Australian National University

About: Australian National University is a education organization based out in Canberra, Australian Capital Territory, Australia. It is known for research contribution in the topics: Population & Galaxy. The organization has 34419 authors who have published 109261 publications receiving 4315448 citations. The organization is also known as: The Australian National University & ANU.

Eight-vertex SOS model and generalized Rogers-Ramanujan-type identities

Abstract: The eight-vertex model is equivalent to a “solid-on-solid” (SOS) model, in which an integer heightl i is associated with each sitei of the square lattice. The Boltzmann weights of the model are expressed in terms of elliptic functions of period 2K, and involve a variable parameter η. Here we begin by showing that the hard hexagon model is a special case of this eight-vertex SOS model, in which η=K/5 and the heights are restricted to the range 1⩽l i⩽4. We remark that the calculation of the sublattice densities of the hard hexagon model involves the Rogers-Ramanujan and related identities. We then go on to consider a more general eight-vertex SOS model, with η=K/r (r an integer) and 1⩽l i⩽r−1. We evaluate the local height probabilities (which are the analogs of the sublattice densities) of this model, and are automatically led to generalizations of the Rogers-Ramanujan and similar identities. The results are put into a form suitable for examining critical behavior, and exponentsβ, α, $$\bar \alpha $$ are obtained.

Online learning with kernels

Abstract: Kernel-based algorithms such as support vector machines have achieved considerable success in various problems in batch setting, where all of the training data is available in advance. Support vector machines combine the so-called kernel trick with the large margin idea. There has been little use of these methods in an online setting suitable for real-time applications. In this paper, we consider online learning in a reproducing kernel Hilbert space. By considering classical stochastic gradient descent within a feature space and the use of some straightforward tricks, we develop simple and computationally efficient algorithms for a wide range of problems such as classification, regression, and novelty detection. In addition to allowing the exploitation of the kernel trick in an online setting, we examine the value of large margins for classification in the online setting with a drifting target. We derive worst-case loss bounds, and moreover, we show the convergence of the hypothesis to the minimizer of the regularized risk functional. We present some experimental results that support the theory as well as illustrating the power of the new algorithms for online novelty detection.

Developing a National Index of Subjective Wellbeing: The Australian Unity Wellbeing Index

Abstract: The Australian Unity Wellbeing Index has beendesigned as a new barometer of Australians'satisfaction with their lives, and life inAustralia. It is based on, and develops, thetheoretical model of subjective wellbeinghomeostasis. The Index comprises two sub-scalesof Personal and National Wellbeing. Data werecollected through a nationally representativesample of 2,000 people in April/May 2001.Factor analysis confirmed the integrity of thetwo sub-scales and, confirming empiricalexpectation, the average level of lifesatisfaction was 75.5 percent of the scalemaximum score. Group comparisons revealed thatall age groups maintained their Personal Indexscore within the normal range. In addition,people in country areas were more satisfiedwith their personal lives than city-dwellers,but less satisfied about the nationalsituation, and people who had recentlyexperienced a strong positive event evidenced arise in wellbeing, whereas those who hadexperienced a strong negative event evidencedwellbeing in the low-normal range. It is arguedthat these data generally support homeostatictheory. However, an unusual result was thatfemales were more satisfied with their ownlives than males. A tentative argument isadvanced that this may represent aconstitutional difference. It is concluded thatthe Australian Unity Wellbeing Index haspotential as a valid, reliable and sensitiveinstrument to monitor national wellbeing.

SPICE: Semantic Propositional Image Caption Evaluation

Abstract: There is considerable interest in the task of automatically generating image captions. However, evaluation is challenging. Existing automatic evaluation metrics are primarily sensitive to n-gram overlap, which is neither necessary nor sufficient for the task of simulating human judgment. We hypothesize that semantic propositional content is an important component of human caption evaluation, and propose a new automated caption evaluation metric defined over scene graphs coined SPICE. Extensive evaluations across a range of models and datasets indicate that SPICE captures human judgments over model-generated captions better than other automatic metrics (e.g., system-level correlation of 0.88 with human judgments on the MS COCO dataset, versus 0.43 for CIDEr and 0.53 for METEOR). Furthermore, SPICE can answer questions such as `which caption-generator best understands colors?' and `can caption-generators count?'