Background:Pneumonia is the leading cause of mortality for children below 5 years of age.The majority of these occur in poor countries with limited access to diagnosis.The World Health Organization (WHO) criterion for...Background:Pneumonia is the leading cause of mortality for children below 5 years of age.The majority of these occur in poor countries with limited access to diagnosis.The World Health Organization (WHO) criterion for pneumonia is the de facto method for diagnosis.It is designed targeting a high sensitivity and uses easy to measure parameters.The WHO criterion has poor specificity.Methods:We propose a method using common measurements (including the WHO parameters) to diagnose pneumonia at high sensitivity and specificity.Seventeen clinical features obtained from 134 subjects were used to create a series of logistic regression models.We started with one feature at a time,and continued building models with increasing number of features until we exhausted all possible combinations.We used a k-fold cross validation method to measure the performance of the models.Results:The sensitivity of our method was comparable to that of the WHO criterion but the specificity was 84%-655% higher.In the 2-11 month age group,the WHO criteria had a sensitivity and specificity of 92.0%±11.6% and 38.1%±18.5%,respectively.Our best model (using the existence of a runny nose,the number of days with runny nose,breathing rate and temperature) performed at a sensitivity of 91.3%±13.0% and specificity of 70.2%±22.80%.In the 12-60 month age group,the WHO algorithm gave a sensitivity of 95.7%±7.6% at a specificity of 9.8%±13.1%,while our corresponding sensitivity and specificity were 94.0%±12.1% and 74.0%±23.3%,respectively (using fever,number of days with cough,heart rate and chest in-drawing).Conclusions:The WHO algorithm can be improved through mathematical analysis of clinical observations and measurements routinely made in the field.The method is simple and easy to implement on a mobile phone.Our method allows the freedom to pick the best model in any arbitrary field scenario (e.g.,when an oximeter is not available).展开更多
This paper explores potential for the RAMpage memory hierarchy to use a microkernel with a small memory footprint, in a specialized cache-speed static RAM （tightly-coupled memory, TCM）. Dreamy memory is DRAM kept in...This paper explores potential for the RAMpage memory hierarchy to use a microkernel with a small memory footprint, in a specialized cache-speed static RAM （tightly-coupled memory, TCM）. Dreamy memory is DRAM kept in low-power mode, unless referenced. Simulations show that a small microkernel suits RAMpage well, in that it achieves significantly better speed and energy gains than a standard hierarchy from adding TCM. RAMpage, in its best 128KB L2 case, gained 11% speed using TCM, and reduced energy 14%. Equivalent conventional hierarchy gains were under 1%. While 1MB L2 was significantly faster against lower-energy cases for the smaller L2, the larger SRAM＇s energy does not justify the speed gain. Using a 128KB L2 cache in a conventional architecture resulted in a best-case overall run time of 2.58s, compared with the best dreamy mode run time （RAMpage without context switches on misses） of 3.34s, a speed penalty of 29%. Energy in the fastest 128KB L2 case was 2.18J vs. 1.50J, a reduction of 31%. The same RAMpage configuration without dreamy mode took 2.83s as simulated, and used 2.393, an acceptable trade-off （penalty under 10%） for being able to switch easily to a lower-energy mode.展开更多
In this paper, we propose a framework that uses localization for multi-objective optimization to simultaneously guide an evolutionary algorithm in both the decision and objective spaces. The localization is built usin...In this paper, we propose a framework that uses localization for multi-objective optimization to simultaneously guide an evolutionary algorithm in both the decision and objective spaces. The localization is built using a limited number of adaptive spheres （local models） in the decision space. These spheres axe usually guided, using some direction information, in the decision space towards the areas with non-dominated solutions. We use a second mechanism to adjust the spheres to specialize on different parts of the Paxeto front by using a guided dominance technique in the objective space. Through this interleaved guidance in both spaces, the spheres will be guided towards different parts of the Paxeto front while also exploring the decision space efficiently. The experimental results showed good performance for the local models using this dual guidance, in comparison with their original version.展开更多
基金Supported in part by the Key Program of the National Natural Science Foundation of China under Grant Nos．60723003，60505008in part by the Natural Science Foundation of Jiangsu Province of China under Grant Nos．BK2007520，BK2006116in part by the Australian Research Council(ARC)Centre for Complex Systems under Grant No．CEO0348249~~
文摘Background:Pneumonia is the leading cause of mortality for children below 5 years of age.The majority of these occur in poor countries with limited access to diagnosis.The World Health Organization (WHO) criterion for pneumonia is the de facto method for diagnosis.It is designed targeting a high sensitivity and uses easy to measure parameters.The WHO criterion has poor specificity.Methods:We propose a method using common measurements (including the WHO parameters) to diagnose pneumonia at high sensitivity and specificity.Seventeen clinical features obtained from 134 subjects were used to create a series of logistic regression models.We started with one feature at a time,and continued building models with increasing number of features until we exhausted all possible combinations.We used a k-fold cross validation method to measure the performance of the models.Results:The sensitivity of our method was comparable to that of the WHO criterion but the specificity was 84%-655% higher.In the 2-11 month age group,the WHO criteria had a sensitivity and specificity of 92.0%±11.6% and 38.1%±18.5%,respectively.Our best model (using the existence of a runny nose,the number of days with runny nose,breathing rate and temperature) performed at a sensitivity of 91.3%±13.0% and specificity of 70.2%±22.80%.In the 12-60 month age group,the WHO algorithm gave a sensitivity of 95.7%±7.6% at a specificity of 9.8%±13.1%,while our corresponding sensitivity and specificity were 94.0%±12.1% and 74.0%±23.3%,respectively (using fever,number of days with cough,heart rate and chest in-drawing).Conclusions:The WHO algorithm can be improved through mathematical analysis of clinical observations and measurements routinely made in the field.The method is simple and easy to implement on a mobile phone.Our method allows the freedom to pick the best model in any arbitrary field scenario (e.g.,when an oximeter is not available).
文摘This paper explores potential for the RAMpage memory hierarchy to use a microkernel with a small memory footprint, in a specialized cache-speed static RAM （tightly-coupled memory, TCM）. Dreamy memory is DRAM kept in low-power mode, unless referenced. Simulations show that a small microkernel suits RAMpage well, in that it achieves significantly better speed and energy gains than a standard hierarchy from adding TCM. RAMpage, in its best 128KB L2 case, gained 11% speed using TCM, and reduced energy 14%. Equivalent conventional hierarchy gains were under 1%. While 1MB L2 was significantly faster against lower-energy cases for the smaller L2, the larger SRAM＇s energy does not justify the speed gain. Using a 128KB L2 cache in a conventional architecture resulted in a best-case overall run time of 2.58s, compared with the best dreamy mode run time （RAMpage without context switches on misses） of 3.34s, a speed penalty of 29%. Energy in the fastest 128KB L2 case was 2.18J vs. 1.50J, a reduction of 31%. The same RAMpage configuration without dreamy mode took 2.83s as simulated, and used 2.393, an acceptable trade-off （penalty under 10%） for being able to switch easily to a lower-energy mode.
基金This work is supported by the Australian Research Council(ARC)Centre for Complex Systems under Grant No.CEO0348249the Postgraduate Research Student Overseas Grant from UNSW@ADFA,University of New South Wales.
文摘In this paper, we propose a framework that uses localization for multi-objective optimization to simultaneously guide an evolutionary algorithm in both the decision and objective spaces. The localization is built using a limited number of adaptive spheres （local models） in the decision space. These spheres axe usually guided, using some direction information, in the decision space towards the areas with non-dominated solutions. We use a second mechanism to adjust the spheres to specialize on different parts of the Paxeto front by using a guided dominance technique in the objective space. Through this interleaved guidance in both spaces, the spheres will be guided towards different parts of the Paxeto front while also exploring the decision space efficiently. The experimental results showed good performance for the local models using this dual guidance, in comparison with their original version.
