The understanding of the dynamics fishing vessels is great interest to characterize spatial distribution effort and define sustainable strategies. It also a prerequisite for anticipating changes in fishermen's activity reaction management rules, economic context, or evolution exploited resources. Analyzing trajectories individual offers promising perspectives describe during trips. A hidden Markov model with two behavioral states (steaming fishing) developed infer sequence non‐observed...

1. The utilisation distribution describes the relative probability of use a spatial unit by an animal. It is natural to think it as long-term consequence animal's short-term movement decisions: accumulation small displacements which, over time, gives rise global patterns space use. However, most models either ignore underlying movement, assuming independenceof observed locations, or are based on simplistic Brownian motion rules. 2. We introduce new continuous-time model animal Langevin...

This paper introduces a new algorithm to approximate smoothed additive functionals for partially observed stochastic differential equations. method relies on recent procedure which allows compute such approximations online, i.e. as the observations are received, and with computational complexity growing linearly number of Monte Carlo samples. online smoother cannot be used directly in case equations since transition density latent data is usually unknown. We prove that similar may still...

We consider online computation of expectations additive state functionals under general path probability measures proportional to products unnormalised transition densities. These densities are assumed be intractable but possible estimate, with or without bias. Using pseudo-marginalisation techniques we able extend the particle-based, rapid incremental smoother (PaRIS) algorithm proposed in [J.Olsson and J.Westerborn. Efficient particle-based smoothing hidden Markov models: The PaRIS...

A longitudinal study was conducted within a cattle ranch in Gabon to determine the diminazene aceturate (Berenil) index (DAI) group of Zebu, raised under low tsetse density; this measure providing an assessment trypanosomiasis risk. The objective evaluate pressure thus informing control methods and management. Twenty female adult Zebu were monitored for 24 weeks during dry season. Blood samples collected on weekly basis subjected parasitological haematological analysis (n = 480), using...

Summary The paper proposes a new model for individuals’ movement in ecology. process is defined as solution to stochastic differential equation whose drift the gradient of multimodal potential surface. This offers flexible approach among popular potential-based models To perform parameter inference, widely used Euler method compared with two other pseudolikelihood procedures and Monte Carlo expectation–maximization based on exact simulation diffusions. Performances all methods are assessed...

Understanding fishing vessel dynamics at a fine spatial scale is of great interest for defining appropriate management plans. Different models have been developed to detect activity from Vessel Monitoring System (VMS) data. While mathematical and statistical methods differ, all rely on the idea that speed over ground provides information activity. However, trawling with constant relative water mass may as well prove winning strategy both technical (to ensure sufficient trawl opening)...

This paper proposes a new Sequential Monte Carlo algorithm to perform online estimation in the context of state space models when either transition density latent or conditional likelihood an observation given is intractable. In this setting, obtaining low variance estimators expectations under posterior distributions unobserved states observations challenging task. Following recent theoretical results for pseudo-marginal sequential smoothers, backward importance sampling step introduced...

This article addresses online variational estimation in state-space models. We focus on learning the smoothing distribution, i.e. joint distribution of latent states given observations, using a approach together with Monte Carlo importance sampling. propose an efficient algorithm for computing gradient evidence lower bound (ELBO) context streaming data, where observations arrive sequentially. Our contributions include computationally ELBO estimator, demonstrated performance offline and true...

Mathematical modeling in systems toxicology enables a comprehensive understanding of the effects pharmaceutical substances on cardiac health. However, complexity these models limits their widespread application early drug discovery. In this paper, we introduce novel approach to solving parameterized action potentials by combining meta-learning techniques with Systems Biology-Informed Neural Networks (SBINNs). The proposed method, HyperSBINN, effectively addresses challenge predicting various...

This paper proposes a new model for individuals movement in ecology. The process is defined as solution to stochastic differential equation whose drift the gradient of multimodal potential surface. offers flexible approach among popular based models To perform parameter inference, widely used Euler method compared with two other pseudo-likelihood procedures and Monte Carlo Expectation Maximization on exact simulation diffusions. Performances all methods are assessed simulated data set...

We consider the problem of state estimation in general state-space models using variational inference. For a generic family defined same backward decomposition as actual joint smoothing distribution, we establish for first time that, under mixing assumptions, approximation expectations additive functionals induces an error which grows at most linearly number observations. This guarantee is consistent with known upper bounds distributions standard Monte Carlo methods. Moreover, propose...

We propose an innovative and generic methodology to analyse individual collective behaviour through trajectory data. The work is motivated by the analysis of GPS trajectories fishing vessels collected from regulatory tracking data in context marine biodiversity conservation ecosystem-based fisheries management. build a low-dimensional latent representation using convolutional neural networks as non-linear mapping. This done training conditional variational auto-encoder taking into account...