talk: A Practitioner’s Introduction to Deep Learning, 1pm Fri 11/17

ACM Tech Talk Series

​A Practitioner’s Introduction to Deep Learning

​Ashwin Kumar Ganesan, PhD student

1:00-2:00pm Friday, 17 November 2017​, ITE325, UMBC

In recent years, Deep Neural Networks have been highly successful at performing a number of tasks in computer vision, natural language processing and artificial intelligence in general. The remarkable performance gains have led to universities and industries investing heavily in this space. This investment creates a thriving open source ecosystem of tools & libraries that aid the design of new architectures, algorithm research as well as data collection.

This talk (and hands-on session) introduce people to some of the basics of machine learning, neural networks and discusses some of the popular neural network architectures. We take a dive into one of the popular libraries, Tensorflow, and an associated abstraction library Keras.

To participate in the hands-on aspects of the workshop, bring a laptop computer with Python installed and install the following libraries using pip.  For windows or (any other OS) consider doing an installation of anaconda that has all the necessary libraries.

  • numpy, scipy & scikit-learn
  • tensorflow / tensoflow-gpu (The first one is the GPU version)
  • matplotlib for visualizations (if necessary)
  • jupyter & ipython (We will use python2.7 in our experiments)

Following are helpful links:

Contact Nisha Pillai (NPillai1 at umbc.edu) with any questions regarding this event.

Volunteer to help in UMBC’s Hour of Code events, Dec 4 and 5

Volunteer to help in UMBC’s Hour of Code events, Dec 4 and 5

Hour of Code is a worldwide effort to celebrate computer science using a one-hour coding activity. The one hour coding activity is designed to show that anyone can learn the basics of coding, and to create a more inclusive computer science community. The CS Education student organization, with support from the CS Matters in Maryland CS education project, will host UMBC’s Hour of Code event on Monday (Dec. 4) and Tuesday (Dec. 5).

On December 4th from 10am-1pm, they will host the Hour of Code event for anyone who wants to try the coding activities on Main Street.

On December 5th from 11am-2pm, they will have special guests. Some students from Lakeland Elementary School will learn to code along with President Freeman Hrabowski.

The CS Education Club needs volunteers to make this event a success. Coding experience is useful, but not necessary. They will host training sessions prior to the event for everyone who volunteers. The training will give you a chance to try our Hour of Code activity along with a Makey Makey activity.  Read about last year’s Hour of Code event at UMBC here.

If you are interested in helping, please visit this link for more information and to sign up.

talk: Ferraro on Understanding What We Read and Share, 1pm Fri 11/10, ITE325, UMBC

 

ACM Faculty Talk Series

Understanding What We Read and Share:
Event Processing from Text and Images

Dr. Frank Ferraro, Assistant Professor, CSEE
1:00-2:00pm Friday, 10 November 2017, ITE 325, UMBC

A goal of natural language processing (NLP) is to design machines with human-like communication and language understanding skills. NLP systems able to represent knowledge and synthesize domain-appropriate responses have the potential to improve many tasks and human-facing applications, like virtual assistants such as Google Now or question answering systems like IBM’s Watson.

In this talk, I will present some of my work—past, on-going, and future—in developing knowledge-aware NLP models. I will discuss how to better (1) encode linguistic- and cognitive science-backed meanings within learned word representations, (2) learn high-level representations for document and discourse understanding, and (3) how to generate compelling, human-like stories from sequences of images.

Frank Ferraro is an assistant professor in the CSEE department at UMBC. His research focuses on natural language processing, computational event semantics, and unlabeled, structured probabilistic modeling over very large corpora. He has published basic and applied research on a number of cross-disciplinary projects, and has papers in areas such as multimodal processing and information extraction, latent-variable syntactic methods and applications, and the induction and evaluation of frames and scripts.

talk: Winning NCCDC, and its practicality in the real world, 12pm 11/3, ITE231

The UMBC Cyber Defense Lab presents

Winning NCCDC, and its practicality in the real world

Bryan Vanek, CSEE, UMBC

12:00noon–1pm Friday, 3 November 2017, ITE 231

The National Collegiate Cyber Defense Competition (NCCDC) takes place every year and gives students an environment where they can develop understanding and operational competency in managing and protecting corporate network infrastructure and business information systems. Competitors participate as the blue team, and try to protect their machines from being infiltrated by the red team, while simultaneously keeping critical services up and running in order for a mock business to stay up and running. After an immense amount of preparation and strife, the UMBC Cyber Defense Team took home its first national title for the competition this year. But what exactly did the team do to prepare for this competition? What exactly happened at the different stages of the competition? And just how practical are these situations in the real world? One of the winning team members will be covering these questions in this week’s CDL, so we hope to see you there!

Bryan Vanek is a UMBC undergraduate computer science major and mathematics minor. In addition to being one of the winning team members for NCCDC, he is currently serving as the president for the UMBC Cyber Defense Team, and is a CWIT T-SITE scholar. He currently works at Interclypse Inc. as a security engineer and software developer, and has had multiple internships and jobs dealing with aspects of computer development and security. Most recently he has completed his second internship at the Department of Defense  in the Summer Internship Program for Information Assurance. Upon graduation he will be returning to the DoD as a member of the Computer Network Operations development Program.

Host: Alan T. Sherman,

New UMBC course : CMSC 201 Computer Science I for Non-CS Disciplines

CMSC 201 Computer Science I for Non-CS Disciplines

 

This spring Dr. Susan Mitchell will teach a special section of CMSC 201 Computer Science I  designed for social and biological sciences *and other majors*. The course will cover the same content and have the same rigor as the regular sections of CMSC 201 and prepare students to continue on to CMSC 202 if they wish.  As with other sections, it fulfills any major’s requirement for CMSC 201. The main difference will be that the assignments and projects will emphasize topics applicable to many non-CS disciplines, such as statistical analysis, working with large data sets, and data visualization. The catalog description is:

An introduction to computer science through problem solving and computer programming. Programming techniques covered by this course include modularity, abstraction, top-down design, specifications documentation, debugging and testing. The core material for this course includes control structures, functions, lists, strings, abstract data types, file I/O, and recursion.

The course will include a lecture from 2:30pm to 3:45pm on Tuesdays and Thursdays in Engineering 027 and a one hour lab on Tuesdays at 11:30am.

Permission from the instructor is required to register for this section. No prior programming experience is required. The only prerequisite is that students must have completed MATH 150, 151 or 152 with a C or better; OR have MATH test placement into MATH 151; OR be concurrently enrolled in MATH 155 or completed it with a C or better.

For permission or questions, email Dr. Susan Mitchell at

talk: DOE Energy Exascale Earth System Model, 2:30 Tue 10/31, ITE325

 

CHMPR Distinguished Lecture Series

Energy Exascale Earth System Model

Dr. Mark Taylor, E3SM Chief Computational Scientist, Sandia National Laboratories

2:30pm Tuesday, 31 October 2017, ITE 325, UMBC

 

Dr. Taylor will present an overview of the DOE Energy Exascale Earth System Model (E3SM), including Sandia’s role in numerical algorithms, parallel scalability, and computational performance. E3SM is designed to run on upcoming next-generation DOE supercomputers. Adapting simulation codes to these new architectures is expected to be more disruptive than the previous transition from vector to massively parallel supercomputers. E3SM development is driven by several grand challenge science questions focused Earth’s cryosphere, biogeochemical and water cycle systems. E3SM has a new land and atmosphere component models branched from the CESM v1.2, coupled to new MPAS ocean, sea ice, and land ice models.

The current performance and throughput challenges of the E3SM high-resolution coupled configuration on several DOE computers will be discussed. Our current focus is on the NERSC Cori system with Intel Xeon Phi architecture, in the longer term we hope to make effective use of the upcoming NVIDIA GPU based system at ORNL. An analysis is presented of the E3SM spectral element atmosphere dycore following the NGGPS dycore computational evaluation protocol, but with an emphasis on the throughput rates needed for climate simulations. For even higher resolution simulations, we will rely on E3SM’s ability to use unstructured grids in all component models. This will allow us to achieve cloud-resolving resolution in select regions of interest seamlessly within the global modeling system.

Mark Taylor is a mathematician who specializes in numerical methods for parallel computing and geophysical flows. He currently serves as Chief Computational Scientist for the DOE’s Accelerated Climate Modeling for Energy (ACME) project. Mark developed the mimetic/conservative formulation of the spectral element method, one of the atmospheric dynamical cores used in the Community Earth System Model (CESM) and the ACME project. Mark received his Ph.D. from New York University’s Courant Institute of Mathematical Sciences in 1992 and has worked at Sandia National Laboratories since 2004. In 2014 he was awarded The Secretary of Energy Achievement Award for his work unifying the DOE climate modeling research community and enabling the development of high-resolution fully-coupled climate-system simulations.

IEEE Fall 2017 Arduino Workshop and Halloween Social, 5-7 Fri 10/27

IEEE Fall 2017 Arduino Workshop and Halloween Social, 5-7 Fri 10/27

The UMBC Student Branch of IEEE (Institute of Electrical and Electronics Engineers) will host a Halloween Social at 5-7:00pm Friday, 27 October 2017 in ITE 233. This is a fun event where attendees can build and take home electronic Halloween decorations. The Halloween Workshop will take place from 5pm to 7pm, and they will be giving out candies, Arduinos and various electronic equipment on a first come, first serve basis. If you would like to participate, please register here.

IEEE is the world’s largest professional association for the advancement of technology. The UMBC graduate and undergraduate student members share technical interests rooted in electrical and computer sciences, engineering and related disciplines. ​At their meetings and events, they present and promote current research trends at UMBC and elsewhere, host skills workshops, and provide our members the opportunity to expand their network of contacts.

Gymama Slaughter: The Art of Powering Implantable Electronics

The Art of Powering Implantable Electronics

UMBC professor Gymama Slaughter give a short talk at the Grit-X event on her recent research on powering implantable devices for medical applications.

The number of smart implantable devices is on the rise, especially as we approach the ramping up of the “internet of things.” A key challenge for implantable electronic devices has been keeping these devices properly and conveniently powered. Current battery technologies are sealed within these devices, thereby forcing the surgical replacement of the device once the battery is depleted. We need an inconspicuous means of powering implantable electronics with imperceptible methods that moves us toward new innovative solutions to the power challenge in implantable devices. A lightweight bio-solution that leverages the biochemical energy from human biological fluids is a step forward for powering these smart implantable technologies.

Open House: UMBC Graduate Cybersecurity and Data Science Programs, 6-7:30 Wed. 10/25

Open House: UMBC Graduate Professional Programs

The Fall Open House for UMBC Professional Programs, including the graduate programs on Cybersecurity and Data Science, takes place this coming Wednesday evening, 25 October 2017, at BWTECH South (map) from 6:00-7:30pm.

Students interested in pursuing such programs (MPS degrees and/or certificates) or just to learn more about the field are encouraged to register and attend. Current students interested in pursuing a BS/MPS option for selected programs (such as Cybersecurity or Data Science) are especially welcome.

Attendees who apply to start in Spring 18 will have their UMBC application fee waived.

The programs represented include:

Program directors for these programs will present in individual breakout sessions and relevant support staff from DPS, the UMBC Graduate School, Veterans Affairs, etc. will be on-hand to provide administrative overviews, answer questions, and mingle. Refreshments will be provided.

for more information, directions and to register, see here.

Nilanjan Banerjee: When What You Wear Understands You

When What You Wear Understands You

Professor Nilanjan Banerjee give a short talk at yher Grit-x event on recent research on systems that use intelligent, wearable sensors to provide better human-computer interfaces and for medical applications.

How can cutting-edge research on textile sensors and wearable radar sensors help us recognize gestures, monitor sleep fragmentation, and diagnose sleep disorders? The Banerjee lab has developed and applied sensors to users with upper extremity mobility impairments, adults suffering from insomnia and restless leg syndrome, and kids with attention deficit/hyperactivity disorder, with the intent to begin answering that question.

1 2 3 4 76