Bringing the world of machine learning to Howard West using Google Cloud Platform

Using Google Cloud Platform tools like Cloud Datalab and TensorFlow, Howard University professor Gloria Washington helps Howard West students grasp the practice of solving problems with the magic of machine learning and power of belief in themselves.

While Gloria Washington, now a computer science professor at Howard University, was getting her Ph.D. at George Washington University, she recalls people questioning whether she was smart enough to get her degree. “I actually had a professor tell me that (I wasn’t smart enough),” Washington said. “That can psychologically really mess someone up, especially if you have all the GRE scores; you have all the grades.” But, the naysayers weren’t going to stop Gloria from realizing her dreams.

"I want to increase the number of black women with Ph.D.'s in computer science"

Gloria Washington, Howard University, Computer Science Professor

A new interest: Human-Computer Interaction

As a National Physical Sciences Consortium fellow at George Washington, Gloria was introduced to Human-Computer Interaction (HCI), a field that focuses on the end user of applications. “I always knew I wanted to work in end user support, mixed in with software engineering, because I wanted to help people,” Washington said. “This is a field I’m really interested in because it’s more than just sitting at your desk and programming. It’s interacting with users, getting their stories, understanding what makes them tick, and developing applications around them.” Through her love for using computer science to help others and her strong will, Gloria persevered—receiving her Ph.D. in HCI.

After graduating and working for organizations like the National Security Agency (NSA) and Booz Allen Hamilton, Gloria realized her true calling was academia. “I realized that my ideas were not my ideas; unfortunately, they were the company’s ideas,” Washington explained. “That’s great—but when you are moving up in the ranks, and you’re really trying to establish yourself as a subject matter expert with a Ph.D., it gets hard because you are not getting recognition for your work.” Gloria decided it was time to make her way to teaching. “I can own the ideas, and I can also influence someone who is like me to get their Ph.D.”

Bringing her passion to Academia at Howard

Since starting her time at Howard, her goal has always been to inspire students and help them to achieve their goals. Gloria focuses on mentoring minorities who are overlooked in higher education. “I want to increase the number of black women with Ph.D.s in computer science,” Washington said. And, only two years after starting as a professor at Howard, she saw her first black female Ph.D. student graduate.

As Gloria continues to instill courage in her students, she still faces challenges in her Howard classroom. “Unfortunately, at Howard, we don't have huge amounts of money, so some of our computers are old,” Washington said. Older hardware lacks the processing power necessary to conduct the types of tasks her computer science students are working on.

So how does an underfunded program make classes more efficient and incorporate advanced technology to aid student learning? Gloria believes the answer lies with the use of Google Cloud Platform (GCP). “If we can connect to GCP, we would be able to get over the hurdle of hardware and the difficulty that creates,” Washington said.

Using Google Cloud Platform to help students build apps with ease

In some of Gloria’s classes at Howard, she typically sets her students up with a local application, which can slow down mobile development. But now, with the use of different GCP products, Gloria can eliminate this issue. “I wanted the students to have access to a backend that actually works and was easy to set up,” Washington said. “I didn’t want the students to focus so much on building a huge database and connecting with PHP. I wanted an API-based system that the students could easily plug into their mobile apps and have it capture the data they were using.”

Gloria talked about the success one of her students had using Firebase, Google’s mobile platform that helps you quickly and easily develop high-quality apps. “She created an application where you take a picture of yourself—whether you're frowning, smiling, or just indifferent—and it would suggest a playlist based on that emotion. The images were held in Firebase so they could be accessed later. It (Firebase) really helped—all she had to do was use the API to connect it to the mobile app.”

"I chose Datalab so we could get up and running really quickly and would not have to download anything."

Gloria Washington, Howard University, Computer Science Professor

Bringing the power of Machine Learning to students

As the lab professor for the inaugural Howard West class on Google’s main campus in Mountain View, CA, Gloria co-taught a computer science class on machine learning with Googler Sally Goldman this summer. Her students learned how to approach machine learning problems through the use of GCP’s Cloud Datalab and TensorFlow.

Datalab is a powerful interactive tool created to explore, analyze, transform, and visualize data as well as build machine learning models that run on GCP's Google Compute Engine. TensorFlow is an open source software library for numerical computation that uses data flow graphs to make it easier for developers to build and train machine learning algorithms. Both GCP tools have helped Gloria and her students in the classroom.

“Datalab was really instrumental because students didn’t have to download TensorFlow with the library and make sure that they had the required supporting libraries on their laptops,” Washington said. “Currently, it (Datalab) runs in Jupiter so if (the students) wanted to ingest data into TensorFlow, they could easily do that with one little call of code and then it would run in the actual interface. It is pretty fast because it connected to the cloud service where it stored a copy of the databases that we had. I am really hoping that I will be able to take Datalab back to Howard University.”

Gloria had a clear reason for choosing to use Datalab this summer. “I chose Datalab so we could get up and running really quickly and would not have to download anything,” Washington said. In less than a day, she was able to get Datalab up and running. Gloria expressed her appreciation for how Datalab makes it very easy for her and her students to use TensorFlow machine learning algorithms and APIs.

“When we first got started, it was really fast. The only thing I did was give the students a chance to get re-familiarized with Python so that they could know how to make certain API calls to TensorFlow,” Washington explained. “If you think in terms of semester weeks, that's really good because if you have to download a library, you have to show them how to call it on a computer and with Datalab, and we didn't have to do that.”

Enabling learning in just 5 weeks

Overall, using Datalab and Tensorflow allowed Gloria to design a summer course that was application- and project-centered and focused on helping students grasp the practice for solving problems using machine learning versus an emphasis on the technology behind the scenes.

“For us, that was really crucial because we only had five weeks for this course. If you spend a week trying to get them to download the libraries and then make sure that they have the right commands to be able to run it from the shell, that just creates a whole level of complexity that we didn't want to deal with. The complexity was really cut down.”

As Gloria heads back to Howard, she cannot imagine teaching without Google’s tools. “I envision being able to use (Datalab and Google Developer Codelabs) for future research projects at Howard,” Washington said. “I know right now, if I had to teach machine learning at Howard without Datalab, I would be mad.”

In addition to enabling her students with technology, Gloria offers advice on how they can overcome obstacles and grow academically.

“I’ve thought a lot about what I would say to my 20-year-old-self,” Washington said. “One person does not determine your success . . . if you take the time and do the work . . . no one can determine your success except for you. Do not let one person’s view of you change what you want to do with your whole life. Join a professional organization where people encourage each other and it's not about competition and grades, it’s about uplifting each other up. Make sure that you do stuff outside of the classroom for yourself, for example, if you are interested in human-centered computing, make sure you go read papers on your own so that no one can say, ‘gosh she doesn't know what she's talking about.’ ”

Organization Profile

Howard West was an immersive 12-week summer experience that brought Howard Computer Science (CS) students to Google for a dedicated residency program, at the Googleplex in Mountain View, California. Howard students participated in an experiential-based curriculum (taught by Howard faculty and Google Engineers) at the Google campus during the final two years of their computer science program. Students formed into project teams directly learning from, and engaging in professional/academic activities at Google. This model provided Howard students the opportunity to be engaged in projects framed in “real world” application, and develop skills that will enable students to be “work ready” upon graduation. The overall goals of the program were to strengthen relationships between Google, Silicon Valley and the African American Community, as well as to increase representation of African Americans in the tech industry.

Products used

Thanks for signing up!

Let us know more about your interests.