Called LOGAN, the deep neural network, i.e., a machine of sorts, can learn to transform the shapes of two different objects, for example, a chair and a table, in a natural way, without seeing any paired transforms between the shapes.
Read morePredicting and controlling disease outbreaks would be easier and more reliable with the wider application of mathematical modelling, according to a new study.
Read moreResearchers have developed a model that recovers valuable data lost from images and video that have been 'collapsed' into lower dimensions.
Read moreResearchers have modeled the fluid dynamics of multi-rotor wind turbines via high-resolution numerical simulations. The simulations demonstrate a clear advantage for a turbine model with four rotors. The researchers found, that the wind turbine wake recovers much faster with multi-rotor turbines, that multi-rotor turbines produce slightly more energy than single-rotor turbines, and that a turbine with four rotors as far apart as possible is the optimal construction.
Read moreMany of today's scientific processes are simulated using computer-driven mathematical models. But for a model to accurately predict how air flow behaves at high speeds, for example, scientists need supplemental real life data. Providing validation data, using up-to-date methods, was a key motivating factor for a recent experimental study.
Read moreHydrologic models that simulate and predict water flow are used to estimate how natural systems respond to different scenarios such as changes in climate, land use, and soil management. The output from these models can inform policy and regulatory decisions regarding water and land management practices. Numerical models have become increasingly easy to employ with advances in computer technology and software with graphical user interface (GUI). While these technologies make the models more accessible, problems can arise if they are used by inexperienced modelers.
Read moreEngineers have solved a 50-year-old puzzle in signal processing. They've formulated the 'inverse chirp z-transform,' an algorithm related to one that's running on your cell phone right now. It took some computing power and some math expertise to do it.
Read moreAn artificial intelligence/machine learning model to predict which scientific advances are likely to eventually translate to the clinic has been developed.
Read moreAs tumor cells multiply, they often spawn tens of thousands of genetic mutations. Figuring out which ones are the most promising to target with immunotherapy is like finding a few needles in a haystack. Now a new model hand-picks those needles so they can be leveraged in more effective, customized cancer vaccines.
Read moreUsing virtual, cloud-based, interconnected computing techniques applied to 51,000 variables, researchers reduced the time needed to assess a cancer patient's tumor profile and suitability for clinical trials from 14 to 4 days. This method also increased two-fold, over a four-year period, the number of cases that could be assessed compared to conventional methods.
Read more