Asif Salekin

Assistant Professor
Department of Electrical Engineering and Computer Science
Syracuse University

asif asalekin-at-syr-dot-edu

asif    asif



asif

Bio

I am an experimental computer scientist. My research takes a multi-disciplinary approach to develop novel and practical human behavioral event sensing technologies that capture observable low-level physical signals from human bodies and surrounding environments and employ new machine learning, signal processing and natural language processing techniques, to rectify the existing sensing technologies. My research exquisition goes beyond the conventional learning or sensing approaches and addresses the research challenges, such as, the uncertainties in physical world sensing, human factors such as the user-context and mobility, limitation of current technologies (i.e., IoT, CPS), and resource constraints of the sensing data and computation platform. My research lies at the intersection of machine learning, internet of things (IoT), and natural language processing. Technologies and Systems that I have developed are human-centric, several of them are attributed to health and wellness, and in general, they are in the scope of ubiquitous computing.

Research Interests

Pervasive and Ubiquitous Computing, Internet of Things (IoT), Machine Learning, Cyber Physical Systems (CPS), Wireless, Connected and Mobile Health

Latest News

May 2019: I will join the Department of Electrical Engineering and Computer Science at Syracuse University as a Tenure-track Assistant Professor from Fall 2019.

I am looking for motivated Ph.D. students to work on developing automated human sensing technologies in the scope of health-care and health assessment by leveraging novel machine learning, NLP and IoT approaches. Interested students should send their curriculum vitae, transcripts, and a 600-word personal statement (detailing research interests, projects, publications, etc.) via email, with subject as 'Ph.D. admission at Syracuse University'.



Education


  • PhD in the Department of Computer Science

    University of Virginia

    Advisor: Professor John A. Stankovic

    Graduation: June 2019

    DIPLOMA VALIDATION:

    CeDiD: 19MA-CKEH-AKNA

    First two letters of the name: AS

    Validation site: http://www.virginia.edu/registrar/cedipl-validate.html

  • Masters in the Department of Computer Science

    University of Virginia

    Advisor: Professor John A. Stankovic

    Graduation: May 2016

    DIPLOMA VALIDATION:

    CeDiD: 16AB-A21E-A8N3

    First two letters of the name: AS

    Validation site: http://www.virginia.edu/registrar/cedipl-validate.html

  • BSc in the Department of Computer Science and Engineering

    Bangladesh University of Engineering and Technology

    Graduation: April 2012



Experience


  • Research Intern

    BHAG Realization Lab

    Nokia Bell Labs, Murray hill, New Jersey, USA

    June 2018 - Aug 2018

  • Research Intern

    Human-Machine Interaction Group, RTC 1.2

    BOSCH Research and Technology Center, California, USA

    May 2017 - Oct 2017

  • Graduate Research Assistant

    University of Virginia

    Advisor: Professor John A. Stankovic

    May 2014 – Present

  • Graduate Teaching Assistant

    University of Virginia

    Aug 2013 – May 2014

    Hold office hours and graded homework and exams for over 300 students in Course: Algorithms.

  • Lecturer

    Ahsanullah University of Science and Technology (AUST)

    Oct 2012 – Aug 2013

    Instructed undergraduate courses including: Algorithms, Network Programming, Operating Systems.

  • Lecturer

    BRAC University (BRACU)

    May 2012 – Oct 2012

    Instructed undergraduate courses including: Operating Systems, Introduction to Programming Language: Java.



Awards


  • Graduate Student Award for Outstanding Research

    Department of Computer Science, UVA, 2018

  • Nominated for best paper award (AsthmaGuide)

    Wireless health 2016

  • Third Annual Public Days Showcase Event (2016).

    The Public Days showcase highlights exemplary scholarship, research, and creative work of the University’s undergraduate and graduate students, as well as post-docs. Our project, AsthmaGuide, was selected to represent the highest achievements of scholarship, research, and creative work from undergraduate and graduate students across Grounds

  • SenSys 2015 Student Grant

    U.S. National Science Foundation (NSF)

  • Wireless Health 2014 Travel Grant

    National Institutes of Health (NIH)

  • Dean’s List for excellent performance in an academic year

    2009-2012

  • BUET Academic Merit List Scholarship for excellent result

    In all terms, 2007-2012

  • Dhaka Education Board Scholarship for H.S.C. result

    2007


Contributions to Funded Research


  • NSF Smart and Connected Health grant, 2018

    Award Number: 1838615

    Amount: $1,200,000

  • DGIST Research and Development Program (CPS Global center)

    Funded by the Ministry of Science, ICT and Future Planning, 2016

    Amount: $180,000




Publication

Peer Reviewed Full Papers | Published

  • A. Salekin, S. Ghaffarzadegan, Z. Feng, and J. Stankovic. A Real-Time Audio Monitoring Framework with Limited Data for Constrained Devices, The 15th International Conference on Distributed Computing in Sensor Systems (DCOSS 2019). [pdf]

  • A. Salekin, Jeremy W. Eberle, Jeffrey J. Glenn, Bethany A. Teachman, and John A. Stankovic. 2018. A Weakly Supervised Learning Framework for Detecting Social Anxiety and Depression, ACM Interactive, Mobile, Wearable, and Ubiquitous Technologies (IMWUT), Vol. 2, No. 2, Article 81 (June 2018), 26 pages. (and Ubicomp 2018) [pdf]

  • A. Salekin, Z. Chen, M. Ahmed, J. Lach, D. Metz, K. de la Haye, B. Bell, and J. Stankovic, Distance Emotion Recognition, ACM Interactive, Mobile, Wearable, and Ubiquitous Technologies (IMWUT), Vol. 1, Issue 3, Sept. 2017, 96:1-96:24 (Ubicomp 2017) [pdf]

  • A. Salekin, H. Wang, K. Williams, and J. Stankovic, DAVE: Detecting Agitated Vocal Events, The IEEE/ACM International Conference on Connected Health: Applications, Systems and Engineering Technologies (CHASE), July 2017. [pdf]

  • A. Salekin, and J. Stankovic, Detection of chronic kidney disease and selecting important predictive attributes, IEEE International Conference on Healthcare Informatics (ICHI), 2016 [pdf]

  • A. Salekin, Ho-Kyeong Ra, Hee Jung Yoon, Jeremy Kim, Shahriar Nirjon, David J. Stone, Sujeong Kim, Jong-Myung Lee, Sang Hyuk Son, and John A. Stankovic, AsthmaGuide: an asthma monitoring and advice ecosystem, IEEE Wireless Health 2016 [pdf]

  • Z. Chen, M. Ahmed, A. Salekin, and John A. Stankovic, ARASID: Artificial Reverberation-Adjusted Indoor Speaker Identification Dealing with Variable Distances, International Conference on Embedded Wireless Systems and Networks (EWSN), 2019.

  • S. Nirjon, I. Emi, A. Mondol, A. Salekin, and J. Stankovic, MOBI-COG: A Mobile Application for Instant Screening of Dementia Using the Mini-COG Test, Wireless Health, Oct. 2014. [pdf]

  • A. Salekin , J. Tabassum, and M. Hasan, Extract and rank web communities, Proceedings of the 3rd International Conference on Web Intelligence, Mining and Semantics (WIMS), 2013 [pdf]

  • MD. Rahman, S. Rahman, R. Hasan, R. Noel, A. Salekin, A Novel Approach for Constructing Emulator for Microsoft Kinect XBOX 360 Sensor in the .NET Platform, International Conference on Intelligent Systems, Modelling and Simulation (ISMS), 2013 [pdf]

  • MD. Rahman, MD. Rahman, A. Salekin, S.H. Chowdhury, S.A. Anik, A Novel Clustering-Based Ensemble Classification Model for Block Learning, International Conference on Pattern Recognition Applications and Methods (ICPRAM), 2013 [pdf]

  • MD. Rahman, MD. Rahman, A. Salekin, A.S. Andalib, A Novel Approach for Generating Clustered Based Ensemble of Classifiers, International Journal of Machine Learning and Computing, Vol. 3, Issue. 1, Page 137, 2013 [pdf]

  • A. Salekin, M. Islam, MD. Rahman, Composite Pattern Matching in Time Series, International Conference on Computer and Information Technology (ICCIT), 2012 [pdf]

  • A. Salekin, MD. Rahman, Pattern matching in time series using combination of neural network and rule based approach, International Conference on Electrical and Computer Engineering (ICECE), 2012 [pdf]

  • A.S. Andalib, A. Salekin, M.R. Islam, M. Abdulla-Al-Shami, Novel approaches for detecting fabric fault using artificial neural network with k-fold validation, International Conference on Computer and Information Technology (ICCIT), 2012 [pdf]


Patent

  • AudioEmergency: Audio Event Detection for Surveillance Systems (submitted)

    A. Salekin, Shabnam Ghaffarzadegan, Zhe Feng


First Author Full Papers | in Submission

  • Sensor Data Integration: A Multidimensional Constrained MIL approach.

Posters, Demos, Workshops, and Newsletters

  • A. Salekin, Jeremy W. Eberle, Jeffrey J. Glenn, Bethany A. Teachman, and John A. Stankovic, I Can Hear it in Your Voice: A Weakly Supervised Machine Learning Framework for Detecting Social Anxiety and Depression Symptoms from Features of Speech, ABCT's 52nd Annual Convention, 2018

  • A. Mondol, H. Ra, A. Salekin, H. Yoon, M. Kubovy, S. Son, J. Stankovic, Poster Abstract: LifeMaps - An Automated Diary System Based on the Structure of Lives, Sensys, 2016

  • Ho-Kyeong Ra, Hee Jung Yoon, A. Salekin, Jin-Hee Lee, John A Stankovic, Sang Hyuk Son, Poster Abstract: Software architecture for efficiently designing cloud applications using node.js, 14th Annual International Conference on Mobile Systems, Applications, and Services Companion, 2016

  • A. Salekin, H. Wang, J. Stankovic, Demo Abstract: KinVocal: Detecting Agitated Vocal Events, SenSys, 2015

  • Ho-Kyeong Ra, A. Salekin, Hee Jung Yoon, Jeremy Kim, Shahriar Nirjon, David J. Stone, Sujeong Kim, Jong-Myung Lee, Sang Hyuk Son and John A. Stankovic, Demo Abstract: AsthmaGuide: An Ecosystem for Asthma Monitoring and Advice, SenSys, 2015

  • R.R. Noel, A. Salekin, R. Islam, S. Rahaman, R. Hasan, H.S. Ferdous, A natural user interface classroom based on Kinect, IEEE Learning Technology Newsletter, Volume 13, October 2011


Teaching Experience


  • Invited lecturer: Two lectures on Machine learning for IoT and CPS

    Course: The Internet of Trillions of Things (Graduate level), UVA, Fall 2018

  • Invited lecturer: A lecture on Smart Connected Health

    Course: Wireless Sensor Networks (Undergraduate level), UVA, Fall 2014

  • Mentoring (Graduate level)

    Two first year PhD students (2015, 2017)

    Three Masters students (2014, 2016, 2018)

  • Graduate Teaching Assistant

    Course: Algorithm

    UVA, Fall, 2013 and Spring, 2014

    Hold office hours and graded homework and exams for over 300 students

  • Lecturer

    Ahsanullah University of Science and Technology (AUST), 2012-2013

    Courses: Algorithm, Network Programming, Operating Systems.

    Responsibility: lecture planning, taught and instructed courses, assessing students, holding office hours, invigilating examinations

  • Lecturer

    BRAC University (BRACU), 2012

    Courses: Operating Systems, Introduction to Programming Language: Java.

    Responsibility: lecture planning, taught and instructed courses, assessing students, holding office hours, invigilating examinations



Invited Talks


  • Machine Learning for Constrained Devices with Limited Training Data

    International Workshop on NEXT-GENERATION CYBER-PHYSICAL SYSTEMS, 2018

  • Invited talk: 'Human Machine Interaction', BR Lab, Nokia Bell Labs, NJ, USA, 2018

  • Invited talk: 'Machine Learning for IOT and CPS', ENSA Lab, Nokia Bell Labs, NJ, USA, 2018

  • Full Paper Presentation (Project: DER), Ubicomp 2017

  • Full Paper Presentation (Project: DAVE), IEEE CHASE, July 2017

  • Invited talk: 'Novel Feature Modeling for Audio Analytics', Human Machine Interaction Lab, BOSCH Research and Technology Center, CA, USA, 2017

  • Poster Presentation: 'LifeMaps - An Automated Diary System Based on the Structure of Lives', Sensys, 2016

  • The Public Days showcase event, UVA (Project: AsthmaGuide and KinVocal), 2016

  • Full Paper Presentation (Project: AsthmaGuide), IEEE Wireless Health 2016

  • Full Paper Presentation (Project: Chronic kidney disease detection), IEEE ICHI, 2016

  • Poster and Demo Presentation (KinVocal: Detecting Agitated Vocal Events), SenSys, 2015

  • Poster and Demo Presentation (AsthmaGuide: An Ecosystem for Asthma Monitoring and Advice), SenSys, 2015

  • UVA Open House, 2014, 2015, 2016, 2017



Skills


  • Programming Language

    Java, Python, C, C++, C#, Android development, Assembly Language (Intel 80x86, MIPS), UNIX shell scripting, SQL, LATEX, TEX

  • Deep Learning

    Tensorflow, keras

  • Natural Language Processing (NLP)

    Apache OpenNLP, gensim, Natural language toolkit (NLTK)

  • Scientific Computing

    Matlab, Octave, R, WEKA

  • Internet of Things (IoT)

    Raspberry pi 3B, UP Board, MATRIX Creator, Sony SmartWatch, Myo Armband, Mycroft.ai

  • Modeling Language

    UML, E-R Diagram

  • Web Programming

    HTML, CSS, PHP, JSP, JavaScript, Ajax, Jquery

  • Database Management

    Oracle, MySQL, JSON, MongoDB

  • Others

    OpenGL, Unix Shell Programming, Verilog HDL, Quartus, Cisco Packet Tracer



SELECTED COURSES


  • UNIVERSITY OF VIRGINIA

    Machine Learning, Spec Top: Computer Science (Machine Learning), Text Mining, Information Retrieval, Theory of Computation, Engineering Logic, Statistics Engrs \& Scientists, The Computational Planet, Cyber Physical Systems, Big Data in Health Research, Smart Cities, Homes, Phones, and Beyond (Seminar Course)

  • BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY

    Artificial Intelligence, Pattern Recognition, Introduction to machine learning, Algorithm, Data Structures, Database, Compiler, Structured Programming Language, Object Oriented Programming Language, Digital System Design, Discrete Mathematics, Concrete Mathematics, Computer Networks, Computer Graphics, Software Engineering and Information System Design, Mathematical Analysis for Computer Science, Theory of Computation

  • HIGH-PERFORMANCE COMPUTING BOOTCAMP (2016)



Referee/Reviewer


  • IMWUT (UBICOMP) 2017, 2018, 2019

  • IFIP PERFORMANCE 2018

  • ISSRE 2018

  • DSN 2019

  • ACM Transactions on Cyber-Physical Systems 2019



Volunteering Experience


  • Student Volunteer

    Wireless Health 2014

    UVA Engineering Alumni Reunions 2014

    Hosting Faculty Candidates (2015 & 2016)

    BUET CSE Festival (2008 & 2011)

    Bangladesh National Math Olympiad 2008

  • Co-founder and General Secretary

    Association of Bangladeshi Students, UVA, 2016-2017

  • Community Action

    Mentored two underprivileged students in their studies, 2014-2016



Selected Projects

Anxiety And Depression:

A Weakly Supervised Learning Framework For Detecting Social Anxiety And Depression


angdep

Mental health problems, such as depression and social anxiety disorder, are often under-diagnosed and under-treated, in part due to difficulties identifying and accessing individuals in need of services. One challenge is requiring individuals to self-report their symptoms, which depressed and anxious individuals may not be motivated to do. These detection methods are also vulnerable to social desirability and other subjective biases. Identifying objective, non-burdensome markers of depression and social anxiety, such as features of speech, could help advance assessment, prevention, and treatment approaches. Prior research examining speech detection methods has focused on fully supervised learning approaches employing strongly labeled data. However, strong labeling of persons high in symptoms or state affect in speech audio data is impractical, in part because it is not possible to identify with high confidence which regions of a long speech indicate the person's symptoms or critical affective state. We propose a weakly supervised deep learning framework for detecting social anxiety and depression from long audio clips. Specifically, we present a novel feature modeling technique named NN2Vec, which identifies and exploits the inherent relationship between speakers' vocal states and symptoms/affective states. In addition, we present a new multiple instance learning adaptation of a BLSTM classifier, named BLSTM-MIL. Our novel framework of using NN2Vec features with the BLSTM-MIL classifier achieves significant higher F-1 scores in detecting speakers high in social anxiety and depression symptoms.

View details »




Sensor Data Integration


Conventional multimodality sensor data integration approaches concatenate features from different sensory streams (from a detection window) and feed the concatenated feature-set to a supervised learning classifier. Since different sensors perceive only a sub-part of an event, a significant part of the sensory signals (from each devices) in an event detection window, contain irrelevant (noisy) information, which makes supervised learning classifiers perform poorly. I have developed a novel multidimensional constrained multiple instance learning neural network approach to integrate information at the decision-level from multi-modality multi-point wearable sensing data and applied on automated human activity detection task. This approach considers each of the sensory streams (signals from sensor devices) in an activity detection window as weakly labeled and extracts knowledge through an integrated weakly supervised learning approach from these combined streams. Conventional multiple instance learning (a form of weakly supervised learning) assumption does not entirely hold up to the characteristics of the sensory streams in human activity event detection windows. Hence, we developed a novel Constrained Multiple Instance Learning approach, incorporating the attributes of sensory stream data to perform weakly supervised learning for the targeted task.

View details »

bds



DER:

Distant emotion recognition


DERPIC

Distant emotion recognition (DER) extends the application of speech emotion recognition to the very challenging situation, that is determined by the variable, speaker to microphone distance. The performance of conventional emotion recognition systems degrades dramatically, as soon as the microphone is moved away from the mouth of the speaker. This is due to a broad variety of effects, such as, background noise, feature distortion with distance, overlapping speech from other speakers, and reverberation. I developed a novel solution for DER, addressing the key challenges by identification and deletion of features from consideration which are significantly distorted by distance, creating a novel feature modeling (knowledge engineering and representation) technique, called Emo2vec and overlapping speech filtering technique, and the use of an LSTM (deep learning) classifier to capture the temporal dynamics of speech states found in emotions. A comprehensive evaluation is conducted on two acted datasets (with artificially generated distance effect), as well as on a new emotional dataset of 12 spontaneous family discussions, with audio recorded from multiple microphones placed in different distances.

View details »




DAVE (KinVocal):

Detecting Agitated Vocal Events


DAVE

A system that continuously monitors and detects agitated vocal events, which is useful for the elderly population suffering from dementia. DAVE, using a novel combination of acoustic signal processing and multiple text mining techniques, automatically detects and records the 8 major vocal agitations for dementia patients as defined by the medical community. This includes cursing or verbal aggression, constant unwarranted request for attention or help, negativism, making verbal sexual advances, crying, screaming, laughing, and talking with repetitive sentences. The novelty of DAVE includes the comprehensiveness of addressing all 8 vocal events, using the text of the vocalizations only when accurate, combining text and acoustic features when necessary, and employing text mining and feature identification. Additionally, to understand the ambiguity of spoken words in natural language, we developed a word sense disambiguation technique, adapting the Lesk algorithm. Unlike many other systems, it does not need any wearable or in-situ sensors (other than Kinect sensor), thus it should not cause discomfort for patients. The scope of this project can be extended to smart homes and security where DAVE can be used to detect several previously undetectable verbal anomalous events.

View details »




AsthmaGuide:

A smartphone and cloud based asthma system


There has been an increased use of wireless sensor networks in the medical sector. AsthmaGuide is a system in which a smart phone is used as a hub for collecting physiological, environmental, human input, picture, and video information from several wireless sensors like sensordrone, electronic stethoscope, pulse oximeter, etc. The data, including data over time, is then displayed and analysed in a cloud web application for both patients and health-care providers to view. AsthmaGuide also provides an advice and alarm infrastructure based on the collected data and parameters set by health-care providers.

View details »

asthmaguide



Detection of Chronic Kidney Disease and Selecting Important Predictive Attributes


CKD-COST

Chronic kidney disease (CKD) is a major public health concern with rising prevalence. In this study we consider 24 predictive parameters and create a machine learning classifier to detect CKD. We evaluate our approach on a dataset of 400 individuals, where 250 of them have CKD. Using our approach we achieve a detection accuracy of 0.993 according to the F1-measure with 0.1084 root mean square error. This is a 56% reduction of mean square error compared to the state of the art (i.e., the CKD-EPI equation: a glomerular filtration rate estimator). We also perform feature selection to determine the most relevant attributes for detecting CKD and rank them according to their predictability. We identify new predictive attributes which have not been used by any previous GFR estimator equations. Finally, we perform a cost-accuracy tradeoff analysis to identify a new CKD detection approach with high accuracy and low cost.

View details »