Semi-supervised model adaptation for statistical machine translation

Statistical machine translation systems are usually trained on large amounts of bilingual text (used to learn a translation model), and also large amounts of monolingual text in the target language (used to train a language model). In this article we explore the use of semi-supervised model adaptation methods for the effective use of monolingual data from the source language in order to improve translation quality. We propose several algorithms with this aim, and present the strengths and weaknesses of each one. We present detailed experimental evaluations on the French–English EuroParl data set and on data from the NIST Chinese–English large-data track. We show a significant improvement in translation quality on both tasks.     

A theoretical study of light emission from nanoscale silicon

In comparing our calculated exciton energies with those obtained from pseudopotential calculations (Ref. 27) and from a previous tight binding calculation (Ref. 30), we stated that the differences between the three semi-empirical calculations arise because of different treatment of the nanocrystal surfaces. This appears not to be correct. Subsequent calculations with variable Si-H parameters have shown that the band gap is actually rather insensitive to the actual value of these. Instead, the important feature appears to be the overall quality of the bulk band structure parameterization. References 27 and 30 use more extensive and higher quality empirical parameterizations for bulk Si than the sp³s∗ model used by us. Repeating our time dependent calculations with an improved sp³d⁵ parameterization results in similar values to those of Refs. 27 and 30 for the exciton energies.¹ The agreement of the sp³s∗ values with experimental photoluminescence energies (Fig. 7) cannot, therefore, be regarded as well understood at this time.^1,2     

New unified laws in fatigue: From the Wöhler’s to the Paris’ regime

Generalizations and unification of the celebrated Paris’ and Wöhler’s laws for fatigue crack propagation are derived by applying the recently developed quantized (or finite) fracture mechanics. In particular, three generalized Paris’, Wöhler’s or unified laws are proposed and compared, demonstrating their applicability for predicting the life time of structures containing from small (the Wöhler’s regime) to large (the Paris’ regime) propagating fatigue cracks.     

Mathematical determination of feeding patterns and its consequence on correlational studies

The cell cycle regulatory proteins, which include cyclin-dependent kinases (cdks), cdk inhibitors (CKIs), cyclins, and the pRB, and E2F families of proteins, constitute a network of interacting factors which govern exit from or passage through the mammalian cell cycle. While the proteins within these families have similar structural characteristics, each family member exhibits distinct expression patterns during embryogenesis and distinct biological activities. In order to begin to understand the tissue-specific roles of these interacting factors, we determined the expression pattern and activity of the pRB, E2F, cyclin, cdk, and CKI families of cell cycle regulatory proteins during retinoic acid-induced (neuronal pathway) and DMSO-induced (cardiac muscle pathway) differentiation of the pluripotent murine embryonal carcinoma cell line, P19. We demonstrate here that P19 terminal differentiation causes lineage-specific changes in the expression and activity of distinct members of the E2F, pRB, cyclin, and CKI families. Furthermore, dynamic changes in the activities of these cell cycle regulatory proteins occur through several overlapping mechanisms, culminating in repression of DNA-binding activity by all of the E2F family members as cells terminally differentiate.     

Attenuated hematopoietic response to granulocyte-macrophage colony-stimulating factor in patients with acquired pulmonary alveolar proteinosis

The pathogenesis of acquired pulmonary alveolar proteinosis (PAP), a rare lung disease characterized by excessive surfactant accumulation within the alveolar space, remains obscure. Gene-targeted mice lacking the hematopoietic growth factor granulocyte-macrophage colony-stimulating factor (GM-CSF) or the signal-transducing beta-common chain of the GM-CSF receptor have impaired surfactant clearance and pulmonary pathology resembling human PAP. We therefore investigated the hematopoietic effects of GM-CSF in patients with PAP. The hematologic response of 5 infants with congenital PAP to 5 microgram/kg/d was of normal magnitude. By contrast, despite normal expression of GM-CSF receptor alpha- and beta-common chains on peripheral blood myelomonocytic cells (n = 6) and normal binding affinity of bone marrow mononuclear cells for GM-CSF (n = 3), each of the 12 patients with acquired PAP treated displayed impaired responses to GM-CSF; 5 microgram/kg/d produced only minor eosinophilia, and doses of 7.5 to 20 microgram/kg were required to induce >/=1.5-fold neutrophil increments in the 3 patients who underwent dose-escalation. However, neutrophilic responses to 5 microgram/kg granulocyte colony-stimulating factor (G-CSF) were normal (n = 4). In vitro, the proportion of hematopoietic progenitors responsive to GM-CSF (16.1% +/- 8.9%; P = .042) or interleukin-3 (IL-3; 19.3% +/- 7.7%; P = .063), both of which utilize the beta-common chain of the GM-CSF receptor complex, were reduced among patients with acquired PAP (n = 4) compared with normal bone marrow donor controls (47.2% +/- 25.9% and 40.9% +/- 18.6%, respectively). In the one individual who had complete resolution of lung disease during the period of study, this was temporally associated with correction of this defective in vitro response to GM-CSF and IL-3 on serial assessment. These data establish that patients with acquired PAP have an associated impaired responsiveness to GM-CSF that is potentially pathogenic in the development of their lung disease. Based on these observations, we propose a model of the pathogenesis of acquired PAP that suggests the disease arises as a consequence of an acquired clonal disorder within the hematopoietic progenitor cell compartment.     

Joint Determination of the State Dimension and Autoregressive Order for Models with Markov Regime Switching

Abstract. This paper is concerned with the problem of joint determination of the state dimension and autoregressive order of models with Markov‐switching parameters. A model selection procedure is proposed which is based on optimization of complexity‐penalized likelihood criteria. The efficacy of the procedure is evaluated by means of Monte Carlo experiments.