Modeling Efficient Classification as a Process of Confidence Assessment and Delegation

In visual object detection and recognition, classifiers have two interesting characteristics: accuracy and speed. Accuracy depends on the complexity of the image features and classifier decision surfaces. Speed depends on the hardware and the computational effort required to use the features and decision surfaces. When attempts to increase accuracy lead to increases in complexity and effort, it is necessary to ask how much are we willing to pay for increased accuracy. For example, if increased computational effort implies quickly diminishing returns in accuracy, then those designing inexpensive surveillance applications cannot aim for maximum accuracy at any cost. It becomes necessary to find trade-offs between accuracy and effort. We study efficient classification of images depicting real-world objects and scenes. Classification is efficient when a classifier can be controlled so that the desired trade-off between accuracy and effort (speed) is achieved and unnecessary computations are avoided on a per input basis. A framework is proposed for understanding and modeling efficient classification of images. Classification is modeled as a tree-like process. In designing the framework, it is important to recognize what is essential and to avoid structures that are narrow in applicability. Earlier frameworks are lacking in this regard. The overall contribution is two-fold. First, the framework is presented, subjected to experiments, and shown to be satisfactory. Second, certain unconventional approaches are experimented with. This allows the separation of the essential from the conventional. To determine if the framework is satisfactory, three categories of questions are identified: trade-off optimization, classifier tree organization, and rules for delegation and confidence modeling. Questions and problems related to each category are addressed and empirical results are presented. For example, related to trade-off optimization, we address the problem of computational bottlenecks that limit the range of trade-offs. We also ask if accuracy versus effort trade-offs can be controlled after training. For another example, regarding classifier tree organization, we first consider the task of organizing a tree in a problem-specific manner. We then ask if problem-specific organization is necessary.

Offloading content routing cost from routers

The publish/subscribe paradigm has lately received much attention. In publish/subscribe systems, a specialized event-based middleware delivers notifications of events created by producers (publishers) to consumers (subscribers) interested in that particular event. It is considered a good approach for implementing Internet-wide distributed systems as it provides full decoupling of the communicating parties in time, space and synchronization. One flavor of the paradigm is content-based publish/subscribe which allows the subscribers to express their interests very accurately. In order to implement a content-based publish/subscribe middleware in way suitable for Internet scale, its underlying architecture must be organized as a peer-to-peer network of content-based routers that take care of forwarding the event notifications to all interested subscribers. A communication infrastructure that provides such service is called a content-based network. A content-based network is an application-level overlay network. Unfortunately, the expressiveness of the content-based interaction scheme comes with a price - compiling and maintaining the content-based forwarding and routing tables is very expensive when the amount of nodes in the network is large. The routing tables are usually partially-ordered set (poset) -based data structures. In this work, we present an algorithm that aims to improve scalability in content-based networks by reducing the workload of content-based routers by offloading some of their content routing cost to clients. We also provide experimental results of the performance of the algorithm. Additionally, we give an introduction to the publish/subscribe paradigm and content-based networking and discuss alternative ways of improving scalability in content-based networks. ACM Computing Classification System (CCS): C.2.4 [Computer-Communication Networks]: Distributed Systems - Distributed applications

Cost-effectiveness of a national initiative to improve hand hygiene compliance using the outcome of healthcare associated staphylococcus aureus bacteraemia

Background The objective is to estimate the incremental cost-effectiveness of the Australian National Hand Hygiene Inititiave implemented between 2009 and 2012 using healthcare associated Staphylococcus aureus bacteraemia as the outcome. Baseline comparators are the eight existing state and territory hand hygiene programmes. The setting is the Australian public healthcare system and 1,294,656 admissions from the 50 largest Australian hospitals are included. Methods The design is a cost-effectiveness modelling study using a before and after quasi-experimental design. The primary outcome is cost per life year saved from reduced cases of healthcare associated Staphylococcus aureus bacteraemia, with cost estimated by the annual on-going maintenance costs less the costs saved from fewer infections. Data were harvested from existing sources or were collected prospectively and the time horizon for the model was 12 months, 2011–2012. Findings No useable pre-implementation Staphylococcus aureus bacteraemia data were made available from the 11 study hospitals in Victoria or the single hospital in Northern Territory leaving 38 hospitals among six states and territories available for cost-effectiveness analyses. Total annual costs increased by $2,851,475 for a return of 96 years of life giving an incremental cost-effectiveness ratio (ICER) of $29,700 per life year gained. Probabilistic sensitivity analysis revealed a 100% chance the initiative was cost effective in the Australian Capital Territory and Queensland, with ICERs of $1,030 and $8,988 respectively. There was an 81% chance it was cost effective in New South Wales with an ICER of $33,353, a 26% chance for South Australia with an ICER of $64,729 and a 1% chance for Tasmania and Western Australia. The 12 hospitals in Victoria and the Northern Territory incur annual on-going maintenance costs of $1.51M; no information was available to describe cost savings or health benefits. Conclusions The Australian National Hand Hygiene Initiative was cost-effective against an Australian threshold of $42,000 per life year gained. The return on investment varied among the states and territories of Australia.

Improving cost-effectiveness of supplementation systems for breeder herds in northern Australia

Low level strategic supplements constitute one of the few options for northern beef producers to increase breeder productivity and profitability. Objectives of the project were to improve the cost-effectiveness of using such supplements and to improve supplement delivery systems. Urea-based supplements fed during the dry season can substantially reduce breeder liveweight loss and increase fertility during severe dry seasons. Also when fed during the late wet season these supplements increased breeder body liveweight and increased fertility of breeders in low body condition. Intake of dry lick supplements fed free choice is apparently determined primarily by the palatability of supplements relative to pasture, and training of cattle appears to be of limited importance. Siting of supplementation points has some effect on supplement intake, but little effect on grazing behaviour. Economic analysis of supplementation (urea, phosphorus or molasses) and weaning strategies was based on the relative efficacy of these strategies to maintain breeder body condition late in the dry season. Adequate body condition of breeders at this time of the year is needed to avoid mortality from under-nutrition and achieve satisfactory fertility of breeders during the following wet season. Supplements were highly cost-effective when they reduced mortality, but economic returns were generally low if the only benefit was increased fertility.

Letter from Moses Levy regarding New York congregation's agreement to pay half cost for return of Reverend H. H. Samuel Cohen

Digital image

Measuring cost and affordability of healthy and less healthy diet globally

The INFORMAS food prices module proposes a step-wise framework to measure the cost and affordability of population diets. The price differential and the tax component of healthy and less healthy foods, food groups, meals and diets will be benchmarked and monitored over time. Results can be used to model or assess the impact of fiscal policies, such as ‘fat taxes’ or subsidies. Key methodological challenges include: defining healthy and less healthy foods, meals, diets and commonly consumed items; including costs of alcohol, takeaways, convenience foods and time; selecting the price metric; sampling frameworks; and standardizing collection and analysis protocols. The minimal approach uses three complementary methods to measure the price differential between pairs of healthy and less healthy foods. Specific challenges include choosing policy relevant pairs and defining an anchor for the lists. The expanded approach measures the cost of a healthy diet compared to the current (less healthy) diet for a reference household. It requires dietary principles to guide the development of the healthy diet pricing instrument and sufficient information about the population’s current intake to inform the current (less healthy) diet tool. The optimal approach includes measures of affordability and requires a standardised measure of household income that can be used for different countries. The feasibility of implementing the protocol in different countries is being tested in New Zealand, Australia and Fiji. The impact of different decision points to address challenges will be investigated in a systematic manner. We will present early insights and results from this work.

Developing standardized methods to assess cost of healthy and unhealthy (current) diets in Australia

Unhealthy diets contribute at least 14% to Australia's disease burden and are driven by ‘obesogenic’ food environments. Compliance with dietary recommendations is particularly poor amongst disadvantaged populations including low socioeconomic groups, those living in rural/remote areas and Aboriginal and Torres Strait Islanders. The perception that healthy foods are expensive is a key barrier to healthy choices and a major determinant of diet-related health inequities. Available state/regional/local data (limited and non-comparable) suggests that, despite basic healthy foods not incurring GST, the cost of healthy food is higher and has increased more rapidly than unhealthy food over the last 15 years in Australia. However, there were no nationally standardised tools or protocols to benchmark, compare or monitor food prices and affordability in Australia. Globally, we are leading work to develop and test approaches to assess the price differential of healthy and less-healthy (current) diets under the food price module of the International Network for Food and Obesity/non-communicable diseases (NCDs) Research, Monitoring and Action Support (INFORMAS). This presentation describes contextualization of the INFORMAS approach to develop standardised Australian tools, survey protocols and data collection and analysis systems. The ‘healthy diet basket’ was based on the Australian Foundation Diet, 1 The ‘current diet basket’ and specific items included in each basket, were based on recent national dietary survey data.2 Data collection methods were piloted. The final tools and protocols were then applied to measure the price and affordability of healthy and less healthy (current) diets of different household groups in diverse communities across the nation. We have compared results for different geographical locations/population subgroups in Australia and assessed these against international INFORMAS benchmarks. The results inform the development of policy and practice, including those relevant to mooted changes to the GST base, to promote nutrition and healthy weight and prevent chronic disease in Australia.

On trade-off solution pairs in a special type of transportation problem

Abstract is not available.

Cost-effectiveness of a nonpharmacological intervention in pediatric burn care

OBJECTIVE To report the cost-effectiveness of a tailored handheld computerized procedural preparation and distraction intervention (Ditto) used during pediatric burn wound care in comparison to standard practice. METHODS An economic evaluation was performed alongside a randomized controlled trial of 75 children aged 4 to 13 years who presented with a burn to the Royal Children's Hospital, Brisbane, Australia. Participants were randomized to either the Ditto intervention (n = 35) or standard practice (n = 40) to measure the effect of the intervention on days taken for burns to re-epithelialize. Direct medical, direct nonmedical, and indirect cost data during burn re-epithelialization were extracted from the randomized controlled trial data and combined with scar management cost data obtained retrospectively from medical charts. Nonparametric bootstrapping was used to estimate statistical uncertainty in cost and effect differences and cost-effectiveness ratios. RESULTS On average, the Ditto intervention reduced the time to re-epithelialize by 3 days at AU$194 less cost for each patient compared with standard practice. The incremental cost-effectiveness plane showed that 78% of the simulated results were within the more effective and less costly quadrant and 22% were in the more effective and more costly quadrant, suggesting a 78% probability that the Ditto intervention dominates standard practice (i.e., cost-saving). At a willingness-to-pay threshold of AU$120, there is a 95% probability that the Ditto intervention is cost-effective (or cost-saving) against standard care. CONCLUSIONS This economic evaluation showed the Ditto intervention to be highly cost-effective against standard practice at a minimal cost for the significant benefits gained, supporting the implementation of the Ditto intervention during burn wound care.

Where and how to conserve: Extending the scope of spatial reserve network design

Ongoing habitat loss and fragmentation threaten much of the biodiversity that we know today. As such, conservation efforts are required if we want to protect biodiversity. Conservation budgets are typically tight, making the cost-effective selection of protected areas difficult. Therefore, reserve design methods have been developed to identify sets of sites, that together represent the species of conservation interest in a cost-effective manner. To be able to select reserve networks, data on species distributions is needed. Such data is often incomplete, but species habitat distribution models (SHDMs) can be used to link the occurrence of the species at the surveyed sites to the environmental conditions at these locations (e.g. climatic, vegetation and soil conditions). The probability of the species occurring at unvisited location is next predicted by the model, based on the environmental conditions of those sites. The spatial configuration of reserve networks is important, because habitat loss around reserves can influence the persistence of species inside the network. Since species differ in their requirements for network configuration, the spatial cohesion of networks needs to be species-specific. A way to account for species-specific requirements is to use spatial variables in SHDMs. Spatial SHDMs allow the evaluation of the effect of reserve network configuration on the probability of occurrence of the species inside the network. Even though reserves are important for conservation, they are not the only option available to conservation planners. To enhance or maintain habitat quality, restoration or maintenance measures are sometimes required. As a result, the number of conservation options per site increases. Currently available reserve selection tools do however not offer the ability to handle multiple, alternative options per site. This thesis extends the existing methodology for reserve design, by offering methods to identify cost-effective conservation planning solutions when multiple, alternative conservation options are available per site. Although restoration and maintenance measures are beneficial to certain species, they can be harmful to other species with different requirements. This introduces trade-offs between species when identifying which conservation action is best applied to which site. The thesis describes how the strength of such trade-offs can be identified, which is useful for assessing consequences of conservation decisions regarding species priorities and budget. Furthermore, the results of the thesis indicate that spatial SHDMs can be successfully used to account for species-specific requirements for spatial cohesion - in the reserve selection (single-option) context as well as in the multi-option context. Accounting for the spatial requirements of multiple species and allowing for several conservation options is however complicated, due to trade-offs in species requirements. It is also shown that spatial SHDMs can be successfully used for gaining information on factors that drive a species spatial distribution. Such information is valuable to conservation planning, as better knowledge on species requirements facilitates the design of networks for species persistence. This methods and results described in this thesis aim to improve species probabilities of persistence, by taking better account of species habitat and spatial requirements. Many real-world conservation planning problems are characterised by a variety of conservation options related to protection, restoration and maintenance of habitat. Planning tools therefore need to be able to incorporate multiple conservation options per site, in order to continue the search for cost-effective conservation planning solutions. Simultaneously, the spatial requirements of species need to be considered. The methods described in this thesis offer a starting point for combining these two relevant aspects of conservation planning.

The role of productivity in the ecological and evolutionary dynamics of predator-prey interaction

Productivity is predicted to drive the ecological and evolutionary dynamics of predator-prey interaction through changes in resource allocation between different traits. However, resources are seldom constantly available and thus temporal variation in productivity could have considerable effect on the species' potential to evolve. To study this, three long-term microbial laboratory experiments were established where Serratia marcescens prey bacteria was exposed to predation of protist Tetrahymena thermophila in different prey resource environments. The consequences of prey resource availability for the ecological properties of the predator-prey system, such as trophic dynamics, stability, and virulence, were determined. The evolutionary changes in species traits and prey genetic diversity were measured. The prey defence evolved stronger in high productivity environment. Increased allocation to defence incurred cost in terms of reduced prey resource use ability, which probably constrained prey evolution by increasing the effect of resource competition. However, the magnitude of this trade-off diminished when measured in high resource concentrations. Predation selected for white, non-pigmented, highly defensive prey clones that produced predation resistant biofilm. The biofilm defence was also potentially accompanied with cytotoxicity for predators and could have been traded off with high motility. Evidence for the evolution of predators was also found in one experiment suggesting that co-evolutionary dynamics could affect the evolution and ecology of predator-prey interaction. Temporal variation in resource availability increased variation in predator densities leading to temporally fluctuating selection for prey defences and resource use ability. Temporal variation in resource availability was also able to constrain prey evolution when the allocation to defence incurred high cost. However, when the magnitude of prey trade-off was small and the resource turnover was periodically high, temporal variation facilitated the formation of predator resistant biofilm. The evolution of prey defence constrained the transfer of energy from basal to higher trophic levels, decreasing the strength of top-down regulation on prey community. Predation and temporal variation in productivity decreased the stability of populations and prey traits in general. However, predation-induced destabilization was less pronounced in the high productivity environment where the evolution of prey defence was stronger. In addition, evolution of prey defence weakened the environmental variation induced destabilization of predator population dynamics. Moreover, protozoan predation decreased the S. marcescens virulence in the insect host moth (Parasemia plantaginis) suggesting that species interactions outside the context of host-pathogen relationship could be important indirect drivers for the evolution of pathogenesis. This thesis demonstrates that rapid evolution can affect various ecological properties of predator-prey interaction. The effect of evolution on the ecological dynamics depended on the productivity of the environment, being most evident in the constant environments with high productivity.

Dispersal and related life history traits in the Glanville fritillary butterfly

Most studies of life history evolution are based on the assumption that species exist at equilibrium and spatially distinct separated populations. In reality, this is rarely the case, as populations are often spatially structured with ephemeral local populations. Therefore, the characteristics of metapopulations should be considered while studying factors affecting life history evolution. Theoretical studies have examined spatial processes shaping the evolution of life history traits to some extent, but there is little empirical data and evidence to investigate model predictions. In my thesis I have tried to bridge the gap between theoretical and empirical studies by using the well-known Glanville fritillary (Melitaea cinxia) metapopulation as a model system. The long-term persistence of classic metapopulations requires sufficient dispersal to establish new local populations to compensate for local extinctions. Previous studies on the Glanville fritillary have shown that females establishing new populations are not a random sample from the metapopulation, but they are in fact more dispersive than females in old populations. Many other life-history traits, such as body size, fecundity and lifespan, may be related to dispersal rate. Therefore, I examined a range of correlated traits for their evolutionary and ecological consequences. I was particularly interested in how the traits vary under natural environmental conditions, hence all studies were conducted in a large (32 x 26 m) outdoor population cage built upon a natural habitat patch. Individuals for the experiments were sampled from newly-established and old populations within a large metapopulation. Results show that females originating from newly-established populations had higher within-habitat patch mobility than females from old populations. I showed that dispersal rate is heritable and that flight activity is related to variation in a gene encoding the glycolytic enzyme phosphoglucose isomerase. Both among-individual and among-population variation in dispersal are correlated with the reproductive performance of females, though I found no evidence for a trade-off between dispersal and fecundity in terms of lifetime egg production or clutch size. Instead, the results suggest that highly dispersive females from newly-established populations have a shorter lifespan than females from old populations, and that dispersive females may pay a cost in terms of reduced lifetime reproductive success due to increased time spent outside habitat patches. In summary, the results of this thesis show that genotype-dependent dispersal rate correlates with other life history traits in the Glanville fritillary, and that the rapid turnover of local populations (extinctions and re-colonisations) is likely to be the mechanism that maintains phenotypic variation in many life history traits at the metapopulation level.

Flight metabolic rate in the Glanville fritillary butterfly

Dispersal is a highly important life history trait. In fragmented landscapes the long-term persistence of populations depends on dispersal. Evolution of dispersal is affected by costs and benefits and these may differ between different landscapes. This results in differences in the strength and direction of natural selection on dispersal in fragmented landscapes. Dispersal has been shown to be a nonrandom process that is associated with traits such as flight ability in insects. This thesis examines genetic and physiological traits affecting dispersal in the Glanville fritillary butterfly (Melitaea cinxia). Flight metabolic rate is a repeatable trait representing flight ability. Unlike in many vertebrates, resting metabolic rate cannot be used as a surrogate of maximum metabolic rate as no strong correlation between the two was found in the Glanville fritillary. Resting and flight metabolic rate are affected by environmental variables, most notably temperature. However, only flight metabolic rate has a strong genetic component. Molecular variation in the much-studied candidate locus phosphoglucose isomerase (Pgi), which encodes the glycolytic enzyme PGI, has an effect on carbohydrate metabolism in flight. This effect is temperature dependent: in low to moderate temperatures individuals with the heterozygous genotype at the single nucleotide polymorphism (SNP) AA111 have higher flight metabolic rate than the common homozygous genotype. At high temperatures the situation is reversed. This finding suggests that variation in enzyme properties is indeed translated to organismal performance. High-resolution data on individual female Glanville fritillaries moving freely in the field were recorded using harmonic radar. There was a strong positive correlation between flight metabolic rate and dispersal rate. Flight metabolic rate explained one third of the observed variation in the one-hour movement distance. A fine-scaled analysis of mobility showed that mobility peaked at intermediate ambient temperatures but the two common Pgi genotypes differed in their reaction norms to temperature. As with flight metabolic rate, heterozygotes at SNP AA111 were the most active genotype in low to moderate temperatures. The results show that molecular variation is associated with variation in dispersal rate through the link of flight physiology under the influence of environmental conditions. The evolutionary pressures for dispersal differ between males and females. The effect of flight metabolic rate on dispersal was examined in both sexes in field and laboratory conditions. The relationship between flight metabolic rate and dispersal rate in the field and flight duration in the laboratory were found to differ between the two sexes. In females the relationship was positive, but in males the longest distances and flight durations were recorded for individuals with low flight metabolic rate. These findings may reflect male investment in mate locating. Instead of dispersing, males with high flight metabolic rate may establish territories and follow a perching strategy when locating females and hence move less on the landscape level. Males with low metabolic rate may be forced to disperse due to low competitive success or may show adaptations to an alternative strategy: patrolling. In the light of life history trade-offs and the rate of living theory having high metabolic rate may carry a cost in the form of shortened lifespan. Experiments relating flight metabolic rate to longevity showed a clear correlation in the opposite direction: high flight metabolic rate was associated with long lifespan. This suggests that individuals with high metabolic rate do not pay an extra physiological cost for their high flight capacity, rather there are positive correlations between different measures of fitness. These results highlight the importance of condition.

Altruism, Trade Policy, and the Optimality of Foreign Aid

There are many studies in the literature that deal with the welfare effects of income transfers between nations in a general equilibrium setting. An important impetus for this extensive literature was the demonstration of the transfer paradox; that the donor country could actually gain from a transfer of income to another, and that the recipient could lose as a result of the gift. The reason for this paradoxical result is that the transfer gives rise to a terms-of-trade effect that may be especially beneficial to the donor and detrimental to the recipient. Subsequently, many papers have established conditions under which this paradox will or will not occur. Early work by Samuelson (1954) was followed by demonstrations of paradoxes by Gale (1974), Ohyama (1974), Brecher and Bhagwati (1982) and Bhagwati, Brecher and Hatta 1983, 1985, and Dixit (1983)) among others.1 More recently, many studies have examined whether or not foreign aid — tied and untied — can be welfare improving for both the donor and the recipient (see, for example, Turunen-Red and Woodland (1988), Kemp and Wong (1993), Schweinberger (1990), Hatzipanayotou and Michael (1995), Lahiri and Raimondos-Moller 1995, 1997, Djajić, Lahiri and Raimondos-Møller 1996a, 1996b, and Lahiri, Raimondos-Møller, Wong and Woodland 1997.2