Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Skip to main content
James Glass

    James Glass

    ABSTRACT Spoken content retrieval refers to directly indexing and retrieving spoken content based on the audio rather than text descriptions. This potentially eliminates the requirement of producing text descriptions for multimedia... more
    ABSTRACT Spoken content retrieval refers to directly indexing and retrieving spoken content based on the audio rather than text descriptions. This potentially eliminates the requirement of producing text descriptions for multimedia content for indexing and retrieval purposes, and is able to precisely locate the exact time the desired information appears in the multimedia. Spoken content retrieval has been very successfully achieved with the basic approach of cascading automatic speech recognition (ASR) with text information retrieval: after the spoken content is transcribed into text or lattice format, a text retrieval engine searches over the ASR output to find desired information. This framework works well when the ASR accuracy is relatively high, but becomes less adequate when more challenging real-world scenarios are considered, since retrieval performance depends heavily on ASR accuracy. This challenge leads to the emergence of another approach to spoken content retrieval: to go beyond the basic framework of cascading ASR with text retrieval in order to have retrieval performances that are less dependent on ASR accuracy. This overview article is intended to provide a thorough overview of the concepts, principles, approaches, and achievements of major technical contributions along this line of investigation. This includes five major directions: 1) Modified ASR for Retrieval Purposes: cascading ASR with text retrieval, but the ASR is modified or optimized for spoken content retrieval purposes; 2) Exploiting the Information not present in ASR outputs: to try to utilize the information in speech signals inevitably lost when transcribed into phonemes and words; 3) Directly Matching at the Acoustic Level without ASR: for spoken queries, the signals can be directly matched at the acoustic level, rather than at the phoneme or word levels, bypassing all ASR issues; 4) Semantic Retrieval of Spoken Content: trying to retrieve spoken content that is semanti- ally related to the query, but not necessarily including the query terms themselves; 5) Interactive Retrieval and Efficient Presentation of the Retrieved Objects: with efficient presentation of the retrieved objects, an interactive retrieval process incorporating user actions may produce better retrieval results and user experiences.
    A pilot study was conducted in the first of two monkeys using either radiolabeled Dm-Na-P or radiolabeled hydrocortisone sodium succinate, together with lidocaine HCl. This study indicated an approximately tenfold increase in the quantity... more
    A pilot study was conducted in the first of two monkeys using either radiolabeled Dm-Na-P or radiolabeled hydrocortisone sodium succinate, together with lidocaine HCl. This study indicated an approximately tenfold increase in the quantity of Dm-Na-P delivered to the test electrodes (4 mA; 20 minutes) whereas the quantity of hydrocortisone delivered from the test electrodes was only marginally (approximately 10%) increased as compared with that from the controls. In terms of an anti-inflammatory activity, the effective dose of Dm-Na-P in all tissue layers underlying the test electrodes was at least tenfold that of the hydrocortisone. Therefore, further trials with hydrocortisone were abandoned. In the second animal, positive test electrodes (5 mA; 20 minutes, were sited over five joints on the right side of the body and matching control electrodes (0 mA; 20 minutes) were placed over corresponding joints on the left side of the body. The control and test electrodes each contained 1.0 ...
    To study a possible interaction of nuclear lamins with chromatin, we examined assembly of lamins A and C at mitotic chromosome surfaces in vitro. When a postmicrosomal supernatant of metaphase CHO cells containing disassembled lamins A... more
    To study a possible interaction of nuclear lamins with chromatin, we examined assembly of lamins A and C at mitotic chromosome surfaces in vitro. When a postmicrosomal supernatant of metaphase CHO cells containing disassembled lamins A and C is incubated with chromosomes isolated from mitotic Chinese hamster ovary cells, lamins A and C undergo dephosphorylation and uniformly coat the chromosome surfaces. Furthermore, when purified rat liver lamins A and C are dialyzed with mitotic chromosomes into a buffer of physiological ionic strength and pH, lamins A and C coat chromosomes in a similar fashion. In both cases a lamin-containing supramolecular structure is formed that remains intact when the chromatin is removed by digestion with micrococcal nuclease and extraction with 0.5 M KCl. Lamins associate with chromosomes at concentrations approximately eightfold lower than the critical concentration at which they self-assemble into insoluble structures in the absence of chromosomes, indicating that chromosome surfaces contain binding sites that promote lamin assembly. These binding sites are destroyed by brief treatment of chromosomes with trypsin or micrococcal nuclease. Together, these data suggest the existence of a specific lamin-chromatin interaction in cells that may be important for nuclear envelope reassembly and interphase chromosome structure.
    The mechanism of spermidine-induced ornithine decarboxylase (ODC, E.C. 4.1.1.17) inactivation was investigated using Chinese hamster ovary (CHO) cells, maintained in serum-free medium, which display a stabilization of ODC owing to the... more
    The mechanism of spermidine-induced ornithine decarboxylase (ODC, E.C. 4.1.1.17) inactivation was investigated using Chinese hamster ovary (CHO) cells, maintained in serum-free medium, which display a stabilization of ODC owing to the lack of accumulation of putrescine and spermidine (Glass and Gerner: Biochem. J., 236:351-357, 1986; Sertich et al.: J. Cell Physiol., 127:114-120, 1986). Treatment of cells with 10 microM exogenous spermidine leads to rapid decay of ODC activity accompanied by a parallel decrease in enzyme protein. Analysis of the decay of [35S]methionine-labeled ODC and separation by two-dimensional electrophoresis revealed no detectable modification in ODC structure during enhanced degradation. Spermidine-mediated inactivation of ODC occurred in a temperature-dependent manner exhibiting pseudo-first-order kinetics over a temperature range of 22-37 degrees C. In cultures treated continuously, an initial lag was observed after treatment with spermidine, followed by a rapid decline in activity as an apparent critical concentration of intracellular spermidine was achieved. Treating cells at 22 degrees C for 3 hours with 10 microM spermidine, followed by removal of exogenous polyamine, and then shifting to varying temperatures, resulted in rates of ODC inactivation identical with that determined with a continuous treatment. Arrhenius analysis showed that polyamine mediated inactivation of ODC occurred with an activation energy of approximately 16 kcal/mol. Treatment of cells with lysosomotrophic agents (NH4Cl, chloroquine, antipain, leupeptin, chymostatin) had no effect on ODC degradation. ODC turnover was not dependent on ubiquitin-dependent proteolysis. Shift of ts85 cells, a temperature-sensitive mutant for ubiquitin conjugation, to 39 degrees C (nonpermissive for ubiquitin-dependent proteolysis) followed by addition of spermidine led to a rapid decline in ODC activity, with a rate similar to that seen at 32 degrees C (the permissive temperature). In contrast, spermidine-mediated ODC degradation was substantially decreased by inhibitors of protein synthesis (cycloheximide, emetine, and puromycin). These data support the hypothesis that spermidine regulates ODC degradation via a mechanism requiring new protein synthesis, and that this occurs via a non-lysosomal, ubiquitin-independent pathway.
    Chinese hamster cells (line CHO) maintained in McCoy's 5A medium (modified) supplemented with insulin (10 micrograms/ml), transferrin (5 micrograms/ml), and ferrous sulfate (1.1 microgram/ml) proliferate at rates similar to... more
    Chinese hamster cells (line CHO) maintained in McCoy's 5A medium (modified) supplemented with insulin (10 micrograms/ml), transferrin (5 micrograms/ml), and ferrous sulfate (1.1 microgram/ml) proliferate at rates similar to cultures growing in the McCoy's medium supplemented with 10% fetal bovine serum. Colony-forming ability is similar in cultures supplemented with either serum or the combination of growth factors. By 6 hours after replacement of serum with growth factors, ornithine decarboxylase (ODCase) activity increases, reaching a maximum value by 24 hours after serum replacement. This maximum is cell density dependent and can exceed a 30-fold increase over enzyme activity in cultures supplemented with serum. The increased enzyme activity is due to a decrease in the turnover rate of the enzyme, based on protein synthesis inhibition studies, and an accumulation of active enzyme molecules rather than an activation of existing molecules, since the catalytic activity of ODCase, determined using the radiolabeled form of alpha-difluoromethylornithine (an enzyme-activated, irreversible inhibitor of ODCase) in concert with supplements. Intracellular putrescine and spermidine levels are substantially decreased when cultures are maintained in medium supplemented with insulin, transferrin, and ferrous sulfate, rather than serum, which is the sole source of exogenous ornithine. Titration of cultures growing in the defined medium with ornithine leads to a decrease in ODCase activity and an increase in intracellular putrescine and spermidine levels. Putrescine- and spermidine-dependent S-adenosyl-L-methionine decarboxylase activities are similar in cultures maintained in either medium. These data demonstrate that some, but not all, aspects of polyamine biosynthesis are affected by the availability of ornithine, the first substrate in the pathway.
    Approximately 1,700 communities granted cable television (CA franchises in the decade of the 1970s; more than 4,000 have done so in the 1980s. As original franchise terms expire, the question of refranchising becomes an increasingly... more
    Approximately 1,700 communities granted cable television (CA franchises in the decade of the 1970s; more than 4,000 have done so in the 1980s. As original franchise terms expire, the question of refranchising becomes an increasingly salient issue in communities: across the country. Although specialized consultants are often hired to analyze technological, financial, and legal aspects of CATV operations, consumer input
    ... 90% decrease from two million pairs in the 1950s (eg landslips onto colonies are a natural but infrequent occurrence ... RSPB Research Report No ... Cunningham, DM and Moors, PJ (1994) The decline of Rockhopper Penguins Eu-dyptes... more
    ... 90% decrease from two million pairs in the 1950s (eg landslips onto colonies are a natural but infrequent occurrence ... RSPB Research Report No ... Cunningham, DM and Moors, PJ (1994) The decline of Rockhopper Penguins Eu-dyptes chrysocome at Campbell Island, Southern ...
    ABSTRACT In this paper, we investigate the use of deep neural networks (DNNs) to generate a stacked bottleneck (SBN) feature representation for low-resource speech recognition. We examine different SBN extraction architectures, and... more
    ABSTRACT In this paper, we investigate the use of deep neural networks (DNNs) to generate a stacked bottleneck (SBN) feature representation for low-resource speech recognition. We examine different SBN extraction architectures, and incorporate low-rank matrix factorization in the final weight layer. Experiments on several low-resource languages demonstrate the effectiveness of the SBN configurations when compared to state-of-the-art hybrid DNN approaches.
    ABSTRACT
    ABSTRACT
    This paper presents a novel approach to open-vocabulary spoken utterance retrieval using confusion networks. If out-of-vocabulary (OOV) words are present in queries and the corpus, word-based indexing will not be sufficient. For this... more
    This paper presents a novel approach to open-vocabulary spoken utterance retrieval using confusion networks. If out-of-vocabulary (OOV) words are present in queries and the corpus, word-based indexing will not be sufficient. For this problem, we apply phone confusion networks and combine them with word confusion networks. With this approach, we can generate a more compact index table that enables robust
    ABSTRACT Second-language learners are often unable to find time for language practice due to constraints in their daily lives. In this paper, we examine how brief moments of waiting during a person's existing social conversations... more
    ABSTRACT Second-language learners are often unable to find time for language practice due to constraints in their daily lives. In this paper, we examine how brief moments of waiting during a person's existing social conversations can be leveraged for second language practice, even if the conversation is exchanged in the first language. We present an instant messaging (IM) prototype, WaitChatter, that supports the notion of wait-learning by displaying contextually relevant foreign language vocabulary and micro-quizzes while the user awaits a response from her conversant. The foreign translations are displayed just-in-time in the context of the conversation to promote incidental learning. In a preliminary study of WaitChatter, we found that participants were able to integrate second language learning into their existing instant messaging activities, and that a particularly opportune time to embed foreign language elements may be immediately after the learner sends a chat message.
    Page 1. The Summit Speech Recognition System: Phonological Modelling and Lexical Access1 S2.3 Victor Zue, James Glass, David Goodine, Michael Phillips, and Stephanie Seneff Spoken Language Systems Group Laboratory for Computer Science... more
    Page 1. The Summit Speech Recognition System: Phonological Modelling and Lexical Access1 S2.3 Victor Zue, James Glass, David Goodine, Michael Phillips, and Stephanie Seneff Spoken Language Systems Group Laboratory for Computer Science Massachusetts Institute of ...

    And 212 more