Processing of speech signals for physical and sensory disabilities.

Assistive technology involving voice communication is used primarily by people who are deaf, hard of hearing, or who have speech and/or language disabilities. It is also used to a lesser extent by people with visual or motor disabilities. A very wide range of devices has been developed for people with hearing loss. These devices can be categorized not only by the modality of stimulation [i.e., auditory, visual, tactile, or direct electrical stimulation of the auditory nerve (auditory-neural)] but also in terms of the degree of speech processing that is used. At least four such categories can be distinguished: assistive devices (a) that are not designed specifically for speech, (b) that take the average characteristics of speech into account, (c) that process articulatory or phonetic characteristics of speech, and (d) that embody some degree of automatic speech recognition. Assistive devices for people with speech and/or language disabilities typically involve some form of speech synthesis or symbol generation for severe forms of language disability. Speech synthesis is also used in text-to-speech systems for sightless persons. Other applications of assistive technology involving voice communication include voice control of wheelchairs and other devices for people with mobility disabilities.

What does voice-processing technology support today?

This paper describes the state of the art in applications of voice-processing technologies. In the first part, technologies concerning the implementation of speech recognition and synthesis algorithms are described. Hardware technologies such as microprocessors and DSPs (digital signal processors) are discussed. Software development environment, which is a key technology in developing applications software, ranging from DSP software to support software also is described. In the second part, the state of the art of algorithms from the standpoint of applications is discussed. Several issues concerning evaluation of speech recognition/synthesis algorithms are covered, as well as issues concerning the robustness of algorithms in adverse conditions.

New trends in natural language processing: statistical natural language processing.

The field of natural language processing (NLP) has seen a dramatic shift in both research direction and methodology in the past several years. In the past, most work in computational linguistics tended to focus on purely symbolic methods. Recently, more and more work is shifting toward hybrid methods that combine new empirical corpus-based methods, including the use of probabilistic and information-theoretic techniques, with traditional symbolic methods. This work is made possible by the recent availability of linguistic databases that add rich linguistic annotation to corpora of natural language text. Already, these methods have led to a dramatic improvement in the performance of a variety of NLP systems with similar improvement likely in the coming years. This paper focuses on these trends, surveying in particular three areas of recent progress: part-of-speech tagging, stochastic parsing, and lexical semantics.

The future of voice-processing technology in the world of computers and communications.

This talk, which was the keynote address of the NAS Colloquium on Human-Machine Communication by Voice, discusses the past, present, and future of human-machine communications, especially speech recognition and speech synthesis. Progress in these technologies is reviewed in the context of the general progress in computer and communications technologies.

Processing and cryopreservation of placental/umbilical cord blood for unrelated bone marrow reconstitution.

Clinical evidence of hematopoietic restoration with placental/umbilical cord blood (PCB) grafts indicates that PCB can be a useful source of hematopoietic stem cells for routine bone marrow reconstitution. In the unrelated setting, human leukocyte antigen (HLA)-matched donors must be obtained for candidate patients and, hence, large panels of frozen HLA-typed PCB units must be established. The large volume of unprocessed units, consisting mostly of red blood cells, plasma, and cryopreservation medium, poses a serious difficulty in this effort because storage space in liquid nitrogen is limited and costly. We report here that almost all the hematopoietic colony-forming cells present in PCB units can be recovered in a uniform volume of 20 ml by using rouleaux formation induced by hydroxyethyl starch and centrifugation to reduce the bulk of erythrocytes and plasma and, thus, concentrate leukocytes. This method multiples the number of units that can be stored in the same freezer space as much as 10-fold depending on the format of the storage system. We have also investigated the proportion of functional stem/progenitor cells initially present that are actually available to the recipient when thawed cryopreserved PCB units are infused. Progenitor cell viability is measurably decreased when thawed cells, still suspended in hypertonic cryopreservative solutions, are rapidly mixed with large volumes of isotonic solutions or plasma. The osmotic damage inflicted by the severe solute concentration gradient, however, can be averted by a simple 2-fold dilution after thawing, providing almost total recovery of viable hematopoietic progenitor cells.

The silver gene of Drosophila melanogaster encodes multiple carboxypeptidases similar to mammalian prohormone-processing enzymes.

The silver (svr) gene of Drosophila melanogaster is required for viability, and severe mutant alleles result in death prior to eclosion. Adult flies homozygous or hemizygous for weaker alleles display several visible phenotypes, including cuticular structures that are pale and silvery in color due to reduced melanization. We have identified and cloned the DNA encoding the svr gene and determined the sequence of several partially overlapping cDNAs derived from svr mRNAs. The predicted amino acid sequence of the polypeptides encoded by these cDNAs indicates that the silver proteins are members of the family of preprotein-processing carboxypeptidases that includes the human carboxypeptidases E, M, and N. One class of svr mRNAs is alternatively spliced to encode at least two polyproteins, each of which is composed of two carboxypeptidase domains.

The hepatitis C virus NS3 serine proteinase and NS4A cofactor: establishment of a cell-free trans-processing assay.

The hepatitis C virus RNA genome encodes a long polyprotein that is proteolytically processed into at least 10 products. The order of these cleavage products in the polyprotein is NH2-C-E1-E2-p7-NS2-NS3-NS4A-NS4B-NS5A-NS5B -COOH. A serine proteinase domain located in the N-terminal one-third of nonstructural protein NS3 mediates cleavage at four downstream sites (the 3/4A, 4A/4B, 4B/5A, and 5A/5B sites). In addition to the proteinase catalytic domain, the NS4A protein is required for processing at the 4B/5A site but not at the 5A/5B site. These cleavage events are likely to be essential for virus replication, making the serine proteinase an attractive antiviral target. Here we describe an in vitro assay where the NS3-4A polyprotein, NS3, the serine proteinase domain (the N-terminal 181 residues of NS3), and the NS4A cofactor were produced by cell-free translation and tested for trans-processing of radiolabeled substrates. Polyprotein substrates, NS4A-4B or truncated NS5A-5B, were cleaved in trans by all forms of the proteinase, whereas NS4A was also required for NS4B-5A processing. Proteolysis was abolished by substitution mutations previously shown to inactivate the proteinase or block cleavage at specific sites in vivo. Furthermore, N-terminal sequence analysis established that cleavage in vitro occurred at the authentic 4A/4B site. Translation in the presence of microsomal membranes enhanced processing for some, but not all, proteinase-substrate combinations. Trans-processing was both time and temperature dependent and was eliminated by treatment with a variety of detergents above their critical micelle concentrations. Among many common proteinase inhibitors tested, only high (millimolar) concentrations of serine proteinase inhibitors tosyllysyl chloromethyl ketone and 4-(2-aminoethyl)benzenesulfonyl fluoride inactivated the NS3 proteinase. This in vitro assay should facilitate purification and further characterization of the viral serine proteinase and identification of molecules which selectively inhibit its activity.

Amyloid precursor protein processing is stimulated by metabotropic glutamate receptors.

Stimulation of muscarinic m1 or m3 receptors can, by generating diacylglycerol and activating protein kinase C, accelerate the breakdown of the amyloid precursor protein (APP) to form soluble, nonamyloidogenic derivatives (APPs), as previously shown. This relationship has been demonstrated in human glioma and neuroblastoma cells, as well as in transfected human embryonic kidney 293 cells and PC-12 cells. We now provide evidence that stimulation of metabotropic glutamate receptors (mGluRs), which also are coupled to phosphatidylinositol 4,5-bisphosphate hydrolysis, similarly accelerates processing of APP into nonamyloidogenic APPs. This process is demonstrated both in hippocampal neurons derived from fetal rats and in human embryonic kidney 293 cells transfected with cDNA expression constructs encoding the mGluR 1 alpha subtype. In hippocampal neurons, both an mGluR antagonist, L-(+)-2-amino-3-phosphonopropionic acid, and an inhibitor of protein kinase C, GF 109203X, blocked the APPs release evoked by glutamate receptor stimulation. Ionotropic glutamate agonists, N-methyl-D-aspartate or S(-)-5-fluorowillardiine, failed to affect APPs release. These data show that selective mGluR agonists that initiate signal-transduction events can regulate APP processing in bona fide primary neurons and transfected cells. As glutamatergic neurons in the cortex and hippocampus are damaged in Alzheimer disease, amyloid production in these regions may be enhanced by deficits in glutamatergic neurotransmission.

A chloroplast processing enzyme involved in precursor maturation shares a zinc-binding motif with a recently recognized family of metalloendopeptidases.

Nuclear-encoded proteins targeted to the chloroplast are typically synthesized with N-terminal transit peptides which are proteolytically removed upon import. Structurally related proteins of 145 and 143 kDa copurify with a soluble chloroplast processing enzyme (CPE) that cleaves the precursor for the major light-harvesting chlorophyll a/b binding protein and have been implicated in the maturation of the small subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase and acyl carrier protein. The 145- and 143-kDa proteins have not been found as a heterodimer and thus may represent functionally independent isoforms encoded by separate genes. Here we describe the primary structure of a 140-kDa polypeptide encoded by cDNAs isolated by using antibodies raised against the 145/143-kDa doublet. The 140-kDa polypeptide contains a transit peptide, and strikingly, a His-Xaa-Xaa-Glu-His zinc-binding motif that is conserved in a recently recognized family of metalloendopeptidases, which includes Escherichia coli protease III, insulin-degrading enzyme, and subunit beta of the mitochondrial processing peptidase. Identity of 25-30%, concentrated near the N terminus of the 140-kDa polypeptide, is found with these proteases. Expression of CPE in leaves is not light dependent. Indeed, transcripts are present in dark-grown plants, and the 145/143-kDa doublet and proteolytic activity are both found in etioplasts, as well as in root plastids. Thus, CPE appears to be a necessary component of the import machinery in photosynthetic and nonphotosynthetic tissues, and it may function as a general stromal processing peptidase in plastids.

The tRNA processing enzyme RNase T is essential for maturation of 5S RNA.

The maturation of 5S RNA in Escherichia coli is poorly understood. Although it is known that large precursors of 5S RNA accumulate in mutant cells lacking the endoribonuclease-RNase E, almost nothing is known about how the mature 5' and 3' termini of these molecules are generated. We have examined 5S RNA maturation in wild-type and single- or multiple-exoribonuclease-deficient cells by Northern blot and primer-extension analysis. Our results indicate that no mature 5S RNA is made in RNase T-deficient strains. Rather, 5S RNA precursors containing predominantly 2 extra nucleotides at the 3' end accumulate. Apparently, these 5S RNAs are functional inasmuch as mutant cells are viable, growing only slightly slower than wild type. Purified RNase T can remove the extra 3' residues, showing that it is directly involved in the trimming reaction. In contrast, mutations affecting other 3' exoribonucleases have no effect on 5S RNA maturation. Approximately 90% of the 5S RNAs in both wild-type and RNase T- cells contain mature 5' termini, indicating that 5' processing is independent of RNase T action. These data identify the enzyme responsible for generating the mature 3' terminus of 5S RNA molecules and also demonstrate that a completely processed 5S RNA molecule is not essential for cell survival.

Image processing via level set curvature flow.

We present a controlled image smoothing and enhancement method based on a curvature flow interpretation of the geometric heat equation. Compared to existing techniques, the model has several distinct advantages. (i) It contains just one enhancement parameter. (ii) The scheme naturally inherits a stopping criterion from the image; continued application of the scheme produces no further change. (iii) The method is one of the fastest possible schemes based on a curvature-controlled approach.

Communicating strategically in the face of terrorism: The Spanish government’s response to the 2004 Madrid bombing attacks

With the re-emergence of insurgency tied to terrorism, governments need to strategically manage their communications. This paper analyzes the effect of the Spanish government’s messaging in the face of the Madrid bombing of March 11, 2004: unlike what happened with the 9/11 bombings in the USA and the 7/07 London attacks, the Spanish media did not support the government’s framing of the events. Taking framing as a strategic action in a discursive form (Pan & Kosicki, 2003), and in the context of the attribution theory of responsibilities, this research uses the “cascading activation” model (Entman, 2003, 2004) to explore how a framing contest was generated in the press. Analysis of the coverage shows that the intended government frame triggered a battle among the different major newspapers, leading editorials to shift their frame over the four days prior to the national elections. This research analyzes strategic contests in framing processes and contributes insight into the interactions among the different sides (government, parties, media, and citizens) to help bring about an understanding of the rebuttal effect of the government’s intended frame. It also helps to develop an understanding of the role of the media and the influence of citizens’ frames on media content.

Investigating and Understanding Student Learning Outcomes in an Online and Face-to-Face Graduate Level Legal Administration Course: An Embedded Mixed Methods Design

Online education is a new teaching and learning medium with few current guidelines for faculty, administrators or students. Its rapid growth over the last decade has challenged academic institutions to keep up with the demand, while also providing a quality education. Our understanding of the factors that determine quality and effective online learning experiences that lead to student learning outcomes is still evolving. There is a lack of consensus on the effectiveness of online versus face-to-face education in the current research. The U.S. Department of Education conducted a meta-analysis in 2009 and concluded that student-learning outcomes in online courses were equal to and, often times, better than face-to-face traditional courses. Subsequent research has found contradictory findings, and further inquiry is necessary. The purpose of this embedded mixed methods design research study is to further our understanding of the factors that create quality and successful educational outcomes in an online course. To achieve this, the first phase of this study measured and compared learning outcomes in an online and in class graduate-level legal administration course. The second phase of the study entailed interviews with those students in both the online and face-to-face sections to understand their perspectives on the factors contributing to learning outcomes. Six themes emerged from the qualitative findings: convenience, higher order thinking, discussions, professor engagement, professor and student interaction, and face-to-face interaction. Findings from this study indicate the factors students perceive as contributing to learning outcomes in an online course are consistent among all students and are supported in the existing literature. Higher order thinking, however, emerged as a stronger theme than indicated in the current research, and the face-to-face nature of the traditional classroom may be more an issue of familiarity than a factor contributing to learning outcomes. As education continues to reach new heights and developments in technology advance, the factors found to contribute to student learning outcomes will be refined and enhanced. These developments will continue to transform the ways in which we deliver and receive knowledge in both traditional and online classrooms. While there is a growing body of research on online education, the field’s evolution has unsettled earlier findings and posed new areas to investigate.

Enhancing QA systems with complex temporal question processing capabilities

This paper presents a multilayered architecture that enhances the capabilities of current QA systems and allows different types of complex questions or queries to be processed. The answers to these questions need to be gathered from factual information scattered throughout different documents. Specifically, we designed a specialized layer to process the different types of temporal questions. Complex temporal questions are first decomposed into simple questions, according to the temporal relations expressed in the original question. In the same way, the answers to the resulting simple questions are recomposed, fulfilling the temporal restrictions of the original complex question. A novel aspect of this approach resides in the decomposition which uses a minimal quantity of resources, with the final aim of obtaining a portable platform that is easily extensible to other languages. In this paper we also present a methodology for evaluation of the decomposition of the questions as well as the ability of the implemented temporal layer to perform at a multilingual level. The temporal layer was first performed for English, then evaluated and compared with: a) a general purpose QA system (F-measure 65.47% for QA plus English temporal layer vs. 38.01% for the general QA system), and b) a well-known QA system. Much better results were obtained for temporal questions with the multilayered system. This system was therefore extended to Spanish and very good results were again obtained in the evaluation (F-measure 40.36% for QA plus Spanish temporal layer vs. 22.94% for the general QA system).

Fénix: a flexible information exchange data model for natural language processing

In this paper we describe Fénix, a data model for exchanging information between Natural Language Processing applications. The format proposed is intended to be flexible enough to cover both current and future data structures employed in the field of Computational Linguistics. The Fénix architecture is divided into four separate layers: conceptual, logical, persistence and physical. This division provides a simple interface to abstract the users from low-level implementation details, such as programming languages and data storage employed, allowing them to focus in the concepts and processes to be modelled. The Fénix architecture is accompanied by a set of programming libraries to facilitate the access and manipulation of the structures created in this framework. We will also show how this architecture has been already successfully applied in different research projects.