Visualization of numerical optimization algorithms

Overview

Visualization of numerical optimization algorithms

Numerical optimization is one of the core math foundations in image processing and machine learning. But I remember in the beginning of my Ph.D. years, the math behind always made me frustrated 🙁 🙁 .

During the winter vacation of 2016, I decided to make a change. I revisited some well-known optimization methods (e.g., Gradient Descent, Newton/Quasi-Newton Method, ALM, etc.), and made a series of GIF visualizations to show how these algorithms behave dynamically. Check out this repository and hope it can help you better understand these algorithms.

Gradient Descent Methods.

Fixed step size: Step size=0.5.

Fixed step size: Step size=0.1.

Fixed step size: Step size=1.

Gradient decent with the Nesterov Momentum.

Steepest Descent Method. The step size is determined by using line-search towards the gradient decent direction. The "zigzag" trajectory may cause slow convergence at ill-conditioned regions.

Conjugate Gradient Descent and Coordinate Descent Methods.

Fletcher-Reeves (FR). The FR conjugate gradient method may have very slow convergence rate if the step size is not well controlled.

Polakhe-Ribiere-Polyak (RPR). The PRP method is usually better than FR for ill-conditioned problems. Note that although it is called a "conjugate" method, the update direction (red line) is usually not vertical to the true gradient direction (black line).

Coordinate Descent. The coordinate descent method selects only one coordinate at one time for update. The well-known LibLinear package incorporates this idea to solve the linear SVM. In ill-conditioned regions, this algorithm may also face the "zigzag-step" problem.

Newton Methods.

Basic Newton Method. The black curve is the contour of the 2nd order approximation of the objective function. As the Hessian matrix at the initial point is non-positive, the optimization is not stable at very early steps.

Levenbery-Marquardt (LM) Method. LM method improves the stability of the basic Newton method by adding a small diagonal matrix to the Hessian matrix. This algorithm also can be seen as an integration of the basic Newton method and the gradient descent method.

Damped Newton Method. Damped Newton method can be viewed as a combination of the basic Newton method and the line-search based method. In spite of the fact that the Hessian matrix may be non-positive, the convergance can still be guaranteed.

Broyden Fletcher Goldfarb Shanno (BFGS). The BFGS method is the representative of quasi-Newton methods. It takes the first order gradients to approximate the Hessian matrix. In this figure, the red curve represents the true second-order information, while the black curve represents an approximated one by using BFGS.

Gaussian Newton Least Square Method (GNLS). The Gaussian-Newton least square method is a classical algorithm for solving nonlinear least squares regression problems. The essence of this algorithm is to use the first order Jacobian matrix (black curve) as an approximation of the Hessian matrix (red curve).

Random search algorithm

Genetic Algorithm (GA). GA is a classical algorithm to solve non-convex optimization problems. The key to this algorithm can be summarized as: "breeding", "mutation" and "natural selection". In this figure, the green scatters represent the descendants and the red ones represent the result of natural selection.

Simulated Annealing Algorithm (SAA). SAA is another kind of classical algorithm to solve nonconvex optimization problems. In this figure, the red curve on the right corresponds to the "temperature" and the blue curve corresponds to the objective function value. The objective function value converges with the decrease of the temperature.

Constrained Optimization Method

Gradient Projection Method (GPM). GPM is the most straight-forward way to solve a constrained optimization problem. In each interation, the gradient is projected to the feasible domain to make the current point satisfies the constraints.

Exterior-Point Penalty Method. The exterior-point penalty method is a classical way to solve constrained optimization problems. The key to this algorithm is to penalize the objective function outside the feasible domain so that to convert the original constrained problem into an unconstrained one. Note that the objectve may become ill-conditioned at the boundary of the constraints.

Inner-Point Barrier Method. The Inner-Point Barrier Method is another classical way to solve constrained optimization problems. Different from the exterior-point penalty methods where the objective is penalized outside the feasible region, the inner-point barrier method constructs a barrier function at the boundary of the feasible domain so that to prevent crossing the boundary. Similar to the exterior-point penalty method, the objectve may become ill-conditioned at the boundary of the constraints.

Lagrange Dual Ascent Method. By adding a Lagrangian multiplier, any constrained problem can be equally converted to an unconstrained max-min problem . In the Lagrange Dual Ascent Method, the variable x and the Lagrangian multiplier coefficient are alternately updated. Note that when the background color changes, the Lagrangian multiplier started to be taken into consideration during the updates.

Augmented Lagrangian Method (ALM). ALM is designed based on the Lagrange Dual Ascent Method by adding a penalty function as Augmented Lagrangian multipliers. ALM is more robust at ill-conditioned regions, e.g., at the boundary of constraints.

"keep Calm and Don't Overfit."

Owner
Zhengxia Zou
Postdoc at the University of Michigan. Research interest: computer vision and applications in remote sensing, self-driving, and video games.
Zhengxia Zou
Personal IMDB Graphs with Bokeh

Personal IMDB Graphs with Bokeh Do you like watching movies and also rate all of them in IMDB? Would you like to look at your IMDB stats based on your

2 Dec 15, 2021
Dimensionality reduction in very large datasets using Siamese Networks

ivis Implementation of the ivis algorithm as described in the paper Structure-preserving visualisation of high dimensional single-cell datasets. Ivis

beringresearch 284 Jan 01, 2023
A little logger for machine learning research

Blinker Blinker provides a fast dispatching system that allows any number of interested parties to subscribe to events, or "signals". Signal receivers

Reinforcement Learning Working Group 27 Dec 03, 2022
Bcc2telegraf: An integration that sends ebpf-based bcc histogram metrics to telegraf daemon

bcc2telegraf bcc2telegraf is an integration that sends ebpf-based bcc histogram

Peter Bobrov 2 Feb 17, 2022
Python Data Validation for Humans™.

validators Python data validation for Humans. Python has all kinds of data validation tools, but every one of them seems to require defining a schema

Konsta Vesterinen 670 Jan 09, 2023
HM02: Visualizing Interesting Datasets

HM02: Visualizing Interesting Datasets This is a homework assignment for CSCI 40 class at Claremont McKenna College. Go to the project page to learn m

Qiaoling Chen 11 Oct 26, 2021
Define fortify and autoplot functions to allow ggplot2 to handle some popular R packages.

ggfortify This package offers fortify and autoplot functions to allow automatic ggplot2 to visualize statistical result of popular R packages. Check o

Sinhrks 504 Dec 23, 2022
Small U-Net for vehicle detection

Small U-Net for vehicle detection Vivek Yadav, PhD Overview In this repository , we will go over using U-net for detecting vehicles in a video stream

Vivek Yadav 91 Nov 03, 2022
Python library that makes it easy for data scientists to create charts.

Chartify Chartify is a Python library that makes it easy for data scientists to create charts. Why use Chartify? Consistent input data format: Spend l

Spotify 3.2k Jan 04, 2023
A deceptively simple plotting library for Streamlit

🍅 Plost A deceptively simple plotting library for Streamlit. Because you've been writing plots wrong all this time. Getting started pip install plost

Thiago Teixeira 192 Dec 29, 2022
A simple Monte Carlo simulation using Python and matplotlib library

Monte Carlo python simulation Install linux dependencies sudo apt update sudo apt install build-essential \ software-properties-commo

Samuel Terra 2 Dec 13, 2021
A set of three functions, useful in geographical calculations of different sorts

GreatCircle A set of three functions, useful in geographical calculations of different sorts. Available for PHP, Python, Javascript and Ruby. Live dem

72 Sep 30, 2022
Massively parallel self-organizing maps: accelerate training on multicore CPUs, GPUs, and clusters

Somoclu Somoclu is a massively parallel implementation of self-organizing maps. It exploits multicore CPUs, it is able to rely on MPI for distributing

Peter Wittek 239 Nov 10, 2022
Some useful extensions for Matplotlib.

mplx Some useful extensions for Matplotlib. Contour plots for functions with discontinuities plt.contour mplx.contour(max_jump=1.0) Matplotlib has pro

Nico Schlömer 519 Dec 30, 2022
An open-source tool for visual and modular block programing in python

PyFlow PyFlow is an open-source tool for modular visual programing in python ! Although for now the tool is in Beta and features are coming in bit by

1.1k Jan 06, 2023
Realtime Web Apps and Dashboards for Python and R

H2O Wave Realtime Web Apps and Dashboards for Python and R New! R Language API Build and control Wave dashboards using R! New! Easily integrate AI/ML

H2O.ai 3.4k Jan 06, 2023
a robust room presence solution for home automation with nearly no false negatives

Argos Room Presence This project builds a room presence solution on top of Argos. Using just a cheap raspberry pi zero w (plus an attached pi camera,

Angad Singh 46 Sep 18, 2022
An open-source plotting library for statistical data.

Lets-Plot Lets-Plot is an open-source plotting library for statistical data. It is implemented using the Kotlin programming language. The design of Le

JetBrains 820 Jan 06, 2023
Visualize large time-series data in plotly

plotly_resampler enables visualizing large sequential data by adding resampling functionality to Plotly figures. In this Plotly-Resampler demo over 11

PreDiCT.IDLab 604 Dec 28, 2022
Generate the report for OCULTest.

Sample report generated in this function Usage example from utils.gen_report import generate_report if __name__ == '__main__': # def generate_rep

Philip Guo 1 Mar 10, 2022