric for various classes such as Other-vehicles, People, and Cars. The advantages of our tracking method lie primarily in tracking small size objects, including small-sized classes like People and Bicyclist, as well as objects from other classes that are at a distance and therefore considered small-sized. These improvements are reflected in fewer ID switches, earlier tracking initiation, and more reliable tracking in complex environments.

 

השתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בדף הנוכחות שיועבר באולם במהלך הסמינר

 

 

 

 

 

 

 

 

 

יום פרויקטים של המחלקה להנדסה תעשייה וניהול

05 באוגוסט 2024, 9:30 - 16:00 
וולפסון  
יום פרויקטים של המחלקה להנדסה תעשייה וניהול

יום פרוייקטים במחלקה להנדסת תעשייה- שמרו את התאריך ה05.08.24

הנכם.ן מוזמנים.ות ליום פרוייקטים בהנדסה תעשייה שיערך ביום שלישי ה05.08.24 ויתקיים בפקולטה להנדסה בוולפסון ובכיתות.

 

נשמח לראותכם בין אורחינו,

 

ארגון עמיתי התעשייה והמחלקה תעשייה מהפקולטה להנדסה. 

 

לינק להזמנה

 

Sub-mm-Wave Silicon Based Source Array

סמינר זה יחשב כסמינר שמיעה לתלמידי תואר שני ושלישי

31 ביולי 2024, 8:30 
ZOOM  
Sub-mm-Wave Silicon Based Source Array

 

-This seminar is considered a hearing seminar for Msc.Phd students-

 

 

You are invited to attend a Zoom Seminar on Wednesday, July 31, 2024  at 08:30

 

Join our Cloud HD Video Meeting

tau-ac-il.zoom.us

 

Title: Sub-mm-Wave Silicon Based Source Arrays

 

Prof. M.-C. Frank ChangUniversity of California, Los Angeles, USA

Prof. Richard Al HadiUniversité du Québec, ÉTS, Montreal, Canada

 

 

Abstract

 This talk presents an overview of research on the design and implementation of a sub-mm-Wave integrated signal generator in silicon technology. The presentation will cover the design methodology, electromagnetic modeling, simulations, and measurement results developed over recent years. A novel beamforming technique, called Y-vector network, will be introduced. This technique minimizes chip area and DC power consumption and eliminates the need for traditional adjustable phase shifters at multi-channel front ends, which typically suffer from high insertion loss at these frequencies. A compact 1x4 Beam Steering Phased Array (BSPA) is validated at 0.55THz with an approximately ±30° steering angle range. Additionally, the talk will present a generator array with a frequency stabilization loop, based on a multi-radiator PLL (MR-PLL) harmonic oscillator designed in standard 65-nm CMOS technology. Achieving the ability to lock multiple frontend oscillators to a common frequency is a challenging design requirement, ensuring precise timing, synchronization, and coherent signal transmission across the chip. The designed MR-PLL addresses this critical requirement, providing a reliable implementation at the targeted mm-Wave frequencies (540-560GHz).

Bio

 Dr. M. C. Frank Chang is the Wintek Chair in Electrical Engineering and Distinguished Professor at UCLA.  Throughout his career, he has focused on the research & development of high-speed semiconductor devices and integrated circuits for radio, radar, imager, spectrometer, and interconnect System-on-Chip applications. He invented the multiband, reconfigurable RF-Interconnects for Chip-Multi-Processor (CMP) inter-core communications and inter-chip CPU-to-Memory communications. He and his students were the 1st to demonstrate CMOS active and passive imagers at 100-180GHz. His Lab also pioneered the development of self-healing 57-64GHz radio-on-a-chip (DARPA’s HEALICS program) with embedded sensors, actuators and self-diagnosis/curing capabilities; and invented the Digitally Controlled Artificial Dielectric (DiCAD) embedded in CMOS technologies to vary transmission-line permittivity in real-time (up to 20X in practice) for realizing reconfigurable multiband/mode radios in (sub)-mm-Wave bands. His UCLA Lab also realized the first CMOS Frequency Synthesizer for Terahertz operation (PLL at 560GHz). More recently, his Lab has devised a Reconfigurable Convolution Neuron Network (RCNN) Accelerator for AIoT applications, spun-off an Edge-AI company Kneron in San Diego, and won IEEE’s 2021 Darlington Best Paper Award. He is a Member of the US National Academy of Engineering. He was also recognized by the IEEE David Sarnoff Award (2006), IET JJ Thomson Medal for Electronics (2017), and IEEE/RSE (Royal Society Edinburgh) James Clerk Maxwell Medal (2023) for his seminal contributions to the heterojunction technology and realizations of (sub)-mm-Wave System-on-Chip with unprecedented bandwidth and re-configurability.

Dr. Richard Al Hadi is an Associate Professor of Electrical Engineering at Université du Québec, ÉTS, Montreal. He received the engineering diploma from Caen's National Graduate School of Engineering in Electronics and Applied Physics and the master’s degree from the University of Caen, France, in 2009. He received the Ph.D. degree, summa cum laude, from the University of Wuppertal, Germany, in 2014. Dr. Al Hadi joined University of California, Los Angeles (UCLA) in 2015 as a postdoctoral research fellow. He was leading the effort at Alcatera Inc, a technology focus company, between 2017-2022. Dr. Al Hadi is a senior IEEE society member and is the co-recipient of the 2012 Jan Van Vessem Award for the Outstanding European Paper at the IEEE International Solid-State Circuit Conference and the 2014 EuCAP best paper award.

כמו שהבנתם.ן, אנחנו ממש heartאוהבים.ות heart טקסים כאן בפקולטה. 

למה? סמול-טוק עם אנשים מעניינים, אוכל טעים ולצפות במהנדסים ומהנדסות מוכשרים במיוחד, בתחילת דרכם המקצועית angel

 לינק להורדת כל התמונות - לחצו כאן 

 

*הלינק תקף לשבוע ימים, אז מהרו להוריד (:  

יום פרויקטים המחלקה להנדסת תעשייה

05 באוגוסט 2024, 14:00 - 18:00 
 
יום פרויקטים המחלקה להנדסת תעשייה

14:00-15:20- מושב ראשון

בניין וולפסון: מעבדה 455 - CIM, מעבדת 424, אולם 020

בניין כיתות: אולם 011

15:20-15:30- הפסקה

15:30-16:50 - מושב שני

בניין וולפסון: מעבדה 455 - CIM, מעבדת 424, אולם 020

בניין כיתות: אולם 011

17:00-18:00- מליאה - פאנל בוגרים ודברי ברכה

דקאן הפקולטה להנדסה, פרופ’ נעם אליעז

ראש המחלקה להנדסת תעשייה, פרופ’ ערן טוך

אולם טאו - בניין וולפסון

 

ד"ר ברכה לאופר

עמודים

אוניברסיטת תל אביב עושה כל מאמץ לכבד זכויות יוצרים. אם בבעלותך זכויות יוצרים בתכנים שנמצאים פה ו/או השימוש שנעשה בתכנים אלה לדעתך מפר זכויות
שנעשה בתכנים אלה לדעתך מפר זכויות נא לפנות בהקדם לכתובת שכאן >>