Deench

Everything for Everyone

Category Archives: computerscience

University of Florida guts computer science department in budget-cutting move

Contrary to some reports, Florida is not eliminating its computer science department. Instead, it is proposing a radical restructuring that will kill the research arm of one department, scattering its faculty and leaving any left behind with an expanded teaching burden and fewer teaching assistants. Whether it would be better to simply kill it isn't clear.

The University of Florida, as with many state institutions, has seen its financial support plummet over the past decade (Forbes indicates it's down 30 percent over the last six years). This has triggered a number of cuts to academic programs; this year, the cuts fell on the Computer & Information Science and Engineering Department, which is home to an active research program. Under the plan proposed by the Dean, a bit less than half of the faculty will be the Electrical and Computer Engineering Department, and a few others will be spread among three other departments.

The faculty that's left in Computer & Information Science and Engineering will be converted into teaching only. All teaching assistants in that department will be eliminated, and their duties will be taken up by the remaining faculty. Realistically, most faculty who are actively engaged in a research program are unlikely to accept that sort of change, and will probably look for work elsewhere. Although this plan would seem to gut the department and leave its remains in chaos, the Dean concludes by stating, "this new mission would allow us to devote more faculty time to grow both the size and excellence of the Computer Science degree program."

The students remaining in the department have organized various protests, from a letter-writing campaign to demonstrations.

Read the comments on this post


Erasing a bit shown to boost entropy

In 1961 Rolf Landauer linked information and thermodynamic entropy by showing that erasing or combining bits of memory must be accompanied by an increase in entropy. For the first time since then, a team of physicists have experimentally verified this principle.

According to Landauer’s principle, any logically irreversible transformation of information results in, at best, some small dissipation of heat. The specific amount depends on the operating temperature—per bit, it amounts to around 3×10-21 joules at room temperature. This energy is the Landauer limit, and controls the maximum energy efficiency of computers (it's similar to the Carnot efficiency in heat engines, both of which are related to entropy).

Measuring such a tiny amount of energy in a memory storage devices is, to say the least, challenging. But now, a team from École Normale Supérieure, the University of Kaiserslautern, and the University of Augsburg has managed to do so.

Read the rest of this article...

Read the comments on this post


Aerial robots swarm the stage at TED [video]

Long Beach, California—Vijay Kumar's videos have already been a hit on YouTube, as people have been fascinated to watch swarms of robotic quadrotors perform various feats, like flying through narrow windows and coasting across a room in formation. But Kumar still had a few tricks up his sleeve when he took the stage at TED, and he seized the opportunity to show some serious ways in which aerial robots will change our world. 

Some could say that aerial robots are already making a huge impact, primarily in military applications where (very) remote humans often pilot drones in hostile territories. 

Kumar, however, envisions aerial robots that can fly themselves and carry out their tasks, on their own, or with minimal human input beyond initial design and programming. His drones offload even more of the job of stabilizing their flight to computers that aren't even on-board the copter (a weight and complexity advantage). Once airborne, the entire flight is computer-controlled.

Read the rest of this article...

Read the comments on this post


Nature Editorial: If you want reproducible science, the software needs to be open source

Modern scientific and engineering research relies heavily on computer programs, which analyze experimental data and run simulations. In fact, you would be hard-pressed to find a scientific paper (outside of pure theory) that didn’t involve code in some way. Unfortunately, most code written for research remains closed, even if the code itself is the subject of a published scientific paper. According to an editorial in Nature, this hinders reproducibility, a fundamental principle of the scientific method.

Reproducibility refers to the ability to repeat some work and obtain similar results. It is especially important when the results are unexpected or appear to defy accepted theories (for example, the recent faster-than-light neutrinos). Scientific papers include detailed descriptions of experimental methods—sometimes down to the specific equipment used—so that others can independently verify results and build upon the work.

Read the rest of this article...

Read the comments on this post


Mendeley API Battle: open genetics-sharing tool declared victorious

In March, the platform-agnostic research management tool, Mendeley, announced an API Battle, held in conjunction with the Public Library of Science. The goal was to spur the community into developing neat apps that use the database that powers Mendeley. Apps were judged by a panel that included Tim O'Reilly (of O'Reilly Media) and Amazon CTO Werner Vogels, based on the following criteria: API key usage, whether or not it goes viral, how much the app contributes to collaboration and transparency, and general coolness.

Late last week (December 1st) the winners were announced, and 1st place went to openSNP, a community-driven project for publicly sharing personal genetic data (such as an individual's 23andMe results). You can read an interview with the winners over at Mendeley's blog. PaperCritic and rOpenSci were the runners-up.

Read the comments on this post