Articles

SDE model calibration through neural posterior estimation

jax
diffrax
sde
probabilistic programming
simulation-based inference
machine learning
quantitative finance
Quantifying uncertainty in inverse problems normally requires a statistical model that can be hard to derive. Simulation-based inference offers an alternative approach where this model is learned from synthetic data. Probabilistic programming libraries facilitate that process.
Jul 7, 2024
Yves Barmaz

SDE model calibration with Diffrax

jax
diffrax
sde
quantitative finance
A discussion of numerical integration and flexible gradient-based calibration of SDE models in the JAX ecosystem, with an example from finance.
Mar 10, 2024
Yves Barmaz

Masked autoregressive flows for stochastic differential equations

probabilistic programming
tensorflow-probability
machine learning
Common methods for solving SDEs are discussed, along with their limitations when it comes to statistical inference. Conditional masked autoregressive flows are presented as a flexible alternative.
Nov 12, 2023
Yves Barmaz

Differentiable and probabilistic programming for risk management and investment decisions

differentiable programming
probabilistic programming
tensorflow-probability
quantitative finance
drug development
risk management
Some applications of differentiable programming in data science beyond the usual machine learning algorithms, with an example in drug development valuation.
Nov 11, 2023
Yves Barmaz

Neural likelihood estimation and simulation-based inference

simulation-based inference
machine learning
tensorflow-probability
A discussion of simulation-based inference fueled by machine learning algorithms, with an example of neural likelihood estimation in TensorFlow Probability.
Jul 14, 2023
Yves Barmaz

Rabbits and velociraptors: scientific modeling on steroids

scientific computing
simulation-based inference
jax
tensorflow-probability
A discussion of differentiable programming and scientific computing in the JAX ecosystem, with an example of simulation-based inference.
Jun 24, 2023
Yves Barmaz

Cleaning Strava data with Kalman smoothers

mountain biking
signal processing
tensorflow-probability
satellite navigation
A tale of mountain biking, paranormal activity and signal processing.
Nov 5, 2022
Yves Barmaz

 

Bayesian Poisson factorization

bayesian modeling
variational inference
tensorflow-probability
machine learning
A probabilistic recommender system implemented with TensorFlow Probability layers.
Feb 23, 2022
Yves Barmaz

Hamiltonian mechanics with JAX

jax
autograd
differential geometry
classical mechanics
An application of automatic differentiation with JAX in classical mechanics.
Dec 4, 2021
Yves Barmaz

Gaussian random walks

random walks
bayesian modeling
time series
tensorflow-probability
A construction of a TensorFlow-Probability distribution implementing Gaussian random walks, with an application to stochastic volatility modeling.
Jun 5, 2021
Yves Barmaz

Thompson sampling for contextual bandits

contextual bandits
reinforcement learning
bayesian modeling
variational inference
probabilistic machine learning
tensorflow-probability
An introduction to Thompson sampling and how to implement it with probabilistic machine learning to tackle contextual bandits.
Feb 9, 2021
Yves Barmaz

Variational inference with TensorFlow-Probability

bayesian modeling
variational inference
tensorflow-probability
An overview of the variational inference APIs available in TensorFlow-Probability.
Feb 1, 2021
Yves Barmaz

Anomaly detection in protocol deviations

bayesian modeling
clinical quality
When math helps you define too many and too few protocol deviations.
Jan 13, 2021
Yves Barmaz

Probabilistic assessment of safety underreporting

bayesian modeling
clinical quality
Adverse event underreporting is a problem that sometimes affects clinical trials. Probabilistic programming is the modern way to address this issue.
Jan 8, 2021
Yves Barmaz
No matching items