Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Rapid growth in nanotechnology is increasing the likelihood of engineered nanomaterials coming into contact with humans and the environment. Nanoparticles interacting with proteins, membranes, cells, DNA and organelles establish a series of nanoparticle/biological interfaces that depend on colloidal forces as well as dynamic biophysicochemical interactions. These interactions lead to the formation of protein coronas, particle wrapping, intracellular uptake and biocatalytic processes that could have biocompatible or bioadverse outcomes. For their part, the biomolecules may induce phase transformations, free energy releases, restructuring and dissolution at the nanomaterial surface. Probing these various interfaces allows the development of predictive relationships between structure and activity that are determined by nanomaterial properties such as size, shape, surface chemistry, roughness and surface coatings. This knowledge is important from the perspective of safe use of nanomaterials.

Understanding biophysicochemical interactions at the nano–bio interface

Worldwide commercial interest in carbon nanotubes (CNTs) is reflected in a production capacity that presently exceeds several thousand tons per year. Currently, bulk CNT powders are incorporated in diverse commercial products ranging from rechargeable batteries, automotive parts, and sporting goods to boat hulls and water filters. Advances in CNT synthesis, purification, and chemical modification are enabling integration of CNTs in thin-film electronics and large-area coatings. Although not yet providing compelling mechanical strength or electrical or thermal conductivities for many applications, CNT yarns and sheets already have promising performance for applications including supercapacitors, actuators, and lightweight electromagnetic shields.

/pdf/carbon-nanotubes-present-and-future-commercial-applications-1nhq3xy4hh.pdf

Carbon Nanotubes: Present and Future Commercial Applications

Skyrmions represent topologically stable field configurations with particle-like properties. We used neutron scattering to observe the spontaneous formation of a two-dimensional lattice of skyrmion lines, a type of magnetic vortex, in the chiral itinerant-electron magnet MnSi. The skyrmion lattice stabilizes at the border between paramagnetism and long-range helimagnetic order perpendicular to a small applied magnetic field regardless of the direction of the magnetic field relative to the atomic lattice. Our study experimentally establishes magnetic materials lacking inversion symmetry as an arena for new forms of crystalline order composed of topologically stable spin states.

Skyrmion Lattice in a Chiral Magnet

Magnetic skyrmions are particle-like nanometre-sized spin textures of topological origin found in several magnetic materials, and are characterized by a long lifetime. Skyrmions have been observed both by means of neutron scattering in momentum space and microscopy techniques in real space, and their properties include novel Hall effects, current-driven motion with ultralow current density and multiferroic behaviour. These properties can be understood from a unified viewpoint, namely the emergent electromagnetism associated with the non-coplanar spin structure of skyrmions. From this description, potential applications of skyrmions as information carriers in magnetic information storage and processing devices are envisaged.

Topological properties and dynamics of magnetic skyrmions

We study the coalescence of nanoscale metal clusters in an inert-gas atmosphere using constant-energy molecular dynamics. The coalescence proceeds via atomic diffusion with the release of surface energy raising the temperature. If the temperature exceeds the melting point of the coalesced cluster, a molten droplet forms. If the temperature falls between the melting point of the larger cluster and those of the smaller clusters, a metastable molten droplet forms and freezes.

Coalescence of nanoscale metal clusters: Molecular-dynamics study

Phishing emails have been used widely in fraud of financial organizations and customers. Phishing email detection has drawn a lot attention for many researchers and malicious detection devices are installed in email servers. However, phishing has become more and more complicated and sophisticated and attack can bypass the filter set by anti-phishing techniques. In this paper, we present a method to build a robust classifier to detect phishing emails using hybrid features and to select features using information gain. We experiment on 10 cross-validations to build an initial classifier which performs well. The experiment also analyses the quality of each feature using information gain and best feature set is selected after a recursive learning process. Experimental result shows the selected features perform as well as the original features. Finally, we test five machine learning algorithms and compare the performance of each. The result shows that decision tree builds the best classifier.

/pdf/detecting-phishing-emails-using-hybrid-features-5dekq9u4hj.pdf

Detecting Phishing Emails Using Hybrid Features

A crystallographic orientation transition and early stage growth characteristics of thin Bi films on HOPG

Current efforts to achieve neuromorphic computation are focused on highly organized architectures, such as integrated circuits and regular arrays of memristors, which lack the complex interconnectivity of the brain and so are unable to exhibit brain-like dynamics. New architectures are required, both to emulate the complexity of the brain and to achieve critical dynamics and consequent maximal computational performance. We show here that electrical signals from self-organized networks of nanoparticles exhibit brain-like spatiotemporal correlations and criticality when fabricated at a percolating phase transition. Specifically, the sizes and durations of avalanches of switching events are power law distributed, and the power law exponents satisfy rigorous criteria for criticality. These signals are therefore qualitatively and quantitatively similar to those measured in the cortex. Our self-organized networks provide a low-cost platform for computational approaches that rely on spatiotemporal correlations, such as reservoir computing, and are an important step toward creating neuromorphic device architectures.

https://advances.sciencemag.org/content/advances/5/11/eaaw8438.full.pdf

Simon Brown

Papers

Coalescence of nanoscale metal clusters: Molecular-dynamics study

Detecting Phishing Emails Using Hybrid Features

A crystallographic orientation transition and early stage growth characteristics of thin Bi films on HOPG

Avalanches and criticality in self-organized nanoscale networks

STM and XPS investigations of bismuth islands on HOPG