scholarly journals State predictive information bottleneck

2021 ◽  
Vol 154 (13) ◽  
pp. 134111
Author(s):  
Dedi Wang ◽  
Pratyush Tiwary
Keyword(s):  
Predictive Information ◽  
Information Bottleneck

scholarly journals Context Based Predictive Information

Entropy ◽  
2019 ◽  
Vol 21 (7) ◽  
pp. 645
Author(s):  
Yuval Shalev ◽  
Irad Ben-Gal
Keyword(s):  
Time Series ◽  
Stock Prices ◽  
Lossy Compression ◽  
Compression Algorithm ◽  
Predictive Information ◽  
Real Dataset ◽  
Information Bottleneck ◽  
The U.S

We propose a new algorithm called the context-based predictive information (CBPI) for estimating the predictive information (PI) between time series, by utilizing a lossy compression algorithm. The advantage of this approach over existing methods resides in the case of sparse predictive information (SPI) conditions, where the ratio between the number of informative sequences to uninformative sequences is small. It is shown that the CBPI achieves a better PI estimation than benchmark methods by ignoring uninformative sequences while improving explainability by identifying the informative sequences. We also provide an implementation of the CBPI algorithm on a real dataset of large banks’ stock prices in the U.S. In the last part of this paper, we show how the CBPI algorithm is related to the well-known information bottleneck in its deterministic version.

scholarly journals Getting High: High Fidelity Simulation of High Granularity Calorimeters with High Speed

2021 ◽  
Vol 5 (1) ◽  
Author(s):  
Erik Buhmann ◽  
Sascha Diefenbacher ◽  
Engin Eren ◽  
Frank Gaede ◽  
Gregor Kasieczka ◽  
...  
Keyword(s):  
Particle Physics ◽  
High Speed ◽  
Deep Neural Networks ◽  
High Fidelity ◽  
High Fidelity Simulation ◽  
Physical Processes ◽  
Electromagnetic Showers ◽  
Information Bottleneck ◽  
First Time ◽  
Processing Network

AbstractAccurate simulation of physical processes is crucial for the success of modern particle physics. However, simulating the development and interaction of particle showers with calorimeter detectors is a time consuming process and drives the computing needs of large experiments at the LHC and future colliders. Recently, generative machine learning models based on deep neural networks have shown promise in speeding up this task by several orders of magnitude. We investigate the use of a new architecture—the Bounded Information Bottleneck Autoencoder—for modelling electromagnetic showers in the central region of the Silicon-Tungsten calorimeter of the proposed International Large Detector. Combined with a novel second post-processing network, this approach achieves an accurate simulation of differential distributions including for the first time the shape of the minimum-ionizing-particle peak compared to a full Geant4 simulation for a high-granularity calorimeter with 27k simulated channels. The results are validated by comparing to established architectures. Our results further strengthen the case of using generative networks for fast simulation and demonstrate that physically relevant differential distributions can be described with high accuracy.

scholarly journals Insulin-mediated muscle microvascular perfusion and its phenotypic predictors in humans

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Kaitlin M. Love ◽  
Linda A. Jahn ◽  
Lee M. Hartline ◽  
James T. Patrie ◽  
Eugene J. Barrett ◽  
...  
Keyword(s):  
Body Mass ◽  
Hdl Cholesterol ◽  
Total Body Weight ◽  
Microvascular Perfusion ◽  
Whole Body ◽  
Glucose Disposal ◽  
Phenotypic Traits ◽  
Predictive Information ◽  
Class 1 ◽  
Insulin Responses

AbstractInsulin increases muscle microvascular perfusion and enhances tissue insulin and nutrient delivery. Our aim was to determine phenotypic traits that foretell human muscle microvascular insulin responses. Hyperinsulinemic euglycemic clamps were performed in 97 adult humans who were lean and healthy, had class 1 obesity without comorbidities, or controlled type 1 diabetes without complications. Insulin-mediated whole-body glucose disposal rates (M-value) and insulin-induced changes in muscle microvascular blood volume (ΔMBV) were determined. Univariate and multivariate analyses were conducted to examine bivariate and multivariate relationships between outcomes, ΔMBV and M-value, and predictor variables, body mass index (BMI), total body weight (WT), percent body fat (BF), lean body mass, blood pressure, maximum consumption of oxygen (VO2max), plasma LDL (LDL-C) and HDL cholesterol, triglycerides (TG), and fasting insulin (INS) levels. Among all factors, only M-value (r = 0.23, p = 0.02) and VO2max (r = 0.20, p = 0.047) correlated with ΔMBV. Conversely, INS (r = − 0.48, p ≤ 0.0001), BF (r = − 0.54, p ≤ 0.001), VO2max (r = 0.5, p ≤ 0.001), BMI (r = − 0.40, p < 0.001), WT (r = − 0.33, p = 0.001), LDL-C (r = − 0.26, p = 0.009), TG (r = − 0.25, p = 0.012) correlated with M-value. While both ΔMBV (p = 0.045) and TG (p = 0.03) provided significant predictive information about M-value in the multivariate regression model, only M-value was uniquely predictive of ΔMBV (p = 0.045). Thus, both M-value and VO2max correlated with ΔMBV but only M-value provided unique predictive information about ΔMBV. This suggests that metabolic and microvascular insulin responses are important predictors of one another, but most metabolic insulin resistance predictors do not predict microvascular insulin responses.

Sign in / Sign up

Export Citation Format

Share Document