Branching Populations of Cells Bearing a Continuous Label

2000 Mathematics Subject Classification: 60J80.

Joint modelling of voicing label and continuous F0 for HMM based speech synthesis

Fundamental frequency, or F0 is critical for high quality speech synthesis in HMM based speech synthesis. Traditionally, F0 values are considered to depend on a binary voicing decision such that they are continuous in voiced regions and undefined in unvoiced regions. Multi-space distribution HMM (MSDHMM) has been used for modelling the discontinuous F0. Recently, a continuous F0 modelling framework has been proposed and shown to be effective, where continuous F0 observations are assumed to always exist and voicing labels are explicitly modelled by an independent stream. In this paper, a refined continuous F0 modelling approach is proposed. Here, F0 values are assumed to be dependent on voicing labels and both are jointly modelled in a single stream. Due to the enforced dependency, the new method can effectively reduce the voicing classification error. Subjective listening tests also demonstrate that the new approach can yield significant improvements on the naturalness of the synthesised speech. A dynamic random unvoiced F0 generation method is also investigated. Experiments show that it has significant effect on the quality of synthesised speech. © 2011 IEEE.

A Team of Continuous-Action Learning Automata for Noise-Tolerant Learning of Half-Spaces

Learning automata are adaptive decision making devices that are found useful in a variety of machine learning and pattern recognition applications. Although most learning automata methods deal with the case of finitely many actions for the automaton, there are also models of continuous-action-set learning automata (CALA). A team of such CALA can be useful in stochastic optimization problems where one has access only to noise-corrupted values of the objective function. In this paper, we present a novel formulation for noise-tolerant learning of linear classifiers using a CALA team. We consider the general case of nonuniform noise, where the probability that the class label of an example is wrong may be a function of the feature vector of the example. The objective is to learn the underlying separating hyperplane given only such noisy examples. We present an algorithm employing a team of CALA and prove, under some conditions on the class conditional densities, that the algorithm achieves noise-tolerant learning as long as the probability of wrong label for any example is less than 0.5. We also present some empirical results to illustrate the effectiveness of the algorithm.

Kernelized log linear models for continuous speech recognition

Large margin criteria and discriminative models are two effective improvements for HMM-based speech recognition. This paper proposed a large margin trained log linear model with kernels for CSR. To avoid explicitly computing in the high dimensional feature space and to achieve the nonlinear decision boundaries, a kernel based training and decoding framework is proposed in this work. To make the system robust to noise a kernel adaptation scheme is also presented. Previous work in this area is extended in two directions. First, most kernels for CSR focus on measuring the similarity between two observation sequences. The proposed joint kernels defined a similarity between two observation-label sequence pairs on the sentence level. Second, this paper addresses how to efficiently employ kernels in large margin training and decoding with lattices. To the best of our knowledge, this is the first attempt at using large margin kernel-based log linear models for CSR. The model is evaluated on a noise corrupted continuous digit task: AURORA 2.0. © 2013 IEEE.

Intermittent chemotherapy plus either intermittent or continuous cetuximab for first-line treatment of patients with KRAS wild-type advanced colorectal cancer (COIN-B): a randomised phase 2 trial

Background: Advanced colorectal cancer is treated with a combination of cytotoxic drugs and targeted treatments. However, how best to minimise the time spent taking cytotoxic drugs and whether molecular selection can refine this further is unknown. The primary aim of this study was to establish how cetuximab might be safely and effectively added to intermittent chemotherapy.

Methods: COIN-B was an open-label, multicentre, randomised, exploratory phase 2 trial done at 30 hospitals in the UK and one in Cyprus. We enrolled patients with advanced colorectal cancer who had received no previous chemotherapy for metastases. Randomisation was done centrally (by telephone) by the Medical Research Council Clinical Trials Unit using minimisation with a random element. Treatment allocation was not masked. Patients were assigned (1:1) to intermittent chemotherapy plus intermittent cetuximab or to intermittent chemotherapy plus continuous cetuximab. Chemotherapy was FOLFOX (folinic acid and oxaliplatin followed by bolus and infused fluorouracil). Patients in both groups received FOLFOX and weekly cetuximab for 12 weeks, then either had a planned interruption (those taking intermittent cetuximab) or planned maintenance by continuing on weekly cetuximab (continuous cetuximab). On RECIST progression, FOLFOX plus cetuximab or FOLFOX was recommenced for 12 weeks followed by further interruption or maintenance cetuximab, respectively. The primary outcome was failure-free survival at 10 months. The primary analysis population consisted of patients who completed 12 weeks of treatment without progression, death, or leaving the trial. We tested BRAF and NRAS status retrospectively. The trial was registered, ISRCTN38375681.

Findings: We registered 401 patients, 226 of whom were enrolled. Results for 169 with KRAS wild-type are reported here, 78 (46%) assigned to intermittent cetuximab and 91 (54%) to continuous cetuximab. 64 patients assigned to intermittent cetuximab and 66 of those assigned to continuous cetuximab were included in the primary analysis. 10-month failure-free survival was 50% (lower bound of 95% CI 39) in the intermittent group versus 52% (lower bound of 95% CI 41) in the continuous group; median failure-free survival was 12·2 months (95% CI 8·8–15·6) and 14·3 months (10·7–20·4), respectively. The most common grade 3–4 adverse events were skin rash (21 [27%] of 77 patients vs 20 [22%] of 92 patients), neutropenia (22 [29%] vs 30 [33%]), diarrhoea (14 [18%] vs 23 [25%]), and lethargy (20 [26%] vs 19 [21%]).

Interpretation: Cetuximab was safely incorporated in two first-line intermittent chemotherapy strategies. Maintenance of biological monotherapy, with less cytotoxic chemotherapy within the first 6 months, in molecularly selected patients is promising and should be validated in phase 3 trials.

An open-label, comparative study of the efficacy and safety of once-daily dose of enoxaparin versus unfractionated heparin in the treatment of proximal lower limb deep-vein thrombosis

Background: Treatment of deep-vein thrombosis (DVT) with a once-daily regimen of enoxaparin, rather than a continuous infusion of unfractionated heparin (UFH) is more convenient and allows for home care in some patients. This study was designed to compare the efficacy and safety of these two regimens for the treatment of patients with proximal lower limb DVT. Methods: 201 patients with proximal lower limb DVT from 13 centers in Brazil were randomized in an open manner to receive either enoxaparin [1.5 mg/kg subcutaneous (s.c.) OD] or intravenous (i.v.) UFH (adjusted to aPTT 1.5-2.5 times control) for 5-10 days. All patients also received warfarin (INR 2-3) for at least 3 months. The primary efficacy endpoint Was recurrent DVT (confirmed by venography or ultrasonography), and safety endpoints included bleeding and serious adverse events. The rate of pulmonary embolism (PE) was also collected. Hospitalization was at the physician's discretion. Results: Baseline patient characteristics were comparable between groups. The duration of hospital stay was significantly shorter with enoxaparin than with UFH (3 versus 7 days). In addition, 36% of patients receiving enoxaparin did not need to be hospitalized, whereas all of the patients receiving UFH were! hospitalized. The treatment duration was slightly longer with enoxaparin (8 versus 7 days). There was a nonsignificant trend toward a reduction in the rate of recurrent DVT with enoxaparin versus UFH, and similar safety. Conclusions: A once-daily regimen of enoxaparin 1.5 mg/kg subcutaneous is at least as effective and safe as conventional treatment with a continuous intravenous infusion of UFH. However, the once daily enoxaparin regimen is easier to administer (subcutaneous versus intravenous), does not require aPTT monitoring, and leads to both a reduced number of hospital admissions and an average 4-day-shorter hospital stay. (C) 2004 Elsevier Ltd. All rights reserved.

Further reductions in nonvertebral fracture rate with long-term denosumab treatment in the FREEDOM open-label extension and influence of hip bone mineral density after 3 years

Limited data exist on the efficacy of long-term therapies for osteoporosis. In osteoporotic postmenopausal women receiving denosumab for 7 years, nonvertebral fracture rates significantly decreased in years 4-7 versus years 1-3. This is the first demonstration of a further benefit on fracture outcomes with long-term therapy for osteoporosis. INTRODUCTION This study aimed to evaluate whether denosumab treatment continued beyond 3 years is associated with a further reduction in nonvertebral fracture rates. METHODS Participants who completed the 3-year placebo-controlled Fracture REduction Evaluation of Denosumab in Osteoporosis every 6 Months (FREEDOM) study were invited to participate in an open-label extension. The present analysis includes 4,074 postmenopausal women with osteoporosis (n = 2,343 long-term; n = 1,731 cross-over) who enrolled in the extension, missed ≤1 dose during their first 3 years of denosumab treatment, and continued into the fourth year of treatment. Comparison of nonvertebral fracture rates during years 1-3 of denosumab with that of the fourth year and with the rate during years 4-7 was evaluated. RESULTS For the combined group, the nonvertebral fracture rate per 100 participant-years was 2.15 for the first 3 years of denosumab treatment (referent) and 1.36 in the fourth year (rate ratio [RR] = 0.64; 95 % confidence interval (CI) = 0.48 to 0.85, p = 0.003). Comparable findings were observed in the groups separately and when nonvertebral fracture rates during years 1-3 were compared to years 4-7 in the long-term group (RR = 0.79; 95 % CI = 0.62 to 1.00, p = 0.046). Fracture rate reductions in year 4 were most prominent in subjects with persisting low hip bone mineral density (BMD). CONCLUSIONS Denosumab treatment beyond 3 years was associated with a further reduction in nonvertebral fracture rate that persisted through 7 years of continuous denosumab administration. The degree to which denosumab further reduces nonvertebral fracture risk appears influenced by the hip bone density achieved with initial therapy.

Metabolic and hormonal response to intermittent high-intensity and continuous moderate intensity exercise in individuals with type 1 diabetes: a randomised crossover study.

AIMS/HYPOTHESIS To investigate exercise-related fuel metabolism in intermittent high-intensity (IHE) and continuous moderate intensity (CONT) exercise in individuals with type 1 diabetes mellitus. METHODS In a prospective randomised open-label cross-over trial twelve male individuals with well-controlled type 1 diabetes underwent a 90 min iso-energetic cycling session at 50% maximal oxygen consumption ([Formula: see text]), with (IHE) or without (CONT) interspersed 10 s sprints every 10 min without insulin adaptation. Euglycaemia was maintained using oral (13)C-labelled glucose. (13)C Magnetic resonance spectroscopy (MRS) served to quantify hepatocellular and intramyocellular glycogen. Measurements of glucose kinetics (stable isotopes), hormones and metabolites complemented the investigation. RESULTS Glucose and insulin levels were comparable between interventions. Exogenous glucose requirements during the last 30 min of exercise were significantly lower in IHE (p = 0.02). Hepatic glucose output did not differ significantly between interventions, but glucose disposal was significantly lower in IHE (p < 0.05). There was no significant difference in glycogen consumption. Growth hormone, catecholamine and lactate levels were significantly higher in IHE (p < 0.05). CONCLUSIONS/INTERPRETATION IHE in individuals with type 1 diabetes without insulin adaptation reduced exogenous glucose requirements compared with CONT. The difference was not related to increased hepatic glucose output, nor to enhanced muscle glycogen utilisation, but to decreased glucose uptake. The lower glucose disposal in IHE implies a shift towards consumption of alternative substrates. These findings indicate a high flexibility of exercise-related fuel metabolism in type 1 diabetes, and point towards a novel and potentially beneficial role of IHE in these individuals. TRIAL REGISTRATION ClinicalTrials.gov NCT02068638 FUNDING: Swiss National Science Foundation (grant number 320030_149321/) and R&A Scherbarth Foundation (Switzerland).