Failed experiments represent a source of training data for machine learning

Machine learning relies on training data to derive patterns useful for making predictions in test set data. Iterative in nature, training of machine learning algorithms relies on large amount of data where the larger the training set, the better would be the pattern recognition capability of the algorithm. Most training data are positive data, that … More Failed experiments represent a source of training data for machine learning

Limit for forward weather forecast may be 14 days

Advances in computational modelling, particularly in understanding how cloud dynamics influence atmospheric circulation and weather, has brought forth an increase in weather prediction accuracy and the number of days of forward weather forecast. The current best-in-class atmospheric circulation and weather models afford a forward weather forecast of 10 days with sufficient accuracy in prediction.   … More Limit for forward weather forecast may be 14 days

Aberration correction lenses help improve resolution in transmission electron microscopy

Using electron scattering and diffraction, transmission electron microscopy is able to achieve resolution higher than that of optical microscopy. However, the effect of aberration due to refraction of electron beams also introduce uncertainty into the image and reduces resolution. Nevertheless, the advent of aberration correction lenses has helped improve the resolution of transmission electron microscopy … More Aberration correction lenses help improve resolution in transmission electron microscopy

Flow batteries could provide for grid level storage

Intermittency in solar and wind power meant that energy storage solutions would need to be used to provide power during times when solar and wind power are not available. However, difficulties exist in using current lithium-ion battery technologies for storing large amount of wind and solar power as grid level storage given problems with scaling. … More Flow batteries could provide for grid level storage

High technology paints for enabling autonomous driving

Radio waves and light sensors are commonly fitted to autonomous vehicles for detecting the speeds and distances of objects around the vehicle. However, dark paints on vehicles tend to absorb the radio waves and infrared signals used by such sensors. Hence, much research has been invested in the creation of multi-layer structured high technology paints … More High technology paints for enabling autonomous driving

Superionic water

A recent article in the New York Times (Link) describes new research on the structure of water. Although much researched, there remains much to be learnt of the molecular mysteries of water. In new research, water was found to behave both as a liquid and solid at high pressure and temperature. Specifically, the material structure … More Superionic water

Obtaining movies of molecular movement of proteins through serial femtosecond crystallography

Seeing is believing; thus, it would be much better if we could observe real-time molecular level movement of proteins during catalysis, where, for example, binding of a substrate to an active site leads to molecular cleavage and enzymatic action. But, this feat is not realized until recent advent of serial femtosecond crystallography (SFX).   In … More Obtaining movies of molecular movement of proteins through serial femtosecond crystallography

Challenges of recycling lithium-ion batteries

With increasing use of lithium ion batteries for personal mobile devices and computers, there is a concomitant need to plan the end of life treatment of these batteries that contain precious metals as well as toxic and flammable substances. Specifically, while lithium ion is thought to be the most important element for recover in the … More Challenges of recycling lithium-ion batteries

Examining surface premelting in colloidal particles

Article in Nature, Vol. 531, 24 March 2017, pp. 485, “Modes of surface premelting in colloidal crystals composed of attractive particles”   Summary of article: Using Landau force field with poly(methyl methacrylate) characteristics, the effect of cohesion forces at low temperature was investigated and revealed the formation of crystalline domains and grain boundaries.   Link … More Examining surface premelting in colloidal particles