The structure-function relationship, where the 3D shape of a protein defines its function, is the theoretical cornerstone behind the success of computer-aided tools for protein engineering, drug design, and drug discovery. The structural dynamics of a protein are critically connected to its function; however, software tools often ignore conformational changes and localized fluctuations in order to reduce the complexity of the modeling process.
This year's SIGMOD Programming Contest (http://sigmod18contest.db.in.tum.de/index.shtml) was about evaluating as fast as possible batches of SPJA (Selection-Projection-Join-Aggregation) queries on a set of immutable relations, under a single-node multicore and in-memory setting. We built from scratch a system called Robin that ranked 1st place on the leaderboard, which was faster by a large margin (40%) than the runner-up solution.
Traditional storage systems interfaces, including POSIX file I/O and block-based interfaces have been a major success. They are easy to understand and use and their decades-long rule have prevented vendor lock-in and encouraged innovation across common interfaces. However, HPC and cloud systems are pushing their scalability, and behind their veneer of simplicity, idiosyncratic "magic" numbers are lurking, destroying performance if applications fail to discover and properly tend to them.
To celebrate the end of the academic year, the Computer Sciences Department invites you to our annual lunch on Monday, April 30th from 11:45 AM - 1:30 PM.
The lunch will be conveniently located over in WID (Wisconsin Inst. for Discovery), in the H.F. DeLuca Forum.
The event is a great opportunity to visit informally with CS faculty and other students, both undergraduates and graduates. At the lunch, we'll announce the winners of the CS Department Awards and Fellowships.
We study an online resource allocation problem where a seller has many items with multiplicities to offer and the items are arranged in a total order. Buyers are interested in buying intervals of items, and have different values for different intervals, drawn from a known distribution. The seller’s goal is to design an online allocation mechanism that maximizes social welfare. Importantly, buyers’ preferences have complementarities, so that recently-developed constant-factor approximations via item prices do not apply, and indeed strong negative results are known.
In the fight against cancer, Standard Imaging delivers precision QA products that help lead to better patient outcomes. We are a market leader specializing in radiotherapy QA technology.
Do you have strong back end engineering skills with a passion for creating exceptional user experiences? We are seeking a creative, talented full-stack engineer to help us build medical device applications for the radiotherapy, quality assurance industry.
Quantitative Risk Management (QRM) is the world’s premier enterprise risk management consulting firm. We
develop industry-leading risk management principles, practices, and models, and provide clients with the advice,
knowledge, and tools necessary to adopt those innovations, thus increasing risk-adjusted returns. For over 20
years, QRM has partnered with clients to enhance their ability to measure risk, identify profitable opportunities, and
We’re looking for a self-motivated iOS Developer to join our startup team. As a leading Developer, you will be working directly with Jia, a team of designers and hundreds of app testers to build an app based on the Rejection Therapy card game.
Even as the cost of DNA sequencing continues to drop, there is still a strong interest among researchers to use techniques such as targeted enrichment to enable more focused sequencing of specific regions in a genome. The lower cost of sequencing allows researchers to increase the number of samples, depth of coverage, or panel size. For this talk, I will focus on three published studies which have used panels of nucleic acid probes for targeted enrichment in unique ways to accomplish their scientific goals.