Programming by Demonstration Using Version Space Algebra

Programming by demonstration enables users to easily personalize their applications, automating repetitive tasks simply by executing a few examples. We formalize programming by demonstration as a machine learning problem: given the changes in the application state that result from the user's demonstrated actions, learn the general program that maps from one application state to the next. We present a methodology for learning in this space of complex functions. First we extend version spaces to learn arbitrary functions, not just concepts. Then we introduce the version space algebra, a method for composing simpler version spaces to construct more complex spaces. Finally, we apply our version space algebra to the text-editing domain and describe an implemented system called SMARTedit that learns repetitive text-editing procedures by example. We evaluate our approach by measuring the number of examples required for the system to learn a procedure thatworks on the remainder of examples, and by an informal user study measuring the effort users spend using our system versus performing the task by hand. The results show that SMARTedit is capable of generalizing correctly from as few as one or two examples, and that users generally save a significant amount of effort when completing tasks with SMARTedit's help.     

生物医用钙磷酸盐微晶玻璃

钙磷酸盐微晶玻璃因其具有生物活性、生物相容性 ,而广泛应用于牙科、骨科的替代及骨组织工程等领域。本文就钙磷酸盐微晶玻璃材料的制备工艺、组成性能、invitro vivo实验及医学应用作了较详细的评述。     

Hybrid bacterial cellulose-collagen membranes production in culture media enriched with antioxidant compounds from plant extracts

Bioactive, structural, and mechanical properties were measured of a cellulose biopolymer produced by Gluconacetobacter xylinus using hydrolyzed collagen and phenolic compounds from teas and grape pomace (GP) in an optimized culture medium. Biopolymers were impregnated with an antibiotic to investigate possible antibacterial activities. Hybrid bacterial cellulose (BC)-collagen membranes obtained from cultures with green tea and a ternary mixture (TM) of teas with GP presented a high concentration of phenolic compounds (879.7 and 1312.8 mg gallic acid equivalent per liter) and antioxidant activity (973.3 and 575.9 mmol/g), respectively. BC membranes included a nanodimensional network of microfibrils, and the addition of hydrolyzed collagen in the TM of teas in situ resulted in thicker structures with improved mechanical properties. BC-collagen membranes pretreated with cephalexin exhibited significant inhibition against Escherichia coli and Staphylococcus aureus. The enrichment of the culture media with plant phenolic compounds and collagen resulted in the formation of hybrid BC membranes in a shorter time when compared to those produced without these components. Membranes modified with bioactive compounds in situ and antibiotics ex situ can be turned into potential products for application in biomaterials with antioxidant activity.     

Comparison between heat treated sapwood and heartwood from Pinus pinaster

Sapwood and heartwood samples of Pinus pinaster were treated in an oven at 190 and 200 °C for 2–6 h. Dimensional stability, measured as Anti Shrinking Efficiency (ASE) between 0 and 65 % relative humidity, durability against fungi, mechanical resistance (MOE and MOR), hardness and chemical composition were determined for treated and untreated sapwood and heartwood. Radial ASE reached 52 % for sapwood and 50 % for heartwood, while tangential ASE reached 50 and 40 %, respectively. MOE increased slightly at the beginning of the treatment, decreasing afterwards. No significant differences were found between sapwood and heartwood. MOR decreased by 50 and 30 % for sapwood and heartwood, respectively. A significant increase in durability against Rhodonia placenta was found for both heartwood and sapwood at the higher temperature (200 °C), but at the lower temperature (190 °C) only heartwood showed good results.     

On the Optimality of the Simple Bayesian Classifier under Zero-One Loss

The simple Bayesian classifier is known to be optimal when attributes are independent given the class, but the question of whether other sufficient conditions for its optimality exist has so far not been explored. Empirical results showing that it performs surprisingly well in many domains containing clear attribute dependences suggest that the answer to this question may be positive. This article shows that, although the Bayesian classifier's probability estimates are only optimal under quadratic loss if the independence assumption holds, the classifier itself can be optimal under zero-one loss (misclassification rate) even when this assumption is violated by a wide margin. The region of quadratic-loss optimality of the Bayesian classifier is in fact a second-order infinitesimal fraction of the region of zero-one optimality. This implies that the Bayesian classifier has a much greater range of applicability than previously thought. For example, in this article it is shown to be optimal for learning conjunctions and disjunctions, even though they violate the independence assumption. Further, studies in artificial domains show that it will often outperform more powerful classifiers for common training set sizes and numbers of attributes, even if its bias is a priori much less appropriate to the domain. This article's results also imply that detecting attribute dependence is not necessarily the best way to extend the Bayesian classifier, and this is also verified empirically.     

Unifying Instance-Based and Rule-Based Induction

Several well-developed approaches to inductive learning low exist, but each has specific limitations that are hard to overcome. Multi-strategy learning attempts to tackle this problem combining multiple methods in one algorithm. This article describes a unification of two widely-used empirical approaches: rule induction and instance-based learning. In the new algorithm, instances are treated as maximally specific rules, and classification is oerformed using a best-match strategy. Rules are learned by gradually generalizing instances until no improvement in apparent accuracy is obtained. Theoretical analysis shows this approach to be efficient. It is implemented in the RISE 3.1 system. In an extensive empirical study, RISE consistently achieves higher accuracies than state-of-the-art representatives of both its parent approaches (PEBLS and CN2), as well as a decision tree learner (C4.5). Lesion studies show that eachoof RISE's components is essential to this performance. Most significantly, in 14 of the 30 domains studied, RISE is more accurate than the best of PEBLS and CN2, showing that a significant synergy can be obtained by combining multiple empirical methods.     

Analysis of Fire Hazard in Campsite Areas

The most common location for campsites is in forested or arborous areas. Given the regular use of combustible material by campers and the frequent use of campfires to cook, campsites become areas with a high propensity for the occurrence of fires. Moreover, fires occurring close to campsites can cause evacuations, disturbing their regular activities and jeopardizing tourism and likewise economic activities. While some campsites have tents and caravans within a secure distance of each other, many other campsites have the tents and caravans very concentrated in a limited area, making it much easier for fire to spread from a tent to the next. Despite the occurrence of several of these events, the associated risk is currently not well studied or documented. The work presented in this paper was performed within the Portuguese research project FireCamp, whose main purpose was the analysis and documentation of the fire risk in campsites. Several topics were addressed and some of them will be presented within this paper, specifically: (1) understanding of the mechanisms associated to burning tents; (2) analysis of the combustibility of typical accessories used in campsites such as tents, sleeping bags or camping mattresses; (3) characterization of a campsite fuel cover and its surroundings via image analysis taken by unmanned aerial vehicles; and (4) modeling of fire spread in campsites. A pilot study was carried out between 2011 and 2014 in the Campsite of Côja, in Arganil—Coimbra—Portugal. Taking into consideration that external fires frequently threaten these sites, complementary to the analysis of the fuel distribution inside the campsite, the land use analysis in the surrounding area is also of great interest. Thus, a fuel map of the Campsite of Côja was produced based on a high-resolution map of the area obtained by aerial photography produced by drones. Based on this fuel map, a stochastic model was set up to predict the fire spread in the covered area. The results obtained in this pilot study are hereby presented. Additionally, several laboratory tests were carried out in order to analyze the high calorific values (HCV) and the time to ignition (TTI) of the materials of typical camping accessories. The values of HCV found are in the range of 22.45–45.45 MJ kg^?1 and the values of TTI found are in the range of 118–224 s. These values show a high propensity for these materials to spread fire. Real tents with and without typical accessories inside (e.g., sleeping bags) were burnt in controlled environment in order to determine the mass loss decay, the increase of temperature and the convective airflows produced. Due to the convective flow inside the tent, these parameters were not considered as valid but these tests were important to understand the mechanisms associated to the burning of tents and the risk associated. The presence of combustible material inside a tent was found to be of high importance in order to sustain the combustion.     

Prediction of the thermal radiative properties of an x-ray mu-tomographied porous silica glass

A Monte Carlo ray tracing procedure is proposed to simulate thermal optical processes in heterogeneous materials. It operates within a detailed 3D image of the material, and it can therefore be used to investigate the relationship between the microstructure, the constituent optical properties, and the macroscopic radiative behavior. The program is applied to porous silica glass. A sample was first characterized by 3D x-ray tomography; then, its normal spectral emittance was calculated and compared with the experimental spectrum measured independently by high-temperature infrared emittance spectroscopy. We conclude with a discussion of the light-scattering mechanisms occurring in the sample.     

Uremic hyperhomocysteinemia: a randomized trial of folate treatment for the prevention of cardiovascular events

Homocysteine is a risk factor for atherosclerosis in the general population, and serum homocysteine levels are almost universally elevated in chronic renal failure patients. When such patients are treated with dialysis, cardiovascular disease accounts for more than 50% of their mortality, which, in some proportion, may be pathophysiologically related to the elevated serum homocysteine levels. From April 2003 to March 2005, we conducted a 2-year, double-blind, randomized, placebo-controlled trial of 186 patients with end-stage kidney disease due to any cause, who were older than 18 years and stable on hemodialysis. Patients were assigned to receive either oral folic acid 10 mg 3 times a week immediately after every dialysis session under nurse supervision or an identical-appearing placebo for the entire study. On admission, plasma total homocysteine (tHcy) levels were above 13.9 micromol/L in 96.7% of patients (median 25.0 micromol/L, range 9.3-104.0 micromol/L). In the placebo group, tHcy levels remained elevated at 6, 12, and 24 months, while oral folate significantly decreased tHcy to a median value of 10.5 (2.8-20.3) micromol/L, (p<0.01). During the study, 38 patients (folic acid group 17 vs. placebo group 21; p=0.47) died from cardiovascular disease. Kaplan-Meier life table analysis dealing with the incidence of cardiovascular events, both fatal and nonfatal (myocardial infarction, arrhythmias, angina, heart failure, cerebrovascular accident), showed that 2 years of folic acid treatment and the lowering of the homocysteine blood levels had no effect on cardiovascular events (p=0.41; hazard ratio 1.24, 95% CI 0.74-2.10). However, the carotid artery intima-media wall thickness measured in a blinded fashion decreased from 1.94 +/- 0.59 mm to 1.67 +/- 0.38 mm (p<0.01) after 2 years of folate therapy. In this short-term study of uremic patients, 2 years of folic acid supplementation normalized the tHcy blood levels in 92.3% of patients but did not change the incidence of cardiovascular events compared with the control group. However, ultrasonography of the common carotid arteries performed at entry and 24 months later showed a significant decrease in intima-media thickness with folate supplementation. This suggests that early folate supplementation may benefit patients with chronic renal failure by preventing cardiovascular deterioration.     

Economic Viability Analysis of Investment Alternatives for the Use of Biogas in a Pig Farming in the Middle East Area of Brazil

This paper presents an economic viability analysis of two investment alternatives for a pig farming located in the city of Jatai, Goias State, Brazil. The first investment alternative consists of the implementation of a biodigester, being the biogas burned in a flare. The second alternative also consists of a biodigester, nevertheless the produced biogas is used to generate electricity. The methodologies to estimate the methane production in anaerobic reactors are presented, as well as those for the quantification of the baseline and project emissions and of the electricity that can be generated. The results showed that the use of biogas for the electricity generation is the alternative more economically viable.