Online Learning versus Offline Learning

We present an off-line variant of the mistake-bound model of learning. This is an intermediate model between the on-line learning model (Littlestone, 1988, Littlestone, 1989) and the self-directed learning model (Goldman, Rivest & Schapire, 1993, Goldman & Sloan, 1994). Just like in the other two models, a learner in the off-line model has to learn an unknown concept from a sequence of elements of the instance space on which it makes guess and test trials. In all models, the aim of the learner is to make as few mistakes as possible. The difference between the models is that, while in the on-line model only the set of possible elements is known, in the off-line model the sequence of elements (i.e., the identity of the elements as well as the order in which they are to be presented) is known to the learner in advance. On the other hand, the learner is weaker than the self-directed learner, which is allowed to choose adaptively the sequence of elements presented to him.We study some of the fundamental properties of the off-line model. In particular, we compare the number of mistakes made by the off-line learner on certain concept classes to those made by the on-line and self-directed learners. We give bounds on the possible gaps between the various models and show examples that prove that our bounds are tight.Another contribution of this paper is the extension of the combinatorial tool of labeled trees to a unified approach that captures the various mistake bound measures of all the models discussed. We believe that this tool will prove to be useful for further study of models of incremental learning.     

Development of Fine Scale PZT Ceramic Fiber/Polymer Shell Composite Transducers

The relic processing technique was used to fabricate fine-scale piezoelectric lead zirconate titanate (PZT) ceramic fiber/polymer shell composites. In this technique sacrificial activated carbon fabrics were soaked in a PZT precursor solution, dried, and heat treated to form piezoceramic relics. Relics were embedded with polymer, which was allowed to cure, and the resulting composites were polished, electroded, and poled. Different facets of the composite- forming process were examined: structural modifications, soaking, firing, and polymer impregnation. The physical and electromechanical properties of the unique resulting composite were evaluated. Optimized PZT shell composites with 39 vol% ceramic exhibited the following property values: K ε200, tan δε5.5%, d ₃₃ε290 pC/N, d _hε100 pC/N, d _h g _hε6000 ± 10^-15 m²N, k _pε0.19, and k _tε0.28.     

Naturally occurring circular permutations in proteins

A pair of proteins is defined to be related by a circular permutationif the N-terminal region of one protein has significant sequencesimilarity to the C-terminal of the other and vice versa. Todetect pairs of proteins that might be related by circular permutation,we implemented a procedure based on a combination of a fastscreening algorithm that we had designed and manual verificationof candidate pairs. The screening algorithm is a variation ofa dynamic programming string matching algorithm, in which oneof the sequences is doubled. This algorithm, although not guaranteedto identify all cases of circular permutation, is a good firstindicator of protein pairs related by permutation events. Thecandidate pairs were further validated first by applicationof an exhaustive string matching algorithm and then by manualinspection using the dotplot visual tool. Screening the wholeSwissprot database, a total of 25 independent protein pairswere identified. These cases are presented here, divided intothree categories depending on the level of functional similarityof the related proteins. To validate our approach and to confirmfurther the small number of circularly permuted protein pairs,a systematic search for cases of circular permutation was carriedout in the Pfam database of protein domains. Even with thismore inclusive definition of a circular permutation, only sevenadditional candidates were found. None of these fitted our originaldefinition of circular permutations. The small number of casesof circular permutation suggests that there is no mechanismof local genetic manipulation that can induce circular permutations;most examples observed seem to result from fusion of functionalunits.     

Modelling the heart as a communication system

Electrical communication between cardiomyocytes can be perturbed during arrhythmia, but these perturbations are not captured by conventional electrocardiographic metrics. We developed a theoretical framework to quantify electrical communication using information theory metrics in two-dimensional cell lattice models of cardiac excitation propagation. The time series generated by each cell was coarse-grained to 1 when excited or 0 when resting. The Shannon entropy for each cell was calculated from the time series during four clinically important heart rhythms: normal heartbeat, anatomical reentry, spiral reentry and multiple reentry. We also used mutual information to perform spatial profiling of communication during these cardiac arrhythmias. We found that information sharing between cells was spatially heterogeneous. In addition, cardiac arrhythmia significantly impacted information sharing within the heart. Entropy localized the path of the drifting core of spiral reentry, which could be an optimal target of therapeutic ablation. We conclude that information theory metrics can quantitatively assess electrical communication among cardiomyocytes. The traditional concept of the heart as a functional syncytium sharing electrical information cannot predict altered entropy and information sharing during complex arrhythmia. Information theory metrics may find clinical application in the identification of rhythm-specific treatments which are currently unmet by traditional electrocardiographic techniques.     

复混肥料中硼含量测定方法的改进

本文提出了复混肥料中微量元素硼的化学分析方法，采用甘露醇强化硼酸后滴定的方法，其测定结果与推荐性国标相符，可以用来指导生产。     

Toxin Content of Commercial Castor Cultivars

The castor plant Ricinus communis L. is the source of castor oil which has numerous uses and is a key chemical feedstock for an array of products from polymers to cosmetics. Although castor was once widely grown throughout the world, the presence of the toxic protein ricin has deterred the re‐introduction of this useful crop. In order to develop a low‐ricin or no‐ricin castor crop, a robust, accurate method for screening castor seed for ricin content is essential. In this study, we validated a sandwich enzyme‐linked immunosorbent assay for the detection of ricin and the homologous Ricinus communis agglutinin (RCA) in castor beans and screened a collection of castor lines including commercial varieties using this assay. Although the average content derived by extracting several seeds provided a fair representation of the ricin content for a given selection, there was enough seed‐to‐seed variation to justify determining the ricin/RCA content of individual seeds. Ricin levels observed ranged from 1.16 to 6.25 % by weight.     

Experimental and theoretical study of skylight polarization transmitted through Snell's window of a flat water surface

The celestial polarization pattern may be scrambled by refraction at the air-water interface. This polarization pattern was examined in shallow waters with a submersible polarimeter, and it was calculated by using land measurements ('semiempirical predictions') and models of the skylight polarization. Semiempirically predicted and measured e-vector orientations were significantly similar. Conversely, predicted percent polarization was correlated but lower than measurements. Percent polarization depended on wavelength, where at high sun altitudes maximal percent polarization generally appeared in the UV and red spectral regions. The wavelength dependency of polarization may lead to differential spectral sensitivity in polarization-sensitive animals according to time and type of activity.     

Membrane-induced step in the activation of Sendai virus fusion protein

Peptides derived from conserved heptad-repeat regions of several viruses have been shown recently to inhibit virus-cell fusion. To find out their possible role in the fusion process, two biologically active heptad-repeat segments of the fusion protein (F) of Sendai virus, SV-150 (residues 150-186), and SV-473 (residues 473-495) were synthesized, fluorescently labeled and spectroscopically characterized for their structure and organization in solution and within the membrane. SV-150 was found to be 50-fold less active than SV-473 in inhibiting Sendai virus-cell fusion. Circular dichroism (CD) spectroscopy revealed that in aqueous solution, the peptides are self-associated and adopt low alpha-helical structure. However, when the two peptides are mixed together, their alpha-helical content significantly increases. Fluorescence studies, CD, and polarized attenuated total reflection infrared (ATR-FTIR) spectroscopy showed that both peptides, alone or as a complex, bind strongly to negatively charged and zwitterionic phospholipid membranes, dissociate therein into alpha-helical monomers, but do not perturb the lipid packing of the membrane. The ability of the peptides to interact with each other in solution may be correlated with antiviral activity, whereas their ability to interact with the membrane, together with their location near the fusion peptide and the transmembrane domain, suggests a revision to the currently accepted model for viral-induced membrane fusion. In the revised model, in the sequence of events associated with viral entry, the two heptad-repeat sequences may assist in bringing the viral and cellular membranes closer, thus facilitating membrane fusion.