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

Nonlinear Programming: A Unified Approach

A surgical instrument is provided, including: at least one articulatable arm having a distal end, a proximal end, and at least one joint region disposed between the distal and proximal ends; an optical fiber bend sensor provided in the at least one joint region of the at least one articulatable arm; a detection system coupled to the optical fiber bend sensor, said detection system comprising a light source and a light detector for detecting light reflected by or transmitted through the optical fiber bend sensor to determine a position of at least one joint region of the at least one articulatable arm based on the detected light reflected by or transmitted through the optical fiber bend sensor; and a control system comprising a servo controller for effectuating movement of the arm.

Robotic surgery system including position sensors using fiber bragg gratings

A surgical instrument is inserted through a guide tube. A telemanipulation system moves the distal end of the surgical instrument in all six Cartesian degrees of freedom independently of any guide tube movements.

Minimally invasive surgical system

Optical fiber gratings have developed into a mature technology with a wide range of applications in various areas, including physical sensing for temperature, strain, acoustic waves and pressure. All of these applications rely on the perturbation of the period or refractive index of a grating inscribed in the fiber core as a transducing mechanism between a quantity to be measured and the optical spectral response of the fiber grating. This paper presents a relatively recent variant of the fiber grating concept, whereby a small tilt of the grating fringes causes coupling of the optical power from the core mode into a multitude of cladding modes, each with its own wavevector and mode field shape. The main consequence of doing so is that the differential response of the modes can then be used to multiply the sensing modalities available for a single fiber grating and also to increase the sensor resolution by taking advantage of the large amount of data available. In particular, the temperature cross-sensitivity and power source fluctuation noise inherent in all fiber grating designs can be completely eliminated by referencing all the spectral measurements to the wavelength and power level of the core mode back-reflection. The mode resonances have a quality factor of 105, and they can be observed in reflection or transmission. A thorough review of experimental and theoretical results will show that tilted fiber Bragg gratings can be used for high resolution refractometry, surface plasmon resonance applications, and multiparameter physical sensing (strain, vibration, curvature, and temperature).

Tilted fiber Bragg grating sensors

Deep neural networks (DNN) have achieved breakthroughs in applications with large sample size. However, when facing high dimension, low sample size (HDLSS) data, such as the phenotype prediction problem using genetic data in bioinformatics, DNN suffers from overfitting and high-variance gradients. In this paper, we propose a DNN model tailored for the HDLSS data, named Deep Neural Pursuit (DNP). DNP selects a subset of high dimensional features for the alleviation of overfitting and takes the average over multiple dropouts to calculate gradients with low variance. As the first DNN method applied on the HDLSS data, DNP enjoys the advantages of the high nonlinearity, the robustness to high dimensionality, the capability of learning from a small number of samples, the stability in feature selection, and the end-to-end training. We demonstrate these advantages of DNP via empirical results on both synthetic and real-world biological datasets.

/pdf/deep-neural-networks-for-high-dimension-low-sample-size-data-xkne98xtjd.pdf

Deep neural networks for high dimension, low sample size data

The current status of the fiber Bragg grating (FBG) sensor technology was reviewed. Owing to their salient advantages, including immunity to electromagnetic interference, lightweight, compact size, high sensitivity, large operation bandwidth, and ideal multiplexing capability, FBG sensors have attracted considerable interest in the past three decades. Among these sensing physical quantities, temperature and strain are the most widely investigated ones. In this paper, the sensing principle of FBG sensors was briefly introduced first. Then, we reviewed the status of research and applications of FBG sensors. As very important for industrial applications, multiplexing and networking of FBG sensors had been introduced briefly. Moreover, as a key technology, the wavelength interrogation methods were also reviewed carefully. Finally, we analyzed the problems encountered in engineering applications and gave a general review on the development of interrogation methods of FBG sensor.

Review of fiber Bragg grating sensor technology

A compact magnetic field sensor has been proposed based on multimode interference effects. It consists of typical multimode interferometer (MMI) immersed into the magnetic fluid (MF) which is formed by a section of square no-core fiber (NCF) spliced between two single-mode fibers. The transmission spectral characteristics of this MMI have been analyzed, and the spectral magnetic response of the proposed sensor has been investigated by immersing the NCF into the MF environment. The transmission response of the interference maxima exhibits a sensitivity of −0.01939 dB/Oe in the relatively linear range. Due to its low cost and compactness, this sensor would find potential applications in the measurement of magnetic field.

Fiber-optic in-line magnetic field sensor based on the magnetic fluid and multimode interference effects

Refractive index sensor based on measuring the transmission power of tilted fiber Bragg grating

In this paper, the mode coupling mechanism of tilted fiber Bragg gratings (TFBGs) is briefly introduced at first. And a general review on the fabrication, theoretical and experimental research development of TFBGs is presented from a worldwide perspective, followed by an introduction of our current research work on TFBGs at the Institute of Modern Optics, Nankai University (IMONK), including TFBG sensors for single-parameter measurements, temperature cross sensitivity of TFBG sensors, and TFBG-based interrogation technique. Finally, we would make a summary of the related key techniques and a remark on prospects of the research and applications of TFBGs.

/pdf/tilted-fiber-bragg-gratings-principle-and-sensing-4c07n58op9.pdf

Bo Liu

Papers

Deep neural networks for high dimension, low sample size data

Review of fiber Bragg grating sensor technology

Fiber-optic in-line magnetic field sensor based on the magnetic fluid and multimode interference effects

Refractive index sensor based on measuring the transmission power of tilted fiber Bragg grating

Tilted fiber Bragg gratings: Principle and sensing applications