arXiv subjects

Panagiotis Angelikopoulos

Publications and source records attributed to Panagiotis Angelikopoulos.

5 recordsLinked to original sources

arXiv2020

Efficient hyperparameter optimization by way of PAC-Bayes bound minimization

Identifying optimal values for a high-dimensional set of hyperparameters is a problem that has received growing attention given its importance to large-scale machine learning applications such as neural architecture search. Recently developed optimization methods can be used to select thousands or even millions of hyperparameters. Such methods often yield overfit models, however, leading to poor performance on unseen data. We argue that this overfitting results from using the standard hyperparameter optimization objective function. Here we present an alternative objective that is equivalent to a Probably Approximately Correct-Bayes (PAC-Bayes) bound on the expected out-of-sample error. We then devise an efficient gradient-based algorithm to minimize this objective; the proposed method has asymptotic space and time complexity equal to or better than other gradient-based hyperparameter optimization methods. We show that this new method significantly reduces out-of-sample error when applied to hyperparameter optimization problems known to be prone to overfitting.

stat.ML↗

arXiv2019

Personalized Radiotherapy Design for Glioblastoma: Integrating Mathematical Tumor Models, Multimodal Scans and Bayesian Inference

Glioblastoma is a highly invasive brain tumor, whose cells infiltrate surrounding normal brain tissue beyond the lesion outlines visible in the current medical scans. These infiltrative cells are treated mainly by radiotherapy. Existing radiotherapy plans for brain tumors derive from population studies and scarcely account for patient-specific conditions. Here we provide a Bayesian machine learning framework for the rational design of improved, personalized radiotherapy plans using mathematical modeling and patient multimodal medical scans. Our method, for the first time, integrates complementary information from high resolution MRI scans and highly specific FET-PET metabolic maps to infer tumor cell density in glioblastoma patients. The Bayesian framework quantifies imaging and modeling uncertainties and predicts patient-specific tumor cell density with confidence intervals. The proposed methodology relies only on data acquired at a single time point and thus is applicable to standard clinical settings. An initial clinical population study shows that the radiotherapy plans generated from the inferred tumor cell infiltration maps spare more healthy tissue thereby reducing radiation toxicity while yielding comparable accuracy with standard radiotherapy protocols. Moreover, the inferred regions of high tumor cell densities coincide with the tumor radioresistant areas, providing guidance for personalized dose-escalation. The proposed integration of multimodal scans and mathematical modeling provides a robust, non-invasive tool to assist personalized radiotherapy design.

cs.CE↗

arXiv2017

Experimental data over quantum mechanics simulations for inferring the repulsive exponent of the Lennard-Jones potential in Molecular Dynamics

The Lennard-Jones (LJ) potential is a cornerstone of Molecular Dynamics (MD) simulations and among the most widely used computational kernels in science. The potential models atomistic attraction and repulsion with century old prescribed parameters ($q=6, \; p=12$, respectively), originally related by a factor of two for simplicity of calculations. We re-examine the value of the repulsion exponent through data driven uncertainty quantification. We perform Hierarchical Bayesian inference on MD simulations of argon using experimental data of the radial distribution function (RDF) for a range of thermodynamic conditions, as well as dimer interaction energies from quantum mechanics simulations. The experimental data suggest a repulsion exponent ($p \approx 6.5$), in contrast to the quantum simulations data that support values closer to the original ($p=12$) exponent. Most notably, we find that predictions of RDF, diffusion coefficient and density of argon are more accurate and robust in producing the correct argon phase around its triple point, when using the values inferred from experimental data over those from quantum mechanics simulations. The present results suggest the need for data driven recalibration of the LJ potential across MD simulations.

physics.chem-ph↗

arXiv2016

Hierarchical Stochastic Model in Bayesian Inference: Theoretical Implications and Efficient Approximation

We classify two types of Hierarchical Bayesian Model found in the literature as Hierarchical Prior Model (HPM) and Hierarchical Stochastic Model (HSM). Then, we focus on studying the theoretical implications of the HSM. Using examples of polynomial functions, we show that the HSM is capable of separating different types of uncertainties in a system and quantifying uncertainty of reduced order models under the Bayesian model class selection framework. To tackle the huge computational cost for analyzing HSM, we propose an efficient approximation scheme based on Importance Sampling and Empirical Interpolation Method. We illustrate our method using two examples - a Molecular Dynamics simulation for Krypton and a pharmacokinetic/pharmacodynamic model for cancer drug.

stat.AP↗

arXiv2016

Langevin Diffusion for Population Based Sampling with an Application in Bayesian Inference for Pharmacodynamics

We propose an algorithm for the efficient and robust sampling of the posterior probability distribution in Bayesian inference problems. The algorithm combines the local search capabilities of the Manifold Metropolis Adjusted Langevin transition kernels with the advantages of global exploration by a population based sampling algorithm, the Transitional Markov Chain Monte Carlo (TMCMC). The Langevin diffusion process is determined by either the Hessian or the Fisher Information of the target distribution with appropriate modifications for non positive definiteness. The present methods is shown to be superior over other population based algorithms, in sampling probability distributions for which gradients are available and is shown to handle otherwise unidentifiable models. We demonstrate the capabilities and advantages of the method in computing the posterior distribution of the parameters in a Pharmacodynamics model, for glioma growth and its drug induced inhibition, using clinical data.

stat.CO↗