MiraiML: asynchronous, autonomous and continuous Machine Learning in Python

Arthur Paulino

Last update: Jul 27, 2022

Overview

MiraiML

Mirai: future in japanese.

MiraiML is an asynchronous engine for continuous & autonomous machine learning, built for real-time usage.

Usage

Install: $ pip install miraiml
Now, inside a Python environment, you can import the main components:

>>> from miraiml import SearchSpace, Config, Engine

You might want to Read the Docs for a better understanding of MiraiML.

Contributing

Please, follow the guidelines if you want to be part of this project.

Comments

Comfortable and clean full status report
Design a way to display the full status of the engine, if requested:

Scores

Ensemble weights

The description of each base model

Would static html files be enough? Maybe a fancy web app?
new feature approved
opened by arthurpaulino 5
API change
Create a Status class with a build_report method and return its instances instead of a dictionary on Engine.request_status

Remove the request_report method from Engine

Split load_data into load_train_data and load_test_data

Return model on MiraiModel.fit

Add a fit boolean parameter on extract_model so the user can choose whether he wants to get a fit model or not

Add a parameter on Config to set an auto stop for the engine

Rename HyperSearchSpace to SearchSpace

new feature better code approved
opened by arthurpaulino 4
Complete example notebook
Add example to index.rst TOC list

Add recommonmark>=0.6.0,<1 and jupyter>=1.0.0,<2 to developer requirements

Generate the markdown file with: python -m jupyter nbconvert --execute --to markdown docs/example.ipynb

Create make directive to compile example

documentation approved
opened by arthurpaulino 1
extract_model documentation

Mention that the extracted model does not output predictions as the engine does because it does not do OOF ensembles for each base model.

Also, mention that it returns None if the engine hasn't completed at least one cycle.
documentation approved

opened by arthurpaulino 1
PipelineComposer

Takes a list of transformers classes (that ends with an estimator class) and a set of parameter names in order to build a wrapper of a pipeline that can be easily instantiated with different parameter values.
new feature approved

opened by arthurpaulino 1
Smarter implementation of MiraiSeeker.seek

Implement a generic code that chooses a search method when random_search is not chosen. This implementation should scale as more search methods are implemented without relying on code update.
easy better code approved

opened by arthurpaulino 1

Improve LightGBM wrapper example

change constructor to:

def __init__(self, **args):
    args.update(dict(
        objective = 'binary',
        metric = 'auc',
        verbosity = -1
    ))
    self.parameters = args
    self.models = None

easy better code approved

opened by arthurpaulino 1

MiraiML: asynchronous, autonomous and continuous Machine Learning in Python

Related tags

Overview

MiraiML

Usage

Contributing

Comments

Comfortable and clean full status report

API change

Complete example notebook

extract_model documentation

PipelineComposer

Smarter implementation of MiraiSeeker.seek

Improve LightGBM wrapper example

Owner

Arthur Paulino

An example project demonstrating how the Autonomous Learning Library can be used to build new reinforcement learning agents.

This solves the autonomous driving issue which is supported by deep learning technology. Given a video, it splits into images and predicts the angle of turning for each frame.

Self-Supervised Pillar Motion Learning for Autonomous Driving (CVPR 2021)

Graph Self-Attention Network for Learning Spatial-Temporal Interaction Representation in Autonomous Driving

Machine Learning From Scratch. Bare bones NumPy implementations of machine learning models and algorithms with a focus on accessibility. Aims to cover everything from linear regression to deep learning.

Vowpal Wabbit is a machine learning system which pushes the frontier of machine learning with techniques such as online, hashing, allreduce, reductions, learning2search, active, and interactive learning.

Learning Continuous Image Representation with Local Implicit Image Function

[CVPR'21] FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space

[ICML 2020] Prediction-Guided Multi-Objective Reinforcement Learning for Continuous Robot Control

On the model-based stochastic value gradient for continuous reinforcement learning

Predicting path with preference based on user demonstration using Maximum Entropy Deep Inverse Reinforcement Learning in a continuous environment

Learning Continuous Signed Distance Functions for Shape Representation

Softlearning is a reinforcement learning framework for training maximum entropy policies in continuous domains. Includes the official implementation of the Soft Actor-Critic algorithm.

Autonomous Ground Vehicle Navigation and Control Simulation Examples in Python

Read Like Humans: Autonomous, Bidirectional and Iterative Language Modeling for Scene Text Recognition

GndNet: Fast ground plane estimation and point cloud segmentation for autonomous vehicles using deep neural networks.

An unofficial personal implementation of UM-Adapt, specifically to tackle joint estimation of panoptic segmentation and depth prediction for autonomous driving datasets.

Code and datasets for the paper "Combining Events and Frames using Recurrent Asynchronous Multimodal Networks for Monocular Depth Prediction" (RA-L, 2021)