Dr. Roush previously worked as an editor-at-large at Xconomy, an online news outlet covering technological innovation from the USA. He served as the managing editor of Xperience, a branch of the website that was dedicated to exploring how technology affected users. Dr. Roush received his PhD in the history and study of science from MIT in 1994.
The Hangout began with Dr. Roush giving brief introductions of the panelists – Dr. Tracy Hammond, Dr. Allen Downey, and Dr. Ari Trachtenberg.
“Welcome to our panelists,” Dr. Roush greeted. “We’d like to start off by having each of you telling us a little about your current research and expand on my very sketchy introductions.”
Dr. Tracy Hammond, the Director of the Sketch Recognition Lab at Texas A&M University since 2006, was first. While at MIT, she worked in sketch recognition with Dr. Randy Davis in the Design Rationale Group.
“I would say that MIT prepared me really well for my faculty life here [at Texas A&M],” she stated.
Dr. Hammond’s introduction touched on various areas of her studies.
“Sketch recognition is the automated recognition of hand drawn diagrams,” Dr. Hammond defined as she explained her studies. She followed this description with a quick video demonstration of what sketch recognition and how understood diagrams can bring a diagram to life, whether they are cars rolling down a hill, physics diagrams, or artistic drawings. The other panelists broke into smiles and appreciative nods at the demonstration of the animation of the car moving down the slope – including an animation of the wheels rolling.
“We've done this in many different domains,” Dr. Hammond continued as she explained her involvement with sketch recognition studies.
One such domain is using programs in an academic setting to automatically correct homework that is hand-drawn. This innovative program can be used over a wide variety of subjects and has been tests in statistics, biology, foreign language (specifically Japanese), and coding. Another area in which sketch recognition is being used is for people learning how to draw.
“We've taken 2D into 3D when we do activity recognition,” Dr. Hammond transitioned into her second area of study.
From sit ups to lunges to jumping jacks, activity recognition has the ultimate goal of mapping and understanding exactly what kind of action a user is doing.
“We've taken that into other domains,” she stated and then discussed an example. “Trying to help people quick smoking by being able to predict when they’re about to take a cigarette and when they are smoking, so we can do something to change those actions.”
In regard to haptics and wearable computing, Dr. Hammond briefly explained GeoTrooper/HaptiMoto, a system that navigates paratroopers/motorcyclists by touch alone with no training.
“I have a unique experience at Texas A&M,” she concluded with a smile. “It’s a great experience, you get to influence a lot of people’s lives. I really love working with the students.”
Next to introduce himself was Dr. Allen Downy, a professor of Computer Science at Olin College of Engineering.
“The mission of the college is to fix engineering education,” he added with a smile. “So it’s attracted a lot of people who are interested in curriculum design and engineering and engineering education.”
When Dr. Roush later asked what the mission meant and what drew Dr. Downy to Olin college, Dr. Downy responded, “I should say that’s not the actual wording of our mission, that’s my paraphrase.”
Dr. Downy went on to explain that one thing that drew him to Olin College was the reflection on his own education experiences and noticing that, sometimes, things did not work well or needed changing. A major part of Dr. Downy’s work consists of building curriculum and revising it. He commented on this work by describing Olin College as a place that is innovative and eager to experiment with new ideas.
“It’s a lot of fun to get out of computer science, where I did my PhD,” he stated. “And have a chance to teach engineering design and modeling and simulation and, currently, statistics and Bayesian statistics and a whole lot of things that I don’t know much about,” he joked.
Dr. Downy then transitioned into discussing a couple of projects he is working on. The first of which is a book series which he refers to as “Think X for all X.” The first of this series of books is Think Python, in which Dr. Downy discusses how to learn the Python programming language. The remainder of the series moves forward with the concept that if a person has basic programming skills, they can use that as a foundation to learn everything else.
“If you can write short programs and implement the algorithms and run experiments,” he explained, “you can learn things very quickly and learn them from top – down.”
Another project described by Dr. Downy is his blog called ‘Probably Overthinking It’, found here. On his blog, Dr. Downy takes ideas and concepts from statistics and applies them to a real world setting. This allows him to make concepts “more accessible, more applied, more interesting than statistics sometimes is.”
Currently, Dr. Downy is working with students on a study about statistically driven analysis of survival predictions for characters in Game of Thrones. This study turned into an article predicting which characters will or will not survive the upcoming seasons.
The final panelist to introduce himself was Ari Trachtenberg, a professor of Computer Engineering, Information and Data Sciences, Networking and Information Systems at Boston University.
“I got my Bachelors degree at MIT,” he began. “And my claim to fame as an undergrad was that I made it to the Tetris high score list. I was in 73rd place, which probably meant I played a little too much Tetris.”
Now at working at Boston University, Dr. Trachtenberg has conducted research in cell phone security – one such study he did with students was to see if a cellphone could identify its owner based on their walk, their gait. Other research he has conducted has been about how to protect cellphones from mode exploits and how to protect and attach the memory of a phone. While studying phone memory, Dr. Trachtenberg has also studied how various ways of storing memory affects the encryption of the stored file.
“My aim is to do the impossible,” he commented. “Or at least the implausible and, until then, I work on the improbable. Which is really the publishable, and really that’s just the fundable.”
Currently, Dr. Trachtenberg is studying a specific type of encryption called deniable encryption.
“Deniable encryption is that idea of being able to encrypt a text with a particular encryption so that, if you want to, you can encrypt it to the original decryption. And if you don’t want to, you can decrypt it with another decryption,” he defined.
As an example, Dr. Trachtenberg described a hypothetical situation in which Person A is cornered by Person B. Person B demands to know what an encrypted text says. Person A has the choice to show Person B something that is completely made up and believable through deniable encryption.
The fake text, of course, must be similar to the original message, but small bits of text transposed to different parts of a message can change the meaning completely. Think about the comma comics – “Let’s eat Grandpa” as compared to “Let’s eat, Grandpa.”
The remainder of the session mainly consisted of panelists answering viewer questions as they came in from a forum. The Hangout and question and answer period in its entirety can be found here.
No comments:
Post a Comment