Category: DataS

1010 with … Nathir Rawashdeh, Weds., Dec. 16

Nathir Rawashdeh (right) and Dan Fuhrmann, Interim Dean, Dept. of Applied Computing

You are invited to spend one-zero-one-zero—that is, ten—minutes with Dr. Nathir Rawashdeh on Wednesday, December 16, from 5:30 to 5:40 p.m.

Rawashdeh is assistant professor of applied computing in the College of Computing at Michigan Tech.

He will present his current research work, including the using artificial intelligence for autonomous driving on snow covered roads, and a mobile robot using ultraviolet light to disinfect indoor spaces. Following, Rawashdeh will field listener questions.

We look forward to spending 1010 minutes with you!

Did you miss last week’s 1010 with Chuck Wallace? Watch the video below.

The 1010 with … series will continue on Wednesday afternoons in the new year on January 6, 13, 20, and 27 … with more to come!

Siva Kakula to Present PhD Defense Dec. 21, 3 pm

Graduate student Siva Krishna Kakula, Computer Science, will present his PhD defense, “Explainable Feature- and Decision-Level Fusion,” on Monday, December 21, 2020, from 3:00 to 5:00 p.m. EST Kakula is advised by Dr. Timothy Havens, College of Computing.

Siva Kakula earned his master of science in computer engineering at Michigan Tech in 2014, and completed a bachelor of technology in civil engineering at IIT Guwahati in 2011. His research interests include machine learning, pattern recognition, and information fusion.

Download the informational flier below.

Lecture Abstract

Information fusion is the process of aggregating knowledge from multiple data sources to produce more consistent, accurate, and useful information than any one individual source can provide. In general, there are three primary sources of data/information: humans, algorithms, and sensors. Typically, objective data—e.g., measurements—arise from sensors. Using these data sources, applications such as computer vision and remote sensing have long been applying fusion at different “levels” (signal, feature, decision, etc.). Furthermore, the daily advancement in engineering technologies like smart cars, which operate in complex and dynamic environments using multiple sensors, are raising both the demand for and complexity of fusion. There is a great need to discover new theories to combine and analyze heterogeneous data arising from one or more sources.

The work collected in this dissertation addresses the problem of feature- and decision-level fusion. Specifically, this work focuses on Fuzzy Choquet Integral (ChI)-based data fusion methods. Most mathematical approaches for data fusion have focused on combining inputs relative to the assumption of independence between them. However, often there are rich interactions (e.g., correlations) between inputs that should be exploited. The ChI is a powerful aggregation tool that is capable modeling these interactions. Consider the fusion of N sources, where there are 2N unique subsets (interactions); the ChI is capable of learning the worth of each of these possible source subsets. However, the complexity of fuzzy integral-based methods grows quickly, as the fusion of N sources requires training 2N-2 parameters; hence, we require a large amount of training data to avoid the problem of over-fitting. This work addresses the over-fitting problem of ChI-based data fusion with novel regularization strategies. These regularization strategies alleviate the issue of over-fitting while training with limited data and also enable the user to consciously push the learned methods to take a predefined, or perhaps known, structure. Also, the existing methods for training the ChI for decision- and feature-level data fusion involve quadratic programming (QP)-based learning approaches that are exorbitant with their space complexity. This has limited the practical application of ChI-based data fusion methods to six or fewer input sources. This work introduces an online training algorithm for learning ChI. The online method is an iterative gradient descent approach that processes one observation at a time, enabling the applicability of ChI-based data fusion on higher dimensional data sets.

In many real-world data fusion applications, it is imperative to have an explanation or interpretation. This may include providing information on what was learned, what is the worth of individual sources, why a decision was reached, what evidence process(es) were used, and what confidence does the system have on its decision. However, most existing machine learning solutions for data fusion are “black boxes,” e.g., deep learning. In this work, we designed methods and metrics that help with answering these questions of interpretation, and we also developed visualization methods that help users better understand the machine learning solution and its behavior for different instances of data.

Sidike Paheding Lecture is Dec. 11, 3 pm

Assistant Professor Sidike Paheding, Applied Computing, will present his lecture, “Deep Neural Networks for UAV and Satellite Remote Sensing Image Analysis,” on Dec. 11, 2020, at 3:00 p.m. via online meeting.

Paheding’s research focuses on the areas of computer vision, machine learning, deep learning, image/video processing, and remote sensing.

The lecture is presented by the Department of Computer Science.

Lecture Abstract

Remote sensing data can provide non-destructive and instantaneous estimates of the earth’s surface over a large area, and has been accepted as a valuable tool for agriculture, weather, forestry, defense, biodiversity, etc. In recent years, deep neural networks (DNN), as a subset of machine learning. for remote sensing has gained significant interest due to advances in algorithm development, computing power, and sensor systems.

This talk will start with remote sensing image enhancement framework, and then primarily focuses on DNN architectures for crop yield prediction and heterogeneous agricultural landscape mapping using UAV and satellite imagery.

Speaker Biography

Paheding is an associate editor of the Springer journal Signal, Image, and Video Processing, ASPRS Journal Photogrammetric Engineering & Remote Sensing, and serves as a guest editor/reviewer for a number of reputed journals. He has advised students at undergraduate, M.S., and Ph.D. levels, and authored/coauthored close to 100 research articles.

Sangyoon Han to Present Chemistry Seminar this Friday, Nov. 13, at 3 pm

A Chemistry Seminar will be presented Friday, September 13, 2020, at 3:00 p.m., via online meeting.

Dr. Sangyoon Han will present his lecture, “Toward Discovery of the Initial Stiffness-Sensing Mechanism by Adherent Cells.” Han is an Assistant Professor in Biomedical Engineering, an Affiliate Assistant Professor in Mechanical Engineering-Engineering Mechanics, and advisor for the Korean Student Association. Han is a member of the ICC’s Center for Data Science.

Lecture Abstract

The stiffness of the extracellular matrix (ECM) determines nearly every aspect of cellular/tissue development and contributes to metastasis of cancer. Adherent cells’ stiffness-sensing of the ECM triggers intracellular signaling that can affect proliferation, differentiation, and migration of the cells. However, biomechanical and molecular mechanisms behind this stiffness sensing have been largely unclear. One critical early event during the stiff-sensing is believed to be a force transmission through integrin-based adhesions, changing the molecular conformation of the molecules comprising the adhesions that link the ECM to the cytoskeleton. To understand this force transmission, my lab develops experimental and computational techniques, which include soft-gel-based substrates, live-cell imaging, computer-vision-based analysis, and inverse mechanics, etc. In this talk, I will talk about how we use soft-gel to quantify the spatial distribution of mechanical force transmitted by a cell, how we use light microscopy and computer vision to analyze the focal adhesions, and how these techniques are related to stiffness sensing. In particular, I will show you new data where cells can transmit different levels of traction forces in response to varying stiffness, even when the activity of the major motor protein, myosin, is inhibited. At the end of the talk, potential molecules responsible for the differential transmission will be discussed. 

Researcher Bio

Sangyoon Han received his Ph.D. in Mechanical Engineering at the University of Washington (UW) in 2012 and did postdoctoral training with Dr. Gaudenz Danuser in the Department of Cell Biology at Harvard Medical School and the University of Texas Southwestern Medical Center for five years until 2017. Before the Ph.D., he received B.S and M.S. degree from Mechanical Engineering at Seoul National University, Seoul, Korea in 2002 and 2004.

He joined Michigan Tech, Biomedical Engineering from fall 2017, and started Mechanobiology Laboratory. His lab’s interests are in understanding the dynamic nature of force modulation occurring across cell adhesions and cytoskeleton that regulate cells’ environmental sensing. His lab develops a minimally-perturbing experimental approach and computational techniques, including soft-gel fabrication, nano-mechanical tools, live-cell microscopy, and image data modeling, to capture the coupling between force modulation and cellular molecular dynamics.

Sidike Paheding Publishes Paper in Expert Systems and Applications Journal

A research paper by Assistant Professor Sidike Paheding, Applied Computing, is to be published in the November 2020 issue of the journal, Expert Systems and Applications.

An in-press version of the paper, “Binary Chemical Reaction Optimization based Feature Selection Techniques for Machine Learning Classification Problems,” is available online.


  • A chemical reaction optimization (CRO) based feature selection (FS) technique is proposed.
  • The proposed CRO based FS technique is improvised using particle swarm optimization.
  • Performance evaluation of proposed techniques on benchmark datasets gives promising results.

Paper Abstract

Feature selection is an important pre-processing technique for dimensionality reduction of high-dimensional data in machine learning (ML) field. In this paper, we propose a binary chemical reaction optimization (BCRO) and a hybrid binary chemical reaction optimization-binary particle swarm optimization (HBCRO-BPSO) based feature selection techniques to optimize the number of selected features and improve the classification accuracy.

Three objective functions have been used for the proposed feature selection techniques to compare their performances with a BPSO and advanced binary ant colony optimization (ABACO) along with an implemented GA based feature selection approach called as binary genetic algorithm (BGA). Five ML algorithms including K-nearest neighbor (KNN), logistic regression, Naïve Bayes, decision tree, and random forest are considered for classification tasks.

Experimental results tested on eleven benchmark datasets from UCI ML repository show that the proposed HBCRO-BPSO algorithm improves the average percentage of reduction in features (APRF) and average percentage of improvement in accuracy (APIA) by 5.01% and 3.83%, respectively over the existing BPSO based feature selection method; 4.58% and 3.12% over BGA; and 4.15% and 2.27% over ABACO when used with a KNN classifier.

Expert Systems With Applications, published by Science Direct/Elsevier, is a refereed international journal whose focus is on exchanging information relating to expert and intelligent systems applied in industry, government, and universities worldwide. The journal’s Impact factor is 5.4.

Innovative, Active, Effective. Introducing Sidike Paheding, Applied Computing

Be Innovative. Be Active. Be Effective. This is College of Computing Assistant Professor Sidike Paheding’s teaching philosophy.

New to the Department of Applied Computing this fall, Paheding’s teaching interests include digital image processing and machine learning. This academic year he is teaching SAT3812 Cyber Security I.

A member of the Institute of Computing and Cybersystems’s Center for Data Sciences, Paheding’s research seeks to develop novel AI-driven technologies. His primary interests are image/video processing, machine learning, deep learning, computer vision, and remote sensing.

Paheding comes to Michigan Tech from Purdue University Northwest, where he was a visiting assistant professor in the ECE department. Prior to that, he was a postdoctoral research associate and assistant research professor in the Remote Sensing Lab at Saint Louis University from 2017 to 2019.

Paheding is an associate editor of the journals, Signal Image and Video Processing (Springer) and Photogrammetric Engineering and Remote Sensing (ASPRS), and topic editor for Remote Sensing. He completed his Ph.D. in electrical engineering at University of Dayton, Ohio.

Computing is a part of my life.

Sidike Paheding, Assistant Professor, Applied Computing

Active Research

Title: Cybersecurity Modules Aligned with Undergraduate Computer Science and Engineering Curricula
Sponsor: NSF
PI at Michigan Tech
Duration: July 2020 – June 2022
Total Award: $159,417.00

Research Abstract

This project aims to serve the national interest by improving how cybersecurity concepts are taught in undergraduate computing curricula. The need to design and maintain cyber-secure computing systems is increasingly important. As a result, the future technology workforce must be trained to have a security mindset, so that they consider cybersecurity during rather than after system design.

This project aims to achieve this goal by building plug-and-play, hands-on cybersecurity modules for core courses in Computer Engineering, and Computer Science and Engineering. The modules will align with the curricula recommended by the Association for Computing Machinery and will be designed for easy adoption into computing programs nationwide. Modules will be designed for integration into both introductory and advanced courses, thus helping students develop in-depth understanding of cybersecurity as they progress through their computing curriculum. It is expected that the project will encourage more students to pursue careers or higher degrees in the field of cybersecurity.

Recent Publications

Sidike, P., Sagan, V., Maimaitijiang, M., Maimaitiyiming, M., Shakoor, N., Burken, J., … & Fritschi, F. B. (2019). dPEN: deep Progressively Expanded Neural Network for mapping heterogeneous agricultural landscape using WorldView-3 satellite imagery. Remote Sensing of Environment, 221, 756-772. [Impact Factor: 9.085]

Sidike, P., Asari, V. K., & Sagan, V. (2018). Progressively Expanded Neural Network (PEN Net) for hyperspectral image classification: A new neural network paradigm for remote sensing image analysis. ISPRS journal of photogrammetry and remote sensing, 146, 161-181. [Impact Factor: 7.319]

Sidike, P., Asari, V. K., & Alam, M. S. (2015). Multiclass object detection with single query in hyperspectral imagery using class-associative spectral fringe-adjusted joint transform correlation. IEEE Transactions on Geoscience and Remote Sensing, 54(2), 1196-1208. [Impact Factor: 5.855]

Maimaitijiang, M., Sagan, V., Sidike, P., Hartling, S., Esposito, F., & Fritschi, F. B. (2020). Soybean yield prediction from UAV using multimodal data fusion and deep learning. Remote Sensing of Environment, 237, 111599. [Impact Factor: 9.085]

Tim Havens: Warm and Fuzzy Machine Learning

What are you doing for supper this Monday night at 6? Grab a bite with Dean Janet Callahan and Associate Professor Tim Havens, director of the Michigan Tech’s Institute of Computing and Cybersystems and associate dean for research in the College of Computing. Get the full scoop and register at

“Nearly everyone has heard the term ‘Deep Learning’ at this point, whether to describe the latest artificial intelligence feat like AlphaGo, autonomous cars, facial recognition, or numerous other latest-and-greatest gadgets and gizmos,” says Havens. “But what is Deep Learning? How does it work? What can it really do—and how are Michigan Tech students advancing the state-of-the-art?”

In this session of Husky Bites, Prof. Havens will talk about everyday uses of machine learning—including the machine learning research going on in his lab: explosive hazards detection, under-ice acoustics detection and classification, social network analysis, connected vehicle distributed sensing, and other stuff.

Joining in will be one of Havens’ former students, Hanieh Deilamsalehy, who earned her PhD in electrical engineering at Michigan Tech. She’s now a machine learning researcher at Adobe. Dr. Deilamsalehy graduated from Michigan Tech in 2017 and headed to Palo Alto to work for Ford as an autonomous vehicle researcher. She left the Bay Area for Seattle to take a job at Microsoft, first as a software engineer, and then as a machine learning scientist. In April she accepted a new machine learning position at Adobe, “in the middle of the pandemic!”

Havens is a Michigan Tech alum, too. He earned his BS in ‘99 and MS in Electrical Engineering in ‘00, then went to the MIT Lincoln Laboratory, where he worked on simulation and modeling of the Airborne Laser System, among other defense-related projects. From there it was the University of Missouri for a PhD in Electrical and Computer Engineering, researching machine learning in ontologies and relational data.

Nowadays, Havens is the William and Gloria Jackson Associate Professor and Associate Dean for Research in the College of Computing. In addition to serving as director of Michigan Tech’s ICC, he also heads up the ICC Center for Data Sciences and runs his own PRIME Lab, too (short for Pattern Recognition and Intelligent Machines Engineering).

“An important goal for many mobile platforms—terrestrial, aquatic, or airborne—is reliable, accurate, and on-time sensing of the world around them.”Tim Havens

Havens has spent the past 12 years developing methods to find explosive hazards, working with the US Army and a research team in his lab. According to a United Nations report, more than 10,000 civilians were killed or injured in armed conflict in Afghanistan in 2019, with improvised explosive devices used in 42 percent of the casualties. Havens is working to help reduce the numbers.

“Our algorithms detect and locate explosive hazards using two different systems: a vehicle-mounted multi-band ground-penetrating radar system and a handheld multimodal sensor system,” Havens explains. “Each of these systems employs multiple sensors, including different frequencies of ground penetrating radar, magnetometers and visible-spectrum cameras. We’ve created methods of integrating the sensor information to automatically find the explosive hazards.” 

As a PhD student at Michigan Tech, Deilamsalehy worked alongside Havens as a research assistant in the ECE department’s Intelligent Robotics Lab (IRLab). “My research was focused on sensor fusion, machine learning and computer vision, fusing the data from IMU, LiDAR, and a vision camera for 3D localization and mapping purposes,” she says. “I used data from a sensor platform in the IRLab, mounted on an unmanned aerial vehicle (UAV), to evaluate my proposed fusion algorithm.”

Havens is also co-advisor to students in the SENSE (Strategic Education through Naval Systems Experience) Enterprise team at Michigan Tech, along with ME-EM Professor Andrew Barnard. Students in SENSE design, build, and test engineering systems in all domains: space, air, land, sea, and undersea. Like all Enterprise teams, SENSE is open to students in any major. 

Prof. Havens, when did you first get into engineering? What sparked your interest?

I first became an engineer at Michigan Tech in the late 90s. What really sparked my interest in what-I-do-now was my introductory signal processing courses. The material in these courses was the first stuff that really ‘spoke’ to me. I have always been a serious musician and the mathematics of waves and filters was so intuitive because of my music knowledge. I loved that this field of study joined together the two things that I really loved: music and math. And I’ve always been a computer geek. I was doing programming work in high school to make extra money; so that side of me has always led me to want to solve problems with computers.

Hometown, Hobbies, Family?

I grew up in Traverse City, Michigan, and came to Tech as a student in the late 90s. I’ve always wanted to come back to the Copper Country; so, it’s great that I was able to return to the institution that gave me the jump start in my career. I live (and currently work from home) in Hancock with my partner, Dr. Stephanie Carpenter (an author and MTU professor), and our two fur children, Rick Slade, the cutest ginger in the entire world, and Jaco, the smartest cat in the entire world. I have a grown son, Sage, who enjoys a fast-paced life in Traverse City. Steph and I enjoy exploring the greater Keweenaw and long discussions about reality television, and I enjoy playing music with all the local talent, fishing (though catching is a challenge), and gradually working through the lumber pile in my garage.

Dr. Deilamsalehy, how did you find engineering? What sparked your interest?

I was born and raised in Tehran, Iran. I have always been into robotics. I was a member of our robotics team in high school and that led me to engineering. I decided to apply to Michigan Tech sort of by chance when a friend of mine told me about it. I looked at the programs in the ECE department, and felt they aligned with my interests. Then soon after I first learned about Michigan Tech, I found out that one of my undergraduate classmates went there. I talked to him, and he also encouraged me to apply. And that’s how I was able to join Michigan Tech for my PhD program. My degree is in electrical engineering but my focus at Michigan Tech involved computer science and designing Machine Learning solutions.

Hobbies and Interests?

I now live in Seattle, famous for outdoor activities—kind of like the UP, but without the cold—so I do lots of mountaineering, biking, rock climbing, and in the winter, skiing. I learned how to ski at Michigan Tech, up on Mont Ripley. It’s steep, and it’s cold! Once you learn skiing on Ripley, you’re good. You can ski just about anywhere.

Tim Schulz to Present Michigan Tech Research Forum Oct. 14

Timothy Schulz

University Professor Timothy Schulz (ECE) will be featured at the Michigan Tech Research Forum (MTRF) at 4:30 p.m. Wednesday, Oct. 14.
Schulz’s presentation is titled “Direct Measurement of Coherent Fields.” Additional details can be found on the MTRF website.

The presentation will be available via Zoom and a limited number of people will be permitted to attend in person, dependent on university guidelines on the date of the event. If you wish to be considered for in-person attendance, complete this form by today (Oct. 9).

Schulz is a member of the ICC’s Center for Data Sciences.

The MTRF is presented by the Office of the Provost in coordination with the Office of the Vice President for Research. The forum showcases and celebrates the work of Michigan Tech researchers and aims to strengthen discussions in our community. All are welcome, including the general public.

Paper by Yakov Nekrich Accepted for ACM-SIAM SODA21 Symposium

A paper by Associate Professor Yakov Nekrich, Computer Science, has been accepted for the 61st ACM-SIAM Symposium on Discrete Algorithms 2021 (SODA21), which will take place virtually January 10-13, 2021.

Nekrich is sole author of the accepted article, “New Data Structures for Orthogonal Range Reporting and Range Minima Queries.” An extended version of the paper is available for download on ArXiv.

The annual ACM-SIAM Symposium on Discrete Algorithms (SODA) is an academic conference in the fields of algorithm design and discrete mathematics. It is considered among the top conferences for research in algorithms.

Paper Abstract

In this paper we present new data structures for two extensively studied variants of the orthogonal range searching problem.
First, we describe a data structure that supports two-dimensional orthogonal range minima queries in O(n) space and O(logεn) time, where n is the number of points in the data structure and ε is an arbitrarily small positive constant. Previously known linear-space solutions for this problem require O(log1+εn) (Chazelle, 1988) or O(lognloglogn) time (Farzan et al., 2012). A modification of our data structure uses space O(nloglogn) and supports range minima queries in time O(loglogn). Both results can be extended to support three-dimensional five-sided reporting queries.

Next, we turn to the four-dimensional orthogonal range reporting problem and present a data structure that answers queries in optimal O(logn/loglogn+k) time, where k is the number of points in the answer. This is the first data structure that achieves the optimal query time for this problem. Our results are obtained by exploiting the properties of three-dimensional shallow cuttings.

The Society for Industrial and Applied Mathematics (SIAM) is an international community of 14,500+ individual members. Almost 500 academic, manufacturing, research and development, service and consulting organizations, government, and military organizations worldwide are institutional members.