Interactive reconstructive student modeling: A machine‐learning approach

Reconstructive bug modeling is a well‐known approach to student modeling in intelligent tutoring systems, suitable for modeling procedural tasks. Domain knowledge is decomposed into the set of primitive operators and the set of conditions of their applicability. Reconstructive modeling is capable of describing errors that come from irregular application of correct operators. The main obstacle to successfulness of this approach is such decomposition of domain knowledge to primitive operators with a very low level of abstraction so that bugs could never occur within them. The other drawback of this modeling scheme is its efficiency because it is usually done offline, due to vast search spaces involved.

This article reports a novel approach to reconstructive modeling based on machine‐learning techniques for inducing procedures from traces. The approach overcomes the problems of reconstructive modeling by its interactive nature. It allows online model generation by using domain knowledge and knowledge about the student to focus the search on the portion of the problem space the student is likely to traverse while solving the problem. Furthermore, the approach is not only incremental, but also truly interactive because it involves the student in explicit dialogs about his or her goals. In such a way, it is possible to determine whether the student knows the operator he or she is trying to apply. Pedagogical actions and the student model are generated interchangeably, thus allowing for dynamic adaptation of instruction, problem generation, and immediate feedback on student's errors. The approach presented is examined in the context of the symbolic integration tutoring system (SINT), an intelligent tutoring system (ITS) for the domain of symbolic integration.     

Neutralizing human monoclonal antibodies binding multiple serotypes of botulinum neurotoxin

Botulism, a disease of humans characterized by prolonged paralysis, is caused by botulinum neurotoxins (BoNTs), the most poisonous substances known. There are seven serotypes of BoNT (A-G) which differ from each other by 34-64% at the amino acid level. Each serotype is uniquely recognized by polyclonal antibodies, which originally were used to classify serotypes. To determine if there existed monoclonal antibodies (mAbs) capable of binding two or more serotypes, we evaluated the ability of 35 yeast-displayed single-chain variable fragment antibodies generated from vaccinated humans or mice for their ability to bind multiple BoNT serotypes. Two such clonally related human mAbs (1B18 and 4E17) were identified that bound BoNT serotype A (BoNT/A) and B or BoNT/A, B, E and F, respectively, with high affinity. Using molecular evolution techniques, it proved possible to both increase affinity and maintain cross-serotype reactivity for the 4E17 mAb. Both 1B18 and 4E17 bound to a relatively conserved epitope at the tip of the BoNT translocation domain. Immunoglobulin G constructed from affinity matured variants of 1B18 and 4E17 were evaluated for their ability to neutralize BoNT/B and E, respectively, in vivo. Both antibodies potently neutralized BoNT in vivo demonstrating that this epitope is functionally important in the intoxication pathway. Such cross-serotype binding and neutralizing mAbs should simplify the development of antibody-based BoNT diagnostics and therapeutics.     

Theoretical and experimental analysis of 1.3-/spl mu/m InGaAsN/GaAs lasers

We present a comprehensive theoretical and experimental analysis of 1.3-/spl mu/m InGaAsN/GaAs lasers. After introducing the 10-band k /spl middot/ p Hamiltonian which predicts transition energies observed experimentally, we employ it to investigate laser properties of ideal and real InGaAsN/GaAs laser devices. Our calculations show that the addition of N reduces the peak gain and differential gain at fixed carrier density, although the gain saturation value and the peak gain as a function of radiative current density are largely unchanged due to the incorporation of N. The gain characteristics are optimized by including the minimum amount of nitrogen necessary to prevent strain relaxation at the given well thickness. The measured spontaneous emission and gain characteristics of real devices are well described by the theoretical model. Our analysis shows that the threshold current is dominated by nonradiative, defect-related recombination. Elimination of these losses would enable laser characteristics comparable with the best InGaAsP/InP-based lasers with the added advantages provided by the GaAs system that are important for vertical integration.     

Optimization of material parameters in 1.3-/spl mu/m InGaAsN-GaAs lasers

We use a 10-band k/spl middot/p Hamiltonian to investigate gain characteristics of 1.3- /spl mu/m InGaAsN-GaAs 7-nm quantum-well lasers as a function of indium and nitrogen content. The parameters used were obtained by comparison with experimental transition energy data and fitting to measured spontaneous-emission line broadening. We conclude that optimum device performance is obtained by including the minimum amount of nitrogen necessary to prevent strain relaxation at the given well thickness.     

Middleware Challenges in Robotic Fleets for Precision Agriculture

The past decade has witnessed a huge increase in the number of proposed middleware solutions for robotic fleets operating in unstructured environments. As a result, it has become difficult to decide which middleware is the most appropriate for a specific application or application domain. In this paper we first extract a set of common and specific challenges that middlewares address, and group them according to the source domain they have originated within. These challenges are derived from a specific precision agriculture use-case based on the robotic fleet for weed control elaborated within the European project RHEA-robot fleets for highly effective agriculture and forestry management. Furthermore, the paper provides an analysis of a number of different middlewares and suggests a set of criteria for systemizing representative solutions. The aim of this analysis is to assist the process of finding an adequate middleware for a specific application domain.     

A quantitative study of radiative, Auger, and defect related recombination processes in 1.3-/spl mu/m GaInNAs-based quantum-well lasers

By measuring the spontaneous emission (SE) from normally operating /spl sim/1.3-/spl mu/m GaInNAs-GaAs-based lasers we have quantitatively determined the variation of each of the current paths present in the devices as a function of temperature from 130 K to 370 K. From the SE measurements we determine how the current I close to threshold, varies as a function of carrier density n, which enables us to separate out the main current paths corresponding to monomolecular (defect-related), radiative or Auger recombination. We find that defect-related recombination forms /spl sim/55% of the threshold current at room temperature (RT). At RT, radiative recombination accounts for /spl sim/20% of I/sub th/ with the remaining /spl sim/25% being due to nonradiative Auger recombination. Theoretical calculations of the threshold carrier, density as a function of temperature were also performed, using a ten-band k /spl middot/ p Hamiltonian. Together with the experimentally determined defect-related, radiative, and Auger currents we deduce the temperature variation of the respective recombination coefficients (A, B, and C). These are compared with theoretical calculations of the coefficients and good agreement is obtained. Our results suggest that by eliminating the dominant defect-related current path, the threshold current density of these GaInNAs-GaAs-based devices would be approximately halved at RT. Such devices could then have threshold current densities comparable with the best InGaAsP/InP-based lasers with the added advantages provided by the GaAs system that are important for vertical integration.     

High-pressure studies of recombination mechanisms in 1.3-/spl mu/m GaInNAs quantum-well lasers

The pressure dependence of the components of the recombination current at threshold in 1.3-/spl mu/m GaInNAs single quantum-well lasers is presented using for the first time high-pressure spontaneous emission measurements up to 13 kbar. It is shown that, above 6 kbar, the rapid increase of the threshold current with increasing pressure is associated with the unusual increase of the Auger-related nonradiative recombination current, while the defect-related monomolecular nonradiative recombination current is almost constant. Theoretical calculations show that the increase of the Auger current can be attributed to a large increase in the threshold carrier density with pressure, which is mainly due to the increase in the electron effective mass arising from the enhanced level-anticrossing between the GaInNAs conduction band and the nitrogen level.     

Combined orchiectomy and external radiotherapy versus radiotherapy alone for nonmetastatic prostate cancer with or without pelvic lymph node involvement: a prospective randomized study

PURPOSE: We compare the combination of orchiectomy and radiotherapy to radiotherapy alone as treatment for pelvic confined prostate cancer, that is T1-4, pN0-3, M0 (TNM classification). MATERIALS AND METHODS: In this prospective study 91 patients with clinically localized prostate cancer were, after surgical lymph node staging, randomized to receive definitive external beam radiotherapy (46) or combined orchiectomy and radiotherapy (45). Patients treated with radiotherapy alone had androgen ablation at clinical disease progression. The effects on progression-free, disease specific and overall survival rates were calculated. RESULTS: After a median followup of 9.3 years (range 6.0 to 11.4) clinical progression was seen in 61% of the radiotherapy only patients (group 1) and in 31% of the combined treatment patients (group 2) (p = 0.005). The mortality was 61 and 38% (p = 0.02), and cause specific mortality was 44 and 27%, respectively (p = 0.06), in groups 1 and 2. The differences in favor of combined treatment were mainly caused by lymph node positive tumors. For node negative tumors there was no significant difference in survival rates. CONCLUSIONS: The progression-free, disease specific and overall survival rates for patients with prostate cancer and pelvic lymph node involvement are significantly better after combined androgen ablation and radiotherapy than after radiotherapy alone. These results strongly suggest that early androgen deprivation is better than deferred endocrine treatment for these patients.     

Independent component analysis applied on gas sensor array measurement data

The performance of gas-sensor array systems is greatly influenced by the pattern recognition scheme applied on the instrument's measurement data. The traditional method of choice is principal component analysis (PCA), aiming for reduction in dimensionality and visualization of multivariate measurement data. PCA, as a second-order statistical tool, performs well in many cases, but lacks the ability to give meaningful representations for non-Gaussian data, which often is a property of gas-sensor array measurement data. If, instead, higher order statistical methods are considered for data analysis, more useful information can be extracted from the data. This paper introduces the higher order statistical method called independent component analysis (ICA) as a novel tool for analysis of gas-sensor array measurement data. A comparison between the performances of PCA and ICA is illustrated both in theory and for two sets of practical measurement data. The described experiments show that ICA is capable of handling sensor drift combined with improved discrimination, dimensionality reduction, and more adequate data representation when compared to PCA.