Search results for " Marko"
showing 10 items of 201 documents
ASR performance prediction on unseen broadcast programs using convolutional neural networks
2018
In this paper, we address a relatively new task: prediction of ASR performance on unseen broadcast programs. We first propose an heterogenous French corpus dedicated to this task. Two prediction approaches are compared: a state-of-the-art performance prediction based on regression (engineered features) and a new strategy based on convolutional neural networks (learnt features). We particularly focus on the combination of both textual (ASR transcription) and signal inputs. While the joint use of textual and signal features did not work for the regression baseline, the combination of inputs for CNNs leads to the best WER prediction performance. We also show that our CNN prediction remarkably …
Expanding the Active Inference Landscape: More Intrinsic Motivations in the Perception-Action Loop
2018
Active inference is an ambitious theory that treats perception, inference and action selection of autonomous agents under the heading of a single principle. It suggests biologically plausible explanations for many cognitive phenomena, including consciousness. In active inference, action selection is driven by an objective function that evaluates possible future actions with respect to current, inferred beliefs about the world. Active inference at its core is independent from extrinsic rewards, resulting in a high level of robustness across e.g.\ different environments or agent morphologies. In the literature, paradigms that share this independence have been summarised under the notion of in…
On resampling schemes for particle filters with weakly informative observations
2022
We consider particle filters with weakly informative observations (or `potentials') relative to the latent state dynamics. The particular focus of this work is on particle filters to approximate time-discretisations of continuous-time Feynman--Kac path integral models -- a scenario that naturally arises when addressing filtering and smoothing problems in continuous time -- but our findings are indicative about weakly informative settings beyond this context too. We study the performance of different resampling schemes, such as systematic resampling, SSP (Srinivasan sampling process) and stratified resampling, as the time-discretisation becomes finer and also identify their continuous-time l…
Adaptive independent sticky MCMC algorithms
2018
In this work, we introduce a novel class of adaptive Monte Carlo methods, called adaptive independent sticky MCMC algorithms, for efficient sampling from a generic target probability density function (pdf). The new class of algorithms employs adaptive non-parametric proposal densities which become closer and closer to the target as the number of iterations increases. The proposal pdf is built using interpolation procedures based on a set of support points which is constructed iteratively based on previously drawn samples. The algorithm's efficiency is ensured by a test that controls the evolution of the set of support points. This extra stage controls the computational cost and the converge…
Conditional particle filters with diffuse initial distributions
2020
Conditional particle filters (CPFs) are powerful smoothing algorithms for general nonlinear/non-Gaussian hidden Markov models. However, CPFs can be inefficient or difficult to apply with diffuse initial distributions, which are common in statistical applications. We propose a simple but generally applicable auxiliary variable method, which can be used together with the CPF in order to perform efficient inference with diffuse initial distributions. The method only requires simulatable Markov transitions that are reversible with respect to the initial distribution, which can be improper. We focus in particular on random-walk type transitions which are reversible with respect to a uniform init…
Unbiased Inference for Discretely Observed Hidden Markov Model Diffusions
2021
We develop a Bayesian inference method for diffusions observed discretely and with noise, which is free of discretisation bias. Unlike existing unbiased inference methods, our method does not rely on exact simulation techniques. Instead, our method uses standard time-discretised approximations of diffusions, such as the Euler--Maruyama scheme. Our approach is based on particle marginal Metropolis--Hastings, a particle filter, randomised multilevel Monte Carlo, and importance sampling type correction of approximate Markov chain Monte Carlo. The resulting estimator leads to inference without a bias from the time-discretisation as the number of Markov chain iterations increases. We give conver…
Mixture Hidden Markov Models for Sequence Data: The seqHMM Package in R
2019
Sequence analysis is being more and more widely used for the analysis of social sequences and other multivariate categorical time series data. However, it is often complex to describe, visualize, and compare large sequence data, especially when there are multiple parallel sequences per subject. Hidden (latent) Markov models (HMMs) are able to detect underlying latent structures and they can be used in various longitudinal settings: to account for measurement error, to detect unobservable states, or to compress information across several types of observations. Extending to mixture hidden Markov models (MHMMs) allows clustering data into homogeneous subsets, with or without external covariate…
Designing a framework for assisting depression severity assessment from facial image analysis
2015
Depression is one of the most common mental disorders affecting millions of people worldwide. Developing adjunct tools aiding depression assessment is expected to impact overall health outcomes and treatment cost reduction. To this end, platforms designed for automatic and non-invasive depression assessment could help in detecting signs of the disease on a regular basis, without requiring the physical presence of a mental health professional. Despite the different approaches that can be found in the literature, both in terms of methods and algorithms, a fully satisfactory system for the automatic assessment of depression severity has not been presented as yet. This paper describes a propose…
A multi-layer method to study genome-scale positions of nucleosomes
2009
AbstractThe basic unit of eukaryotic chromatin is the nucleosome, consisting of about 150 bp of DNA wrapped around a protein core made of histone proteins. Nucleosomes position is modulated in vivo to regulate fundamental nuclear processes. To measure nucleosome positions on a genomic scale both theoretical and experimental approaches have been recently reported. We have developed a new method, Multi-Layer Model (MLM), for the analysis of nucleosome position data obtained with microarray-based approach. The MLM is a feature extraction method in which the input data is processed by a classifier to distinguish between several kinds of patterns. We applied our method to simulated-synthetic and…
Cartels Uncovered
2018
How many cartels are there? The answer is important in assessing the efficiency of competition policy. We present a Hidden Markov Model that answers the question, taking into account that often we do not know whether a cartel exists in an industry or not. Our model identifies key policy parameters from data generated under different competition policy regimes and may be used with time-series or panel data. We take the model to data from a period of legal cartels - Finnish manufacturing industries 1951 - 1990. Our estimates suggest that by the end of the period, almost all industries were cartelized.