Learning the learning rate for prediction with expert advice

Most standard algorithms for prediction with expert advice depend on a parameter called the learning rate. This learning rate needs to be large enough to fit the data well, but small enough to prevent overfitting. For the exponential weights algorithm, a sequence of prior work has established theoretical guarantees for higher and higher data-dependent tunings of the learning rate, which allow for increasingly aggressive learning. But in practice such theoretical tunings often still perform worse (as measured by their regret) than ad hoc tuning with an even higher learning rate. To close the gap between theory and practice we introduce an approach to learn the learning rate. Up to a factor that is at most (poly)logarithmic in the number of experts and the inverse of the learning rate, our method performs as well as if we would know the empirically best learning rate from a large range that includes both conservative small values and values that are much higher than those for which formal guarantees were previously available. Our method employs a grid of learning rates, yet runs in linear time regardless of the size of the grid.

Second-order quantile methods for experts and combinatorial games

We aim to design strategies for sequential decision making that adjust to the difficulty of the learning problem. We study this question both in the setting of prediction with expert advice, and for more general combinatorial decision tasks. We are not satisfied with just guaranteeing minimax regret rates, but we want our algorithms to perform significantly better on easy data. Two popular ways to formalize such adaptivity are second-order regret bounds and quantile bounds. The underlying notions of 'easy data', which may be paraphrased as "the learning problem has small variance" and "multiple decisions are useful", are synergetic. But even though there are sophisticated algorithms that exploit one of the two, no existing algorithm is able to adapt to both. In this paper we outline a new method for obtaining such adaptive algorithms, based on a potential function that aggregates a range of learning rates (which are essential tuning parameters). By choosing the right prior we construct efficient algorithms and show that they reap both benefits by proving the first bounds that are both second-order and incorporate quantiles.

Temporal characteristics of melanopsin inputs to the human pupil light reflex

Rods, cones and melanopsin containing intrinsically photosensitive retinal ganglion cells (ipRGCs) operate in concert to regulate pupil diameter. The temporal properties of intrinsic ipRGC signalling are distinct to those of rods and cones, including longer latencies and sustained signalling after light offset. We examined whether the melanopsin mediated post-illumination pupil response (PIPR) and pupil constriction were dependent upon the inter-stimulus interval (ISI) between successive light pulses and the temporal frequency of sinusoidal light stimuli. Melanopsin excitation was altered by variation of stimulus wavelength (464 nm and 638 nm lights) and irradiance (11.4 and 15.2 log photons cm(-2) s(-1)). We found that 6s PIPR amplitude was independent of ISI and temporal frequency for all melanopsin excitation levels, indicating complete summation. In contrast to the PIPR, the maximum pupil constriction increased with increasing ISI with high and low melanopsin excitation, but time to minimum diameter was slower with high melanopsin excitation only. This melanopsin response to briefly presented pulses (16 and 100 ms) slows the temporal response of the maximum pupil constriction. We also demonstrate that high melanopsin excitation attenuates the phasic peak-trough pupil amplitude compared to conditions with low melanopsin excitation, indicating an interaction between inner and outer retinal inputs to the pupil light reflex. We infer that outer retina summation is important for rapidly controlling pupil diameter in response to short timescale fluctuations in illumination and may occur at two potential sites, one that is presynaptic to extrinsic photoreceptor input to ipRGCs, or another within the pupil control pathway if ipRGCs have differential temporal tuning to extrinsic and intrinsic signalling.

Deconstructing expected passenger experience in airports

The effect of passenger satisfaction on airport profitability has been widely acknowledged in the aviation industry. As a result, there has been much attention directed towards developing a deeper understanding of the factors that influence the passenger experience. In this paper, we explore passenger experience from a novel perspective - that of the activities expected to be undertaken by passengers while in the airport terminal building. Using the Taxonomy of Passenger Experience (TOPA) as our framework, we look at the pre-travel interview data of 48 participants. The results of our analysis are used to construct an activity-centred account of the expected passenger experience for international departures. Our exploration of the expected passenger experienced revealed that not all of the TOPA activities have an equal impact on the passengers' expected experience. The processing, consumptive, preparatory and queuing activity groups featured most prominently in passengers' accounts of their upcoming airport experiences. Of these, the preparatory category was found to have the most direct impact on passenger satisfaction. Additionally, our analysis indicated that utilising queue time to prepare passengers for upcoming processing activities could have a positive effect on both satisfaction and processing efficiency. A further outcome of this research was the observation that "shopping" did not form a part of the expected experience of any of the interviewed participants. The outcomes of this study can be used by airports to assist in the management of passengers' expected experience in the terminal building. As passenger expectations and passenger satisfaction are intrinsically linked, understanding which activities have the most impact on satisfaction provides a basis from which alternate design choices can be evaluated when constructing, or fine-tuning, airport terminal designs.

Opsin clines in butterflies suggest novel roles for insect photopigments

Opsins are ancient molecules that enable animal vision by coupling to a vitamin-derived chromophore to form lightsensitive photopigments. The primary drivers of evolutionary diversification in opsins are thought to be visual tasks related to spectral sensitivity and color vision. Typically, only a few opsin amino acid sites affect photopigment spectral sensitivity. We show that opsin genes of the North American butterfly Limenitis arthemis have diversified along a latitudinal cline, consistent with natural selection due to environmental factors. We sequenced single nucleotide(SNP) polymorphisms in the coding regions of the ultraviolet (UVRh), blue (BRh), and long-wavelength (LWRh) opsin genes from ten butterfly populations along the eastern United States and found that a majority of opsin SNPs showed significant clinal variation. Outlier detection and analysis of molecular variance indicated that many SNPs are under balancing selection and show significant population structure. This contrasts with what we found by analysing SNPs in the wingless and EF-1 alpha loci, and from neutral amplified fragment length polymorphisms, which show no evidence of significant locus-specific or genome-wide structure among populations. Using a combination of functional genetic and physiological approaches, including expression in cell culture, transgenic Drosophila, UV-visible spectroscopy, and optophysiology, we show that key BRh opsin SNPs that vary clinally have almost no effect on spectral sensitivity. Our results suggest that opsin diversification in this butterfly is more consistent with natural selection unrelated to spectral tuning. Some of the clinally varying SNPs may instead play a role in regulating opsin gene expression levels or the thermostability of the opsin protein. Lastly, we discuss the possibility that insect opsins might have important, yet-to-be elucidated, adaptive functions in mediating animal responses to abiotic factors, such as temperature or photoperiod.

Carbon nanotube-based super nanotubes: Tunable thermal conductivity in three dimensions

Advances in nanomaterials/nanostructures offer the possibility of fabricating multifunctional materials for use in engineering applications. Carbon nanotube (CNT)-based nanostructures are a representative building block for these multifunctional materials. Based on a series of in silico studies, we investigated the possibility of tuning the thermal conductivity of a three-dimensional CNT-based nanostructure: a single-walled CNT-based super-nanotube. The thermal conductivity of the super-nanotubes was shown to vary with different connecting carbon rings and super-nanotubes with longer constituent single-walled CNTs and larger diameters had a smaller thermal conductivity. The inverse of the thermal conductivity of the super-nanotubes showed a good linear relationship with the inverse of the length. The thermal conductivity was approximately proportional to the inverse of the temperature, but was insensitive to the axial strain as a result of the Poisson ratio. These results provide a fundamental understanding of the thermal conductivity of the super-nanotubes and will guide their future design/fabrication and engineering applications.

How children learn to talk and how parents can help them

From within the womb, children are tuning in to the sounds of their native language.

Fast growth of polycrystalline graphene by chemical vapor deposition of ethanol on copper

High conductive graphene films can be grown on metal foils by chemical vapor deposition (CVD). We here analyzed the use of ethanol, an economic precursor, which results also safer than commonly-used methane. A comprehensive range of process parameters were explored in order to obtain graphene films with optimal characteristics in view of their use in optoelectronics and photovoltaics. Commercially-available and electro-polished copper foils were used as substrates. By finely tuning the CVD conditions, we obtained few-layer (2-4) graphene films with good conductivity (-500 Ohm/sq) and optical transmittance around 92-94% at 550 nm on unpolished copper foils. The growth on electro-polished copper provides instead predominantly mono-layer films with lower conductivity (>1000 Ohm/sq) and with a transmittance of 97.4% at 550 nm. As for the device properties, graphene with optimal properties as transparent conductive film were produced by CVD on standard copper with specific process conditions.

Fractal and multifractal properties of a family of fractal networks

In this work, we study the fractal and multifractal properties of a family of fractal networks introduced by Gallos et al (2007 Proc. Nat. Acad. Sci. USA 104 7746). In this fractal network model, there is a parameter e which is between 0 and 1, and allows for tuning the level of fractality in the network. Here we examine the multifractal behavior of these networks, the dependence relationship of the fractal dimension and the multifractal parameters on parameter e. First, we find that the empirical fractal dimensions of these networks obtained by our program coincide with the theoretical formula given by Song et al (2006 Nature Phys. 2 275). Then from the shape of the τ(q) and D(q) curves, we find the existence of multifractality in these networks. Last, we find that there exists a linear relationship between the average information dimension 〈D(1)〉 and the parameter e.

The root locus method: application to linear stability analysis and design of cooperative car-following models

A global framework for linear stability analyses of traffic models, based on the dispersion relation root locus method, is presented and is applied taking the example of a broad class of car-following (CF) models. This approach is able to analyse all aspects of the dynamics: long waves and short wave behaviours, phase velocities and stability features. The methodology is applied to investigate the potential benefits of connected vehicles, i.e. V2V communication enabling a vehicle to send and receive information to and from surrounding vehicles. We choose to focus on the design of the coefficients of cooperation which weights the information from downstream vehicles. The coefficients tuning is performed and different ways of implementing an efficient cooperative strategy are discussed. Hence, this paper brings design methods in order to obtain robust stability of traffic models, with application on cooperative CF models

Altered nociceptive, endocrine, and dorsal horn neuron responses in rats following a neonatal immune challenge

Summary The neonatal period is characterized by significant plasticity where the immune, endocrine, and nociceptive systems undergo fine-tuning and maturation. Painful experiences during this period can result in long-term alterations in the neurocircuitry underlying nociception, including increased sensitivity to mechanical or thermal stimuli. Less is known about the impact of neonatal exposure to mild inflammatory stimuli, such as lipopolysaccharide (LPS), on subsequent inflammatory pain responses. Here we examine the impact of neonatal LPS exposure on inflammatory pain sensitivity and HPA axis activity during the first three postnatal weeks. Wistar rats were injected with LPS (0.05 mg/kg IP, Salmonella enteritidis) or saline on postnatal days (PNDs) 3 and 5 and later subjected to the formalin test at PNDs 7, 13, and 22. One hour after formalin injection, blood was collected to assess corticosterone responses. Transverse spinal cord slices were also prepared for whole-cell patch clamp recording from lumbar superficial dorsal horn neurons (SDH). Brains were obtained at PND 22 and the hypothalamus was isolated to measure glucocorticoid (GR) and mineralocorticoid receptor (MR) transcript expression using qRT-PCR. Behavioural analyses indicate that at PND 7, no significant differences were observed between saline- or LPS-challenged rats. At PND 13, LPS-challenged rats exhibited enhanced licking (p < .01), and at PND 22, increased flinching in response to formalin injection (p < .05). LPS-challenged rats also displayed increased plasma corticosterone at PND 7 and PND 22 (p < .001) but not at PND 13 following formalin administration. Furthermore, at PND 22 neonatal LPS exposure induced decreased levels of GR mRNA and increased levels of MR mRNA in the hypothalamus. The intrinsic properties of SDH neurons were similar at PND 7 and PND 13. However, at PND 22, ipsilateral SDH neurons in LPS-challenged rats had a lower input resistance compared to their saline-challenged counterparts (p < .05). These data suggest neonatal LPS exposure produces developmentally regulated changes in formalin-induced behavioural responses, corticosterone levels, and dorsal horn neuron properties following noxious stimulation later in life. These findings highlight the importance of immune activation during the neonatal period in shaping pain sensitivity later in life. This programming involves both spinal cord neurons and the HPA axis.

Metal-free graphitic carbon nitride as mechano-catalyst for hydrogen evolution reaction

Graphitic carbon nitride (g-C3N4), as a promising metal-free catalyst for photo-catalytic and electrochemical water splitting, has recently attracted tremendous research interest. However, the underlying catalytic mechanism for the hydrogen evolution reaction (HER) is not fully understood. By using density functional theory calculations, here we have established that the binding free energy of hydrogen atom (ΔGH∗0) on g-C3N4 is very sensitive to mechanical strain, leading to substantial tuning of the HER performance of g-C3N4 at different coverages. The experimentally-observed high HER activity in N-doped graphene supported g-C3N4 (Zheng et al., 2014) is actually attributed to electron-transfer induced strain. A more practical strategy to induce mechanical strain in g-C3N4 is also proposed by doping a bridge carbon atom in g-C3N4 with an isoelectronic silicon atom. The calculated ΔGH∗0 on the Si-doped g-C3N4 is ideal for HER. Our results indicate that g-C3N4 would be an excellent metal-free mechano-catalyst for HER and this finding is expected to guide future experiments to efficiently split water into hydrogen based on the g-C3N4 materials.

Pedagogical tools designed to promote engagement and active learning in the biomedical sciences and allied health disciplines

Perceiving students, science students especially, as mere consumers of facts and information belies the importance of a need to engage them with the principles underlying those facts and is counter-intuitive to the facilitation of knowledge and understanding. Traditional didactic lecture approaches need a re-think if student classroom engagement and active learning are to be valued over fact memorisation and fact recall. In our undergraduate biomedical science programs across Years 1, 2 and 3 in the Faculty of Health at QUT, we have developed an authentic learning model with an embedded suite of pedagogical strategies that foster classroom engagement and allow for active learning in the sub-discipline area of medical bacteriology. The suite of pedagogical tools we have developed have been designed to enable their translation, with appropriate fine-tuning, to most biomedical and allied health discipline teaching and learning contexts. Indeed, aspects of the pedagogy have been successfully translated to the nursing microbiology study stream at QUT. The aims underpinning the pedagogy are for our students to: (1) Connect scientific theory with scientific practice in a more direct and authentic way, (2) Construct factual knowledge and facilitate a deeper understanding, and (3) Develop and refine their higher order flexible thinking and problem solving skills, both semi-independently and independently. The mindset and role of the teaching staff is critical to this approach since for the strategy to be successful tertiary teachers need to abandon traditional instructional modalities based on one-way information delivery. Face-to-face classroom interactions between students and lecturer enable realisation of pedagogical aims (1), (2) and (3). The strategy we have adopted encourages teachers to view themselves more as expert guides in what is very much a student-focused process of scientific exploration and learning. Specific pedagogical strategies embedded in the authentic learning model we have developed include: (i) interactive lecture-tutorial hybrids or lectorials featuring teacher role-plays as well as class-level question-and-answer sessions, (ii) inclusion of “dry” laboratory activities during lectorials to prepare students for the wet laboratory to follow, (iii) real-world problem-solving exercises conducted during both lectorials and wet laboratory sessions, and (iv) designing class activities and formative assessments that probe a student’s higher order flexible thinking skills. Flexible thinking in this context encompasses analytical, critical, deductive, scientific and professional thinking modes. The strategic approach outlined above is designed to provide multiple opportunities for students to apply principles flexibly according to a given situation or context, to adapt methods of inquiry strategically, to go beyond mechanical application of formulaic approaches, and to as much as possible self-appraise their own thinking and problem solving. The pedagogical tools have been developed within both workplace (real world) and theoretical frameworks. The philosophical core of the pedagogy is a coherent pathway of teaching and learning which we, and many of our students, believe is more conducive to student engagement and active learning in the classroom. Qualitative and quantitative data derived from online and hardcopy evaluations, solicited and unsolicited student and graduate feedback, anecdotal evidence as well as peer review indicate that: (i) our students are engaging with the pedagogy, (ii) a constructivist, authentic-learning approach promotes active learning, and (iii) students are better prepared for workplace transition.

Catalytic transformation of aliphatic alcohols to corresponding esters in O2 under neutral conditions using visible-light irradiation

Selective oxidation of aliphatic alcohols under mild and base-free conditions is a challenging process for organic synthesis. Herein, we report a one-pot process for the direct oxidative esterification of aliphatic alcohols that is significantly enhanced by visible-light irradiation at ambient temperatures. The new methodology uses heterogenerous photocatalysts of gold–palladium alloy nanoparticles on a phosphate-modified hydrotalcite support and molecular oxygen as a benign oxidant. The alloy photocatalysts can absorb incident light, and the light-excited metal electrons on the surface of metal nanoparticles can activate the adsorbed reactant molecules. Tuning the light intensity and wavelength of the irradiation can remarkably change the reaction activity. Shorter wavelength light (<550 nm) drives the reaction more efficiently than light of longer wavelength (e.g., 620 nm), especially at low temperatures. The phosphate-exchanged hydrotalcite support provides sufficient basicity (and buffer) for the catalytic reactions; thus, the addition of base is not required. The photocatalysts are efficient and readily recyclable. The findings reveal the first example of using “green” oxidants and light energy to drive direct oxidative esterification of aliphatic alcohols under base-free, mild conditions.

Bayesian synthetic likelihood

Having the ability to work with complex models can be highly beneficial, but the computational cost of doing so is often large. Complex models often have intractable likelihoods, so methods that directly use the likelihood function are infeasible. In these situations, the benefits of working with likelihood-free methods become apparent. Likelihood-free methods, such as parametric Bayesian indirect likelihood that uses the likelihood of an alternative parametric auxiliary model, have been explored throughout the literature as a good alternative when the model of interest is complex. One of these methods is called the synthetic likelihood (SL), which assumes a multivariate normal approximation to the likelihood of a summary statistic of interest. This paper explores the accuracy and computational efficiency of the Bayesian version of the synthetic likelihood (BSL) approach in comparison to a competitor known as approximate Bayesian computation (ABC) and its sensitivity to its tuning parameters and assumptions. We relate BSL to pseudo-marginal methods and propose to use an alternative SL that uses an unbiased estimator of the exact working normal likelihood when the summary statistic has a multivariate normal distribution. Several applications of varying complexity are considered to illustrate the findings of this paper.