scholarly journals On the effectiveness of gated echo state networks for data exhibiting long-term dependencies

Author(s):  
Daniele Di Sarli ◽  
Claudio Gallicchio ◽  
Alessio Micheli

In the context of recurrent neural networks, gated architectures such as the GRU have contributed to the development of highly accurate machine learning models that can tackle long-term dependencies in the data. However, the training of such networks is performed by the expensive algorithm of gradient descent with backpropagation through time. On the other hand, reservoir computing approaches such as Echo State Networks (ESNs) can produce models that can be trained efficiently thanks to the use of fixed random parameters, but are not ideal for dealing with data presenting long-term dependencies. We explore the problem of employing gated architectures in ESNs from both theoretical and empirical perspectives. We do so by deriving and evaluating a necessary condition for the non-contractivity of the state transition function, which is important to overcome the fading-memory characterization of conventional ESNs. We find that using pure reservoir computing methodologies is not sufficient for effective gating mechanisms, while instead training even only the gates is highly effective in terms of predictive accuracy.

Universe ◽  
2019 ◽  
Vol 5 (4) ◽  
pp. 91
Author(s):  
Valentina Raskina ◽  
Filip Křížek

The ALICE (A Large Ion Collider Experiment) experiment at CERN will upgrade its Inner Tracking System (ITS) detector. The new ITS will consist of seven coaxial cylindrical layers of ALPIDE silicon sensors which are based on Monolithic Active Pixel Sensor (MAPS) technology. We have studied the radiation hardness of ALPIDE sensors using a 30 MeV proton beam provided by the cyclotron U-120M of the Nuclear Physics Institute of the Czech Academy of Sciences in Řež. In this paper, these long-term measurements will be described. After being irradiated up to the total ionization dose 2.7 Mrad and non-ionizing energy loss 2.7 × 10 13 1 MeV n eq · cm - 2 , ALPIDE sensors fulfill ITS upgrade project technical design requirements in terms of detection efficiency and fake-hit rate.


2021 ◽  
pp. 000348942110155
Author(s):  
Leonard Haller ◽  
Khush Mehul Kharidia ◽  
Caitlin Bertelsen ◽  
Jeffrey Wang ◽  
Karla O’Dell

Objective: We sought to identify risk factors associated with long-term dysphagia, characterize changes in dysphagia over time, and evaluate the incidence of otolaryngology referrals for patients with long-term dysphagia following anterior cervical discectomy with fusion (ACDF). Methods: About 56 patients who underwent ACDF between May 2017 to February 2019 were included in the study. All patients were assessed for dysphagia using the Eating Assessment Tool (EAT-10) survey preoperatively and late postoperatively (≥1 year). Additionally, 28 patients were assessed for dysphagia early postoperatively (2 weeks—3 months). Demographic data, medical comorbidities, intraoperative details, and post-operative otolaryngology referral rates were collected from electronic medical records. Results: Of the 56 patients enrolled, 21 patients (38%) had EAT-10 scores of 3 or more at long-term follow-up. None of the demographics, comorbidities, or surgical factors assessed were associated with long-term dysphagia. Patients who reported no long-term dysphagia had a mean EAT-10 score of 6.9 early postoperatively, while patients with long-term symptoms had a mean score of 18.1 ( P = .006). Of the 21 patients who reported persistent dysphagia symptoms, 3 (14%) received dysphagia testing or otolaryngology referrals post-operatively. Conclusion: Dysphagia is a notable side effect of ACDF surgery, but there are no significant demographics, comorbidities, or surgical risk factors that predict long-term dysphagia. Early postoperative characterization of dysphagia using the EAT-10 questionnaire can help predict long-term symptoms. There is inadequate screening and otolaryngology follow-up for patients with post-ACDF dysphagia.


2020 ◽  
Vol 41 (Supplement_2) ◽  
Author(s):  
B Zafrir ◽  
R Jaffe ◽  
H Sliman ◽  
O Barnett-Griness ◽  
W Saliba

Abstract Background Lymphopenia has been shown to be associated with adverse prognosis in chronic disease states that are related to immune dysregulation. Purpose We aimed to determine the association between lymphopenia and all-cause mortality in patients presenting to coronary angiography with or without acute coronary syndromes (ACS). We also investigated whether elevated red blood cell distribution width (RDW), an established cardiovascular prognostic marker, further refines risk stratification and improves predictive accuracy beyond lymphocytes count. Methods Retrospective cohort analysis of patients undergoing coronary angiography for evaluation or treatment of coronary artery disease between 2003 and 2018. Long-term mortality risk associated with relative (1000–1500 /μL) or severe (<1000 /μL) lymphopenia was analyzed using Cox proportional hazards regression models, adjusting for comorbidities, ACS and RDW. Results Overall, 15179 patients underwent coronary angiography, at a mean age of 65±12 years. On cross-sectional analysis, lymphopenia was associated with kidney disease, cancer, heart failure and presentation with ACS, but lower rates of smoking and obesity. During a median follow-up of 7 (IQR 3.5–11.5) years, 4253 patients died. Compared to normal lymphocytes count (1500–5000 /μL), the multivariable adjusted hazard ratio (HR) (95% confidence interval) for mortality was 1.31 (1.21–1.41) and 1.97 (1.75–2.22) for relative and severe lymphopenia, respectively. The increase in mortality associated with severe lymphopenia was significant in patients presenting with non-ACS [HR 2.18 (1.74–2.73)], ST-segment elevation myocardial infarction (STEMI) [HR 1.59 (1.15–2.21)], or unstable angina/non-STEMI [HR 2.00 (1.70–2.34)]; p-for-interaction 0.626. The association of lymphopenia with mortality remained significant after additional adjustment to RDW. High RDW (>14.5%) was associated with increased mortality risk in each of the lymphocytes count groups, and improved the predictive accuracy with AUC increase from 0.609 (0.601–0.616) to 0.646 (0.639–0.654) (p<0.001). Conclusions Lymphopenia is associated with increased risk for long-term mortality in patients undergoing coronary angiography, regardless of coronary presentation. High RDW may enhance the predictive ability of lymphopenia. Lymphocyte count and mortality risk Funding Acknowledgement Type of funding source: None


Genetics ◽  
1999 ◽  
Vol 153 (2) ◽  
pp. 1009-1020 ◽  
Author(s):  
J A Woolliams ◽  
P Bijma ◽  
B Villanueva

Abstract Long-term genetic contributions (ri) measure lasting gene flow from an individual i. By accounting for linkage disequilibrium generated by selection both within and between breeding groups (categories), assuming the infinitesimal model, a general formula was derived for the expected contribution of ancestor i in category q (μi(q)), given its selective advantages (si(q)). Results were applied to overlapping generations and to a variety of modes of inheritance and selection indices. Genetic gain was related to the covariance between ri and the Mendelian sampling deviation (ai), thereby linking gain to pedigree development. When si(q) includes ai, gain was related to E[μi(q)ai], decomposing it into components attributable to within and between families, within each category, for each element of si(q). The formula for μi(q) was consistent with previous index theory for predicting gain in discrete generations. For overlapping generations, accurate predictions of gene flow were obtained among and within categories in contrast to previous theory that gave qualitative errors among categories and no predictions within. The generation interval was defined as the period for which μi(q), summed over all ancestors born in that period, equaled 1. Predictive accuracy was supported by simulation results for gain and contributions with sib-indices, BLUP selection, and selection with imprinted variation.


2021 ◽  
Vol 17 (2) ◽  
pp. 1-25
Author(s):  
Dat Tran ◽  
Christof Teuscher

Emerging memcapacitive nanoscale devices have the potential to perform computations in new ways. In this article, we systematically study, to the best of our knowledge for the first time, the computational capacity of complex memcapacitive networks, which function as reservoirs in reservoir computing, one of the brain-inspired computing architectures. Memcapacitive networks are composed of memcapacitive devices randomly connected through nanowires. Previous studies have shown that both regular and random reservoirs provide sufficient dynamics to perform simple tasks. How do complex memcapacitive networks illustrate their computational capability, and what are the topological structures of memcapacitive networks that solve complex tasks with efficiency? Studies show that small-world power-law (SWPL) networks offer an ideal trade-off between the communication properties and the wiring cost of networks. In this study, we illustrate the computing nature of SWPL memcapacitive reservoirs by exploring the two essential properties: fading memory and linear separation through measurements of kernel quality. Compared to ideal reservoirs, nanowire memcapacitive reservoirs had a better dynamic response and improved their performance by 4.67% on three tasks: MNIST, Isolated Spoken Digits, and CIFAR-10. On the same three tasks, compared to memristive reservoirs, nanowire memcapacitive reservoirs achieved comparable performance with much less power, on average, about 99× , 17×, and 277×, respectively. Simulation results of the topological transformation of memcapacitive networks reveal that that topological structures of the memcapacitive SWPL reservoirs did not affect their performance but significantly contributed to the wiring cost and the power consumption of the systems. The minimum trade-off between the wiring cost and the power consumption occurred at different network settings of α and β : 4.5 and 0.61 for Biolek reservoirs, 2.7 and 1.0 for Mohamed reservoirs, and 3.0 and 1.0 for Najem reservoirs. The results of our research illustrate the computational capacity of complex memcapacitive networks as reservoirs in reservoir computing. Such memcapacitive networks with an SWPL topology are energy-efficient systems that are suitable for low-power applications such as mobile devices and the Internet of Things.


Author(s):  
Theodosia Bartzela ◽  
Björn Theuerkauf ◽  
Elisabeth Reichardt ◽  
Malte Spielmann ◽  
Charlotte Opitz

Abstract Objectives To clinically characterize patients and family members with cleft lip and/or palate (CL/P) and associated congenital malformations or syndromes and propose possible inheritance patterns. Materials and methods An observational study of patients with CL/P, including medical and family history and intra- and extra-oral examination of their family members, was performed. Results Two hundred sixty-six patients, 1257 family members, and 42 pedigrees were included in the study. The distribution of patients according to the cleft type was 57.9% with CLP, 25.2% with cleft palate (CPO), and 12.8% with cleft lip with/without alveolus (CL/A). Seventy-four (27.8%) patients had associated malformations, and 24 (9.2%) a syndrome. The skeletal (27.7%), cardiovascular (19.3%) systems, and eyes (22.9%) were most commonly affected. Pierre Robin Sequence (7 patients) and van der Woude (4) were the most common syndromes. The majority of patients with CPO (19/24) had an associate syndrome. The families had an average of 2.45 affected members. Conclusion Individual and interfamilial phenotypic variability in patients with CL/P makes the understanding of etiopathogenesis challenging. Clinical relevance The overall prevalence of individuals with CL/P and their pedigrees with associated malformations and syndromes emphasize the need for early identification, interdisciplinary, and long-term planning.


2016 ◽  
Vol 283 (1823) ◽  
pp. 20152404 ◽  
Author(s):  
Jorge Velázquez ◽  
Robert B. Allen ◽  
David A. Coomes ◽  
Markus P. Eichhorn

Plant sizes within populations often exhibit multimodal distributions, even when all individuals are the same age and have experienced identical conditions. To establish the causes of this, we created an individual-based model simulating the growth of trees in a spatially explicit framework, which was parametrized using data from a long-term study of forest stands in New Zealand. First, we demonstrate that asymmetric resource competition is a necessary condition for the formation of multimodal size distributions within cohorts. By contrast, the legacy of small-scale clustering during recruitment is transient and quickly overwhelmed by density-dependent mortality. Complex multi-layered size distributions are generated when established individuals are restricted in the spatial domain within which they can capture resources. The number of modes reveals the effective number of direct competitors, while the separation and spread of modes are influenced by distances among established individuals. Asymmetric competition within local neighbourhoods can therefore generate a range of complex size distributions within even-aged cohorts.


2017 ◽  
Vol 107 (3) ◽  
pp. 362-368 ◽  
Author(s):  
Wayne M. Jurick ◽  
Otilia Macarisin ◽  
Verneta L. Gaskins ◽  
Eunhee Park ◽  
Jiujiang Yu ◽  
...  

Botrytis cinerea causes gray mold and is an economically important postharvest pathogen of fruit, vegetables, and ornamentals. Fludioxonil-sensitive B. cinerea isolates were collected in 2011 and 2013 from commercial storage in Pennsylvania. Eight isolates had values for effective concentrations for inhibiting 50% of mycelial growth of 0.0004 to 0.0038 μg/ml for fludioxonil and were dual resistant to pyrimethanil and thiabendazole. Resistance was generated in vitro, following exposure to a sublethal dose of fludioxonil, in seven of eight dual-resistant B. cinerea isolates. Three vigorously growing B. cinerea isolates with multiresistance to postharvest fungicides were further characterized and found to be osmosensitive and retained resistance in the absence of selection pressure. A representative multiresistant B. cinerea strain caused decay on apple fruit treated with postharvest fungicides, which confirmed the in vitro results. The R632I mutation in the Mrr1 gene, associated with fludioxonil resistance in B. cinerea, was not detected in multipostharvest fungicide-resistant B. cinerea isolates, suggesting that the fungus may be using additional mechanisms to mediate resistance. Results from this study show for the first time that B. cinerea with dual resistance to pyrimethanil and thiabendazole can also rapidly develop resistance to fludioxonil, which may pose control challenges in the packinghouse environment and during long-term storage.


Sign in / Sign up

Export Citation Format

Share Document