Bandit-Based Algorithms for Budgeted Learning

We explore the problem of budgeted machine learning, in which the learning algorithm has free access to the training examples’ labels but has to pay for each attribute that is specified. This learning model is appropriate in many areas, including medical applications. We present new algorithms for choosing which attributes to purchase of which examples in the budgeted learning model based on algorithms for the multi-armed bandit problem. All of our approaches outperformed the current state of the art. Furthermore, we present a new means for selecting an example to purchase after the attribute is selected, instead of selecting an example uniformly at random, which is typically done. Our new example selection method improved performance of all the algorithms we tested, both ours and those in the literature.

Balancing Exploration and Exploitation: A New Algorithm for Active Machine Learning

Active machine learning algorithms are used when large numbers of unlabeled examples are available and getting labels for them is costly (e.g. requiring consulting a human expert). Many conventional active learning algorithms focus on refining the decision boundary, at the expense of exploring new regions that the current hypothesis misclassifies. We propose a new active learning algorithm that balances such exploration with refining of the decision boundary by dynamically adjusting the probability to explore at each step. Our experimental results demonstrate improved performance on data sets that require extensive exploration while remaining competitive on data sets that do not. Our algorithm also shows significant tolerance of noise.

AVITROL 1973

The Federal Registration for Avitrol to be used in standing field corn restricts use to application under supervision of Governmental agencies trained in bird control. The majority of current Avitrol labels state, "For use by or under supervision of government agencies or pest control operators. Not for sale to the public." This slight difference has already caused much discussion and the matter isn't resolved yet. There are different interpretations of this statement. Also for crops the material is only registered for field corn. There is a need for the material to be registered for sweet corn and sunflowers, although I do not believe the latter were grown as widely in Ohio in 1973 as in the past few years.

High-throughput non-destructive nuclear magnetic resonance method to measure intramuscular fat content in beef

High intake of saturated fat from meats has been associated with cardiovascular disease, cancer, diabetes, and others diseases. In this paper, we are introducing a simple, high-throughput, and non-destructive low-resolution nuclear magnetic resonance method that has the potential to analyze the intramuscular fat content (IMF) in more than 1,000 beef portions per hour. The results can be used in nutritional fact labels, replacing the currently used average value. The method is based on longitudinal (T(1)) and transverse (T(2)) relaxation time information obtained by a continuous wave-free precession (CWFP) sequence. CWFP yields a higher correlation coefficient (r=0.9) than the conventional Carr-Purcell-Meiboom- Gill (CPMG) method (r=-0.25) for IMF in beef and is just as fast and a simpler pulse sequence than CPMG. The method can also be applied to other meat products.

A dog at the keyboard: using arbitrary signs to communicate requests

As a consequence of domestication, dogs have a special readiness for communication with humans. We here investigate whether a dog might be able to acquire and consistently produce a set of arbitrary signs in her communication with humans, as was demonstrated in ""linguistic"" individuals of several species. A female mongrel dog was submitted to a training schedule in which, after basic command training and after acquiring the verbal labels of rewarding objects or activities, she learned to ask for such objects or activities by selecting lexigrams and pressing keys on a keyboard. Systematic records taken during spontaneous interaction with one of the experimenters showed that lexigrams were used in an appropriate, intentional way, in accordance with the immediate motivational context. The dog only utilized the keyboard in the experimenter`s presence and gazed to him more frequently after key pressing than before, an indication that lexigram use did have communicative content. Results suggest that dogs may be able to learn a conventional system of signs associated to specific objects and activities, functionally analogous to spontaneous soliciting behaviors and point to the potential fruitfulness of the keyboard/lexigram procedure for studying dog communication and cognition. This is the first report to systematically analyze the learning of arbitrary sign production in dogs.

Particle Competition and Cooperation in Networks for Semi-Supervised Learning

Semi-supervised learning is one of the important topics in machine learning, concerning with pattern classification where only a small subset of data is labeled. In this paper, a new network-based (or graph-based) semi-supervised classification model is proposed. It employs a combined random-greedy walk of particles, with competition and cooperation mechanisms, to propagate class labels to the whole network. Due to the competition mechanism, the proposed model has a local label spreading fashion, i.e., each particle only visits a portion of nodes potentially belonging to it, while it is not allowed to visit those nodes definitely occupied by particles of other classes. In this way, a "divide-and-conquer" effect is naturally embedded in the model. As a result, the proposed model can achieve a good classification rate while exhibiting low computational complexity order in comparison to other network-based semi-supervised algorithms. Computer simulations carried out for synthetic and real-world data sets provide a numeric quantification of the performance of the method.

Structural and Luminescence Properties of Silica-Based Hybrids Containing New Silylated-Diketonato Europium(III) Complex

A new betadiketonate ligand displaying a trimethoxysilyl group as grafting function and a diketone moiety as complexing site (TTA-Si = 4,4,4-trifluoro-2-(3-trimethoxysilyl)propyl)-1-3-butanedione (C4H3S)COCH[(CH2)(3)Si(OCH3)(3)]COCF3) and its highly luminescent europium(III) complex [Eu(TTA-Si)(3)] have been synthesized and fully characterized. Luminescent silica-based hybrids have been prepared as well with this new complex grafted on the surface of dense silica nanoparticles (28 +/- 3 nm) or on mesoporous silica particles. The covalent bonding of Eu(TTA-Si)(3) inside the core of uniform silica nanoparticles (40 +/- 5 nm) was also achieved. Luminescence properties are discussed in relation to the europium chemical environment involved in each of the three hybrids. The general methodology proposed allowed high grafting ratios and overcame chelate release and tendency to agglomeration, and it could be applied to any silica matrix (in the core or at the surface, nanosized or not, dense or mesoporous) and therefore numerous applications such as luminescent markers and luminophors could be foreseen.

The types of Lygistorrhinidae and Mycetophilidae (Diptera: Bibionomorpha) in the KwaZulu-Natal Museum, Pietermaritzburg, South Africa

An annotated list of the type specimens of Lygistorrhinidae and Mycetophilidae (Diptera: Bibionomorpha) at the KwaZulu-Natal Museum, Pietermaritzburg, South Africa is provided. Information on 54 type specimens, three lygistorrhinids and 51 mycetophilids, with details of labels and actual preservation of the specimens is furnished. Locality data are georeferenced and habitus images of type specimens are provided.

Plasma kinetics of an LDL-like nanoemulsion and lipid transfer to HDL in subjects with glucose intolerance

OBJECTIVE: Glucose intolerance is frequently associated with an altered plasma lipid profile and increased cardiovascular disease risk. Nonetheless, lipid metabolism is scarcely studied in normolipidemic glucose-intolerant patients. The aim of this study was to investigate whether important lipid metabolic parameters, such as the kinetics of LDL free and esterified cholesterol and the transfer of lipids to HDL, are altered in glucose-intolerant patients with normal plasma lipids. METHODS: Fourteen glucose-intolerant patients and 15 control patients were studied; none of the patients had cardiovascular disease manifestations, and they were paired for age, sex, race and co-morbidities. A nanoemulsion resembling a LDL lipid composition (LDE) labeled with C-14-cholesteryl ester and H-3-free cholesterol was intravenously injected, and blood samples were collected over a 24-h period to determine the fractional clearance rate of the labels by compartmental analysis. The transfer of free and esterified cholesterol, triglycerides and phospholipids from the LDE to HDL was measured by the incubation of the LDE with plasma and radioactivity counting of the supernatant after chemical precipitation of non-HDL fractions. RESULTS: The levels of LDL, non-HDL and HDL cholesterol, triglycerides, apo A1 and apo B were equal in both groups. The 14 C-esterified cholesterol fractional clearance rate was not different between glucose-intolerant and control patients, but the H-3-free- cholesterol fractional clearance rate was greater in glucose-intolerant patients than in control patients. The lipid transfer to HDL was equal in both groups. CONCLUSION: In these glucose-intolerant patients with normal plasma lipids, a faster removal of LDE free cholesterol was the only lipid metabolic alteration detected in our study. This finding suggests that the dissociation of free cholesterol from lipoprotein particles occurs in normolipidemic glucose intolerance and may participate in atherogenic signaling.

Semi-supervised learning guided by the modularity measure in complex networks

Semi-supervised learning techniques have gained increasing attention in the machine learning community, as a result of two main factors: (1) the available data is exponentially increasing; (2) the task of data labeling is cumbersome and expensive, involving human experts in the process. In this paper, we propose a network-based semi-supervised learning method inspired by the modularity greedy algorithm, which was originally applied for unsupervised learning. Changes have been made in the process of modularity maximization in a way to adapt the model to propagate labels throughout the network. Furthermore, a network reduction technique is introduced, as well as an extensive analysis of its impact on the network. Computer simulations are performed for artificial and real-world databases, providing a numerical quantitative basis for the performance of the proposed method.

Metabolic oligosaccharide engineering of Plasmodium falciparum intraerythrocytic stages allows direct glycolipid analysis by mass spectrometry

A recent addition to the arsenal of tools for glycome analysis is the use of metabolic labels that allow covalent tagging of glycans with imaging probes. In this work we show that N-azidoglucosamine was successfully incorporated into glycolipidic structures of Plasmodium falciparum intraerythrocytic stages. The ability to tag glycoconjugates selectively with a fluorescent reporter group permits TLC detection of the glycolipids providing a new method to quantify dynamic changes in the glycosylation pattern and facilitating direct mass spectrometry analyses. Presence of glycosylphosphatidylinositol and glycosphingolipid structures was determined in the different extracts. Furthermore, the fluorescent tag was used as internal matrix for the MALDI experiment making even easier the analysis. (C) 2012 Elsevier B.V. All rights reserved.

Structural effect of cationic amphiphiles in diacetylenic photopolymerizable membranes

Liposomes have been an excellent option as drug delivery systems, since they are able of incorporating lipophobic and/or lipophilic drugs, reduce drug side effects, increase drug targeting, and control delivery. Also, in the last years, their use reached the field of gene therapy, as non-viral vectors for DNA delivery. As a strategy to increase system stability, the use of polymerizable phospholipids has been proposed in liposomal formulations. In this work, through differential scanning calorimetry (DSC) and electron spin resonance (ESR) of spin labels incorporated into the bilayers, we structurally characterize liposomes formed by a mixture of the polymerizable lipid diacetylenic phosphatidylcholine 1,2-bis(10,12-tricosadiynoyl)-sn-glycero-3-phosphocholine (DC8,9PC) and the zwitterionic lipid 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), in a 1:1 molar ratio. It is shown here that the polymerization efficiency of the mixture (c.a. 60%) is much higher than that of pure DC8,9PC bilayers (c.a. 20%). Cationic amphiphiles (CA) were added, in a final molar ratio of 1:1:0.2 (DC8,9PC:DMPC:CA), to make the liposomes possible carriers for genetic material, due to their electrostatic interaction with negatively charged DNA. Three amphiphiles were tested, 1,2-dioleoyl-3-trimetylammonium-propane (DOTAP), stearylamine (SA) and trimetyl (2-miristoyloxietyl) ammonium chloride (MCL), and the systems were studied before and after UV irradiation. Interestingly, the presence of the cationic amphiphiles increased liposomes polymerization. MCL displaying the strongest effect. Considering the different structural effects the three cationic amphiphiles cause in DC8,9PC bilayers, there seem to be a correlation between the degree of DC8,9PC polymerization and the packing of the membrane at the temperature it is irradiated (gel phase). Moreover, at higher temperatures, in the bilayer fluid phase, more polymerized membranes are significantly more rigid. Considering that the structure and stability of liposomes at different temperatures can be crucial for DNA binding and delivery, we expect the study presented here contributes to the production of new carrier systems with potential applications in gene therapy. (C) 2012 Elsevier Ireland Ltd. All rights reserved.

Current issues regarding health surveillance of fluoride concentrations in foods

This study analyzed the position of the Federal (Brazil), State (Sao Paulo), and municipal (Bauru, Sao Paulo) governments, civil society representatives, the regulated sector, and research associations concerning issues with fluoride content in foods. Analysis of the interviews (N = 15) used a qualitative methodology (collective subject discourse theory). Various central ideas were identified, including the need for stronger health surveillance in monitoring and controlling fluoride levels, educational measures, and more research in the area. The study concludes that the health surveillance approach to fluoride levels in foods is necessary, but still incipient. There is a mismatch between research output and surveillance. Regulation alone does not suffice to solve all the issues. Health risk communication and health education measures need to be implemented. Issues with fluoride on food labels need further research for the intervention to be effective.

Aqueous dispersions of DMPG in low salt contain leaky vesicles

Aqueous dispersions of dimyristoyl phosphatidylglycerol (DMPG), at low ionic strength, display uncommon thermal behavior. Models for such behavior need to assign a form to the lipid aggregate. Although most studies accept the presence of lipid vesicles in the lipid gel and fluid phases, this is still controversial. With electron spin resonance (ESR) spectra of spin labels incorporated into DMPG aggregates, quantification of [C-14]sucrose entrapped by the aggregates, and viscosity measurements, we demonstrate the existence of leaky vesicles in dispersions of DMPG at low ionic strength, in both gel and fluid phases of the lipid. As a control system, the ubiquitous lipid dimyristoyl phosphatidylcholine (DMPC) was used. For DMPG in the gel phase, spin labeling only indicated the presence of lipid bilayers, strongly suggesting that DMPG molecules are organized as vesicles and not micelles or bilayer fragments (bicelles), as the latter has a non-bilayer structure at the edges. Quantification of [C-14]sucrose entrapping by DMPG aggregates revealed the presence of highly leaky vesicles. Due to the short hydrocarbon chains (C-14 atoms), DMPC vesicles were also found to be partially permeable to sucrose, but not as much as DMPG vesicles. Viscosity measurements, with the calculation of the intrinsic viscosiiy of the lipid aggregate, showed that DMPG vesicles are rather similar in the gel and fluid phases, and quite different from aggregates observed along the gel-fluid transition. Taken together, our data strongly supports that DMPG forms leaky vesicles at both gel and fluid phases. (C) 2011 Elsevier Ireland Ltd. All rights reserved.

Network-based stochastic semisupervised learning

Semisupervised learning is a machine learning approach that is able to employ both labeled and unlabeled samples in the training process. In this paper, we propose a semisupervised data classification model based on a combined random-preferential walk of particles in a network (graph) constructed from the input dataset. The particles of the same class cooperate among themselves, while the particles of different classes compete with each other to propagate class labels to the whole network. A rigorous model definition is provided via a nonlinear stochastic dynamical system and a mathematical analysis of its behavior is carried out. A numerical validation presented in this paper confirms the theoretical predictions. An interesting feature brought by the competitive-cooperative mechanism is that the proposed model can achieve good classification rates while exhibiting low computational complexity order in comparison to other network-based semisupervised algorithms. Computer simulations conducted on synthetic and real-world datasets reveal the effectiveness of the model.