An implementation of the methods presented in Causal-BALD: Deep Bayesian Active Learning of Outcomes to Infer Treatment-Effects from Observational Data.

Related tags

Deep Learning causal-bald
Overview

causal-bald

| Abstract | Installation | Example | Citation | Reproducing Results DUE

An implementation of the methods presented in Causal-BALD: Deep Bayesian Active Learning of Outcomes to Infer Treatment-Effects from Observational Data.

Evolution of CATE function with Causal BALD acquisition strategy

Abstract

Estimating personalized treatment effects from high-dimensional observational data is essential in situations where experimental designs are infeasible, unethical or expensive. Existing approaches rely on fitting deep models on outcomes observed for treated and control populations, but when measuring the outcome for an individual is costly (e.g. biopsy) a sample efficient strategy for acquiring outcomes is required. Deep Bayesian active learning provides a framework for efficient data acquisition by selecting points with high uncertainty. However, naive application of existing methods selects training data that is biased toward regions where the treatment effect cannot be identified because there is non-overlapping support between the treated and control populations. To maximize sample efficiency for learning personalized treatment effects, we introduce new acquisition functions grounded in information theory that bias data acquisition towards regions where overlap is satisfied, by combining insights from deep Bayesian active learning and causal inference. We demonstrate the performance of the proposed acquisition strategies on synthetic and semi-synthetic datasets IHDP and CMNIST and their extensions which aim to simulate common dataset biases and pathologies.

Installation

$ git clone [email protected]:[anon]/causal-bald.git
$ cd causal-bald
$ conda env create -f environment.yml
$ conda activate causal-bald

[Optional] For developer mode

$ pip install -e .

Example

Active learning loop

First run using random acquisition:

causal-bald \
    active-learning \
        --job-dir experiments/ \
        --num-trials 5 \
        --step-size 10 \
        --warm-start-size 100 \
        --max-acquisitions 38 \
        --acquisition-function random \
        --temperature 0.25 \
        --gpu-per-trial 0.2 \
    ihdp \
        --root assets/ \
    deep-kernel-gp

Now run using $\mu\rho\textrm{-BALD}$ acquisition.

causal-bald \
    active-learning \
        --job-dir experiments/ \
        --num-trials 5 \
        --step-size 10 \
        --warm-start-size 100 \
        --max-acquisitions 38 \
        --acquisition-function mu-rho \
        --temperature 0.25 \
        --gpu-per-trial 0.2 \
    ihdp \
        --root assets/ \
    deep-kernel-gp

Evaluation

Evaluate PEHE at each acquisition step

causal-bald \
    evaluate \
        --experiment-dir experiments/active_learning/ss-10_ws-100_ma-38_af-random_temp-0.25/ihdp/deep_kernel_gp/kernel-Matern32_ip-100-dh-200_do-1_dp-3_ns--1.0_dr-0.1_sn-0.95_lr-0.001_bs-100_ep-500/ \
        --output-dir experiments/due/ihdp \
    pehe
causal-bald \
    evaluate \
        --experiment-dir experiments/active_learning/ss-10_ws-100_ma-38_af-mu-rho_temp-0.25/ihdp/deep_kernel_gp/kernel-Matern32_ip-100-dh-200_do-1_dp-3_ns--1.0_dr-0.1_sn-0.95_lr-0.001_bs-100_ep-500/ \
        --output-dir experiments/due/ihdp \
    pehe

Plot results

causal-bald \
    evaluate \
        --experiment-dir experiments/due/ihdp \
    plot-convergence \
        -m mu-rho \
        -m random

Plotting convergence of acquisitions. Comparing random and mu-rho for example code

Citation

If you find this code helpful for your work, please cite our paper Paper as

@article{jesson2021causal,
  title={Causal-BALD: Deep Bayesian Active Learning of Outcomes to Infer Treatment-Effects from Observational Data},
  author={Jesson, Andrew and Tigas, Panagiotis and van Amersfoort, Joost and Kirsch, Andreas and Shalit, Uri and Gal, Yarin},
  journal={Advances in Neural Information Processing Systems},
  volume={35},
  year={2021}
}

Reprodcuing Results Due

IHDP

$\mu\rho$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 200 --step-size 10 --warm-start-size 100 --max-acquisitions 38 --acquisition-function mu-rho --temperature 0.25 --gpu-per-trial 0.2 ihdp --root assets/ deep-kernel-gp
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-100_ma-38_af-mu-rho_temp-0.25/ihdp/deep_kernel_gp/kernel-Matern32_ip-100-dh-200_do-1_dp-3_ns--1.0_dr-0.1_sn-0.95_lr-0.001_bs-100_ep-500/ --output-dir experiments/due/ihdp pehe

$\mu$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 200 --step-size 10 --warm-start-size 100 --max-acquisitions 38 --acquisition-function mu --temperature 0.25 --gpu-per-trial 0.2 ihdp --root assets/ deep-kernel-gp
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-100_ma-38_af-mu_temp-0.25/ihdp/deep_kernel_gp/kernel-Matern32_ip-100-dh-200_do-1_dp-3_ns--1.0_dr-0.1_sn-0.95_lr-0.001_bs-100_ep-500/ --output-dir experiments/due/ihdp pehe

$\mu\pi$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 200 --step-size 10 --warm-start-size 100 --max-acquisitions 38 --acquisition-function mu-pi --temperature 0.25 --gpu-per-trial 0.2 ihdp --root assets/ deep-kernel-gp
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-100_ma-38_af-mu-pi_temp-0.25/ihdp/deep_kernel_gp/kernel-Matern32_ip-100-dh-200_do-1_dp-3_ns--1.0_dr-0.1_sn-0.95_lr-0.001_bs-100_ep-500/ --output-dir experiments/due/ihdp pehe

$\rho$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 200 --step-size 10 --warm-start-size 100 --max-acquisitions 38 --acquisition-function rho --temperature 0.25 --gpu-per-trial 0.2 ihdp --root assets/ deep-kernel-gp
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-100_ma-38_af-rho_temp-0.25/ihdp/deep_kernel_gp/kernel-Matern32_ip-100-dh-200_do-1_dp-3_ns--1.0_dr-0.1_sn-0.95_lr-0.001_bs-100_ep-500/ --output-dir experiments/due/ihdp pehe

$\pi$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 200 --step-size 10 --warm-start-size 100 --max-acquisitions 38 --acquisition-function pi --temperature 0.25 --gpu-per-trial 0.2 ihdp --root assets/ deep-kernel-gp
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-100_ma-38_af-pi_temp-0.25/ihdp/deep_kernel_gp/kernel-Matern32_ip-100-dh-200_do-1_dp-3_ns--1.0_dr-0.1_sn-0.95_lr-0.001_bs-100_ep-500/ --output-dir experiments/due/ihdp pehe

$\tau$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 200 --step-size 10 --warm-start-size 100 --max-acquisitions 38 --acquisition-function tau --temperature 0.25 --gpu-per-trial 0.2 ihdp --root assets/ deep-kernel-gp
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-100_ma-38_af-tau_temp-0.25/ihdp/deep_kernel_gp/kernel-Matern32_ip-100-dh-200_do-1_dp-3_ns--1.0_dr-0.1_sn-0.95_lr-0.001_bs-100_ep-500/ --output-dir experiments/due/ihdp pehe

Random

causal-bald active-learning --job-dir experiments/ --num-trials 200 --step-size 10 --warm-start-size 100 --max-acquisitions 38 --acquisition-function random --temperature 0.25 --gpu-per-trial 0.2 ihdp --root assets/ deep-kernel-gp
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-100_ma-38_af-random_temp-0.25/ihdp/deep_kernel_gp/kernel-Matern32_ip-100-dh-200_do-1_dp-3_ns--1.0_dr-0.1_sn-0.95_lr-0.001_bs-100_ep-500/ --output-dir experiments/due/ihdp pehe

Sundin

causal-bald active-learning --job-dir experiments/ --num-trials 200 --step-size 10 --warm-start-size 100 --max-acquisitions 38 --acquisition-function sundin --temperature 1.0 --gpu-per-trial 0.2 ihdp --root assets/ deep-kernel-gp
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-100_ma-38_af-sundin_temp-1.0/ihdp/deep_kernel_gp/kernel-Matern32_ip-100-dh-200_do-1_dp-3_ns--1.0_dr-0.1_sn-0.95_lr-0.001_bs-100_ep-500/ --output-dir experiments/due/ihdp pehe

Plot Results

causal-bald \
    evaluate \
        --experiment-dir experiments/due/ihdp \
    plot-convergence \
        -m mu-rho \
        -m mu \
        -m mu-pi \
        -m rho \ \
        -m pi
        -m tau \
        -m random \
        -m sundin

Synthetic

Synthetic dataset

Synthetic: $\mu\rho$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 40 --step-size 10 --warm-start-size 10 --max-acquisitions 31 --acquisition-function mu-rho --temperature 0.25 --gpu-per-trial 0.2 synthetic deep-kernel-gp --kernel RBF --dim-hidden 100 --num-inducing-points 20 --negative-slope 0.0 --batch-size 200 --dropout-rate 0.2
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-10_ma-31_af-mu-rho_temp-0.25/synthetic/deep_kernel_gp/kernel-RBF_ip-20-dh-100_do-1_dp-3_ns-0.0_dr-0.2_sn-0.95_lr-0.001_bs-200_ep-500/ --output-dir experiments/due/synthetic pehe

Synthetic: $\mu$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 40 --step-size 10 --warm-start-size 10 --max-acquisitions 31 --acquisition-function mu --temperature 0.25 --gpu-per-trial 0.2 synthetic deep-kernel-gp --kernel RBF --dim-hidden 100 --num-inducing-points 20 --negative-slope 0.0 --batch-size 200 --dropout-rate 0.2
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-10_ma-31_af-mu_temp-0.25/synthetic/deep_kernel_gp/kernel-RBF_ip-20-dh-100_do-1_dp-3_ns-0.0_dr-0.2_sn-0.95_lr-0.001_bs-200_ep-500/ --output-dir experiments/due/ihdp pehe

Synthetic: $\mu\pi$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 40 --step-size 10 --warm-start-size 10 --max-acquisitions 31 --acquisition-function mu-pi --temperature 0.25 --gpu-per-trial 0.2 synthetic deep-kernel-gp --kernel RBF --dim-hidden 100 --num-inducing-points 20 --negative-slope 0.0 --batch-size 200 --dropout-rate 0.2
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-10_ma-31_af-mu-pi_temp-0.25/synthetic/deep_kernel_gp/kernel-RBF_ip-20-dh-100_do-1_dp-3_ns-0.0_dr-0.2_sn-0.95_lr-0.001_bs-200_ep-500/ --output-dir experiments/due/synthetic pehe

Synthetic: $\rho$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 40 --step-size 10 --warm-start-size 10 --max-acquisitions 31 --acquisition-function rho --temperature 0.25 --gpu-per-trial 0.2 synthetic deep-kernel-gp --kernel RBF --dim-hidden 100 --num-inducing-points 20 --negative-slope 0.0 --batch-size 200 --dropout-rate 0.2
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-10_ma-31_af-rho_temp-0.25/synthetic/deep_kernel_gp/kernel-RBF_ip-20-dh-100_do-1_dp-3_ns-0.0_dr-0.2_sn-0.95_lr-0.001_bs-200_ep-500/ --output-dir experiments/due/synthetic pehe

Synthetic: $\pi$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 40 --step-size 10 --warm-start-size 10 --max-acquisitions 31 --acquisition-function pi --temperature 0.25 --gpu-per-trial 0.2 synthetic deep-kernel-gp --kernel RBF --dim-hidden 100 --num-inducing-points 20 --negative-slope 0.0 --batch-size 200 --dropout-rate 0.2
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-10_ma-31_af-pi_temp-0.25/synthetic/deep_kernel_gp/kernel-RBF_ip-20-dh-100_do-1_dp-3_ns-0.0_dr-0.2_sn-0.95_lr-0.001_bs-200_ep-500/ --output-dir experiments/due/synthetic pehe

Synthetic: $\tau$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 40 --step-size 10 --warm-start-size 10 --max-acquisitions 31 --acquisition-function tau --temperature 0.25 --gpu-per-trial 0.2 synthetic deep-kernel-gp --kernel RBF --dim-hidden 100 --num-inducing-points 20 --negative-slope 0.0 --batch-size 200 --dropout-rate 0.2
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-10_ma-31_af-tau_temp-0.25/synthetic/deep_kernel_gp/kernel-RBF_ip-20-dh-100_do-1_dp-3_ns-0.0_dr-0.2_sn-0.95_lr-0.001_bs-200_ep-500/ --output-dir experiments/due/synthetic pehe

Synthetic: Random

causal-bald active-learning --job-dir experiments/ --num-trials 40 --step-size 10 --warm-start-size 10 --max-acquisitions 31 --acquisition-function random --temperature 0.25 --gpu-per-trial 0.2 synthetic deep-kernel-gp --kernel RBF --dim-hidden 100 --num-inducing-points 20 --negative-slope 0.0 --batch-size 200 --dropout-rate 0.2
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-10_ma-31_af-random_temp-0.25/synthetic/deep_kernel_gp/kernel-RBF_ip-20-dh-100_do-1_dp-3_ns-0.0_dr-0.2_sn-0.95_lr-0.001_bs-200_ep-500/ --output-dir experiments/due/synthetic pehe

Synthetic: Sundin

causal-bald active-learning --job-dir experiments/ --num-trials 40 --step-size 10 --warm-start-size 10 --max-acquisitions 31 --acquisition-function sundin --temperature 1.0 --gpu-per-trial 0.2 synthetic deep-kernel-gp --kernel RBF --dim-hidden 100 --num-inducing-points 20 --negative-slope 0.0 --batch-size 200 --dropout-rate 0.2
causal-bald evaluate --experiment-dir experiments/active_learning/ss-10_ws-10_ma-31_af-sundin_temp-1.0/synthetic/deep_kernel_gp/kernel-RBF_ip-20-dh-100_do-1_dp-3_ns-0.0_dr-0.2_sn-0.95_lr-0.001_bs-200_ep-500/ --output-dir experiments/due/synthetic pehe

Synthetic: Plot Results

causal-bald \
    evaluate \
        --experiment-dir experiments/due/synthetic \
    plot-convergence \
        -m mu-rho \
        -m mu \
        -m mu-pi \
        -m rho \ \
        -m pi
        -m tau \
        -m random \
        -m sundin

CMNIST

CMNIST dataset

CMNIST: $\mu\rho$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 10 --step-size 50 --warm-start-size 250 --max-acquisitions 56 --acquisition-function mu-rho --temperature 0.25 --gpu-per-trial 0.5 cmnist --root assets/ deep-kernel-gp --kernel RBF --depth 2 --dropout-rate 0.05 --spectral-norm 3.0 --batch-size 64
causal-bald evaluate --experiment-dir experiments/active_learning/ss-50_ws-250_ma-56_af-mu-rho_temp-0.25/cmnist/deep_kernel_gp/kernel-RBF_ip-100-dh-200_do-1_dp-2_ns--1.0_dr-0.05_sn-3.0_lr-0.001_bs-64_ep-500/ --output-dir experiments/due/cmnist pehe

CMNIST: $\mu$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 10 --step-size 50 --warm-start-size 250 --max-acquisitions 56 --acquisition-function mu --temperature 0.25 --gpu-per-trial 0.5 cmnist --root assets/ deep-kernel-gp --kernel RBF --depth 2 --dropout-rate 0.05 --spectral-norm 3.0 --batch-size 64
causal-bald evaluate --experiment-dir experiments/active_learning/ss-50_ws-250_ma-56_af-mu_temp-0.25/cmnist/deep_kernel_gp/kernel-RBF_ip-100-dh-200_do-1_dp-2_ns--1.0_dr-0.05_sn-3.0_lr-0.001_bs-64_ep-500/ --output-dir experiments/due/ihdp pehe

CMNIST: $\mu\pi$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 10 --step-size 50 --warm-start-size 250 --max-acquisitions 56 --acquisition-function mu-pi --temperature 0.25 --gpu-per-trial 0.5 cmnist --root assets/ deep-kernel-gp --kernel RBF --depth 2 --dropout-rate 0.05 --spectral-norm 3.0 --batch-size 64
causal-bald evaluate --experiment-dir experiments/active_learning/ss-50_ws-250_ma-56_af-mu-pi_temp-0.25/cmnist/deep_kernel_gp/kernel-RBF_ip-100-dh-200_do-1_dp-2_ns--1.0_dr-0.05_sn-3.0_lr-0.001_bs-64_ep-500/ --output-dir experiments/due/cmnist pehe

CMNIST: $\rho$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 10 --step-size 50 --warm-start-size 250 --max-acquisitions 56 --acquisition-function rho --temperature 0.25 --gpu-per-trial 0.5 cmnist --root assets/ deep-kernel-gp --kernel RBF --depth 2 --dropout-rate 0.05 --spectral-norm 3.0 --batch-size 64
causal-bald evaluate --experiment-dir experiments/active_learning/ss-50_ws-250_ma-56_af-rho_temp-0.25/cmnist/deep_kernel_gp/kernel-RBF_ip-100-dh-200_do-1_dp-2_ns--1.0_dr-0.05_sn-3.0_lr-0.001_bs-64_ep-500/ --output-dir experiments/due/cmnist pehe

CMNIST: $\pi$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 10 --step-size 50 --warm-start-size 250 --max-acquisitions 56 --acquisition-function pi --temperature 0.25 --gpu-per-trial 0.5 cmnist --root assets/ deep-kernel-gp --kernel RBF --depth 2 --dropout-rate 0.05 --spectral-norm 3.0 --batch-size 64
causal-bald evaluate --experiment-dir experiments/active_learning/ss-50_ws-250_ma-56_af-pi_temp-0.25/cmnist/deep_kernel_gp/kernel-RBF_ip-100-dh-200_do-1_dp-2_ns--1.0_dr-0.05_sn-3.0_lr-0.001_bs-64_ep-500/ --output-dir experiments/due/cmnist pehe

CMNIST: $\tau$-BALD

causal-bald active-learning --job-dir experiments/ --num-trials 10 --step-size 50 --warm-start-size 250 --max-acquisitions 56 --acquisition-function tau --temperature 0.25 --gpu-per-trial 0.5 cmnist --root assets/ deep-kernel-gp --kernel RBF --depth 2 --dropout-rate 0.05 --spectral-norm 3.0 --batch-size 64
causal-bald evaluate --experiment-dir experiments/active_learning/ss-50_ws-250_ma-56_af-tau_temp-0.25/cmnist/deep_kernel_gp/kernel-RBF_ip-100-dh-200_do-1_dp-2_ns--1.0_dr-0.05_sn-3.0_lr-0.001_bs-64_ep-500/ --output-dir experiments/due/cmnist pehe

CMNIST: Random

causal-bald active-learning --job-dir experiments/ --num-trials 10 --step-size 50 --warm-start-size 250 --max-acquisitions 56 --acquisition-function random --temperature 0.25 --gpu-per-trial 0.5 cmnist --root assets/ deep-kernel-gp --kernel RBF --depth 2 --dropout-rate 0.05 --spectral-norm 3.0 --batch-size 64
causal-bald evaluate --experiment-dir experiments/active_learning/ss-50_ws-250_ma-56_af-random_temp-0.25/cmnist/deep_kernel_gp/kernel-RBF_ip-100-dh-200_do-1_dp-2_ns--1.0_dr-0.05_sn-3.0_lr-0.001_bs-64_ep-500/ --output-dir experiments/due/cmnist pehe

CMNIST: Sundin

causal-bald active-learning --job-dir experiments/ --num-trials 10 --step-size 50 --warm-start-size 250 --max-acquisitions 56 --acquisition-function sundin --temperature 1.0 --gpu-per-trial 0.5 cmnist --root assets/ deep-kernel-gp --kernel RBF --depth 2 --dropout-rate 0.05 --spectral-norm 3.0 --batch-size 64
causal-bald evaluate --experiment-dir experiments/active_learning/ss-50_ws-250_ma-56_af-sundin_temp-1.0/cmnist/deep_kernel_gp/kernel-RBF_ip-100-dh-200_do-1_dp-2_ns--1.0_dr-0.05_sn-3.0_lr-0.001_bs-64_ep-500/ --output-dir experiments/due/cmnist pehe

CMNIST: Plot Results

causal-bald \
    evaluate \
        --experiment-dir experiments/due/cmnist \
    plot-convergence \
        -m mu-rho \
        -m mu \
        -m mu-pi \
        -m rho \ \
        -m pi
        -m tau \
        -m random \
        -m sundin
You might also like...
JudeasRx - graphical app for doing personalized causal medicine using the methods invented by Judea Pearl et al.
JudeasRx - graphical app for doing personalized causal medicine using the methods invented by Judea Pearl et al.

JudeasRX Instructions Read the references given in the Theory and Notation section below Fire up the Jupyter Notebook judeas-rx.ipynb The notebook dra

Robert R. Tucci 19 Nov 7, 2022
A PyTorch-based open-source framework that provides methods for improving the weakly annotated data and allows researchers to efficiently develop and compare their own methods.
A PyTorch-based open-source framework that provides methods for improving the weakly annotated data and allows researchers to efficiently develop and compare their own methods.

Knodle (Knowledge-supervised Deep Learning Framework) - a new framework for weak supervision with neural networks. It provides a modularization for se

null 93 Nov 6, 2022
Implementation of temporal pooling methods studied in [ICIP'20] A Comparative Evaluation Of Temporal Pooling Methods For Blind Video Quality Assessment

Implementation of temporal pooling methods studied in [ICIP'20] A Comparative Evaluation Of Temporal Pooling Methods For Blind Video Quality Assessment

Zhengzhong Tu 5 Sep 16, 2022
pyhsmm - library for approximate unsupervised inference in Bayesian Hidden Markov Models (HMMs) and explicit-duration Hidden semi-Markov Models (HSMMs), focusing on the Bayesian Nonparametric extensions, the HDP-HMM and HDP-HSMM, mostly with weak-limit approximations.
pyhsmm - library for approximate unsupervised inference in Bayesian Hidden Markov Models (HMMs) and explicit-duration Hidden semi-Markov Models (HSMMs), focusing on the Bayesian Nonparametric extensions, the HDP-HMM and HDP-HSMM, mostly with weak-limit approximations.

Bayesian inference in HSMMs and HMMs This is a Python library for approximate unsupervised inference in Bayesian Hidden Markov Models (HMMs) and expli

Matthew Johnson 527 Dec 4, 2022
Hierarchical-Bayesian-Defense - Towards Adversarial Robustness of Bayesian Neural Network through Hierarchical Variational Inference (Openreview)
Hierarchical-Bayesian-Defense - Towards Adversarial Robustness of Bayesian Neural Network through Hierarchical Variational Inference (Openreview)

Towards Adversarial Robustness of Bayesian Neural Network through Hierarchical V

LBK 20 Dec 2, 2022
Deep Learning: Architectures & Methods Project: Deep Learning for Audio Super-Resolution
Deep Learning: Architectures & Methods Project: Deep Learning for Audio Super-Resolution

Deep Learning: Architectures & Methods Project: Deep Learning for Audio Super-Resolution Figure: Example visualization of the method and baseline as a

Oliver Hahn 16 Dec 23, 2022
Deep Learning Models for Causal Inference

Extensive tutorials for learning how to build deep learning models for causal inference using selection on observables in Tensorflow 2.

Bernard J Koch 151 Dec 31, 2022
Official implementation of GraphMask as presented in our paper Interpreting Graph Neural Networks for NLP With Differentiable Edge Masking.

GraphMask This repository contains an implementation of GraphMask, the interpretability technique for graph neural networks presented in our ICLR 2021

Michael Schlichtkrull 29 Sep 2, 2022
Official implementation of the network presented in the paper
Official implementation of the network presented in the paper "M4Depth: A motion-based approach for monocular depth estimation on video sequences"

M4Depth This is the reference TensorFlow implementation for training and testing depth estimation models using the method described in M4Depth: A moti

Michaël Fonder 76 Jan 3, 2023
Comments
  • added errorbars debug plot

    added errorbars debug plot

    Usage

    causal-bald evaluate --experiment-dir /scratch-ssd/andson/experiments/causal-bald/training/hcmnist/deep_kernel_gp/kernel-Matern32_ip-50-dh-200_do-1_dp-3_ns-0.1_dr-0.05_sn-3.0_lr-0.0002_bs-64_ep-1000/ plot-errorbars
    opened by anndvision 0
  • Hcmnist

    Hcmnist

    Runs with:

    PYTHONPATH=../../ python main.py active-learning --job-dir hcmnist --num-trials 40 --step-size 10 --warm-start-size 10 --max-acquisitions 31 --acquisition-function random --temperature 1.0 --gpu-per-trial 1 hcmnist --root ../../data deep-kernel-gp --kernel RBF --dim-hidden 100 --num-inducing-points 20 --negative-slope 0.0 --batch-size 200 --dropout-rate 0.2
    opened by y0ast 0
Owner
Andrew Jesson
PhD in Machine Learning at University of Oxford @OATML
Andrew Jesson
GitHub
CausalNLP is a practical toolkit for causal inference with text as treatment, outcome, or "controlled-for" variable.

CausalNLP CausalNLP is a practical toolkit for causal inference with text as treatment, outcome, or "controlled-for" variable. Install pip install -U

Arun S. Maiya 95 Jan 3, 2023
This repository contains numerical implementation for the paper Intertemporal Pricing under Reference Effects: Integrating Reference Effects and Consumer Heterogeneity.

This repository contains numerical implementation for the paper Intertemporal Pricing under Reference Effects: Integrating Reference Effects and Consumer Heterogeneity.

Hansheng Jiang 6 Nov 18, 2022
Causal-Adversarial-Instruments - PyTorch Implementation for Developing Library of Investigating Adversarial Examples on A Causal View by Instruments

Causal-Adversarial-Instruments This is a PyTorch Implementation code for develop

LBK 26 Dec 28, 2022
Bald-to-Hairy Translation Using CycleGAN

GANiry: Bald-to-Hairy Translation Using CycleGAN Official PyTorch implementation of GANiry. GANiry: Bald-to-Hairy Translation Using CycleGAN, Fidan Sa

Fidan Samet 10 Oct 27, 2022
Python Library for learning (Structure and Parameter) and inference (Statistical and Causal) in Bayesian Networks.

pgmpy pgmpy is a python library for working with Probabilistic Graphical Models. Documentation and list of algorithms supported is at our official sit

pgmpy 2.2k Jan 3, 2023
From a body shape, infer the anatomic skeleton.

OSSO: Obtaining Skeletal Shape from Outside (CVPR 2022) This repository contains the official implementation of the skeleton inference from: OSSO: Obt

Marilyn Keller 166 Dec 28, 2022
An Image compression simulator that uses Source Extractor and Monte Carlo methods to examine the post compressive effects different compression algorithms have.

ImageCompressionSimulation An Image compression simulator that uses Source Extractor and Monte Carlo methods to examine the post compressive effects o

James Park 1 Dec 11, 2021
Python package facilitating the use of Bayesian Deep Learning methods with Variational Inference for PyTorch

PyVarInf PyVarInf provides facilities to easily train your PyTorch neural network models using variational inference. Bayesian Deep Learning with Vari

null 342 Dec 2, 2022
This repository contains the code used for Predicting Patient Outcomes with Graph Representation Learning (https://arxiv.org/abs/2101.03940).

Predicting Patient Outcomes with Graph Representation Learning This repository contains the code used for Predicting Patient Outcomes with Graph Repre

Emma Rocheteau 76 Dec 22, 2022
Bayesian-Torch is a library of neural network layers and utilities extending the core of PyTorch to enable the user to perform stochastic variational inference in Bayesian deep neural networks

Bayesian-Torch is a library of neural network layers and utilities extending the core of PyTorch to enable the user to perform stochastic variational inference in Bayesian deep neural networks. Bayesian-Torch is designed to be flexible and seamless in extending a deterministic deep neural network architecture to corresponding Bayesian form by simply replacing the deterministic layers with Bayesian layers.

Intel Labs 210 Jan 4, 2023