MLPro Extension Hub

With this release, we provide numerous minor improvements and fixes. Indipendently, we moved on preparing our upcoming sub-frameworks for supervised learning (MLPro-SL) and online machine learning (MLPro-OA). However, the latter ones are not yet ready to use. Nevertheless, we would like to highlight our new systematics for 3rd party extensions called MLPro Extension Hub. It is described in detail in our online documentation in Section "Extension Hub". Feel free to add your own extensions that can range from nice new examples or cool ML applications up to your complete frameworks for specific topics. Let's share our developments and reach more together.

What's Changed

MLPro-SL - Supervised Learning

• SL: Refactoring PyTorch's Helper Functions by @steveyuwono in #707
• SL: Debugging _adapt_offline() method by @steveyuwono in #746

MLPro-OA - Online Adaptivity

• OA: Online adaptivity12 by @detlefarend in #621
• OA: Initial implementation of core classes by @detlefarend in #710

MLPro-BF - Basic Functions

• BF: Add howto examples MuJoCo with image from camera by @rizkydiprasetya in #721
• BF: Stream sampler by @steveyuwono in #715
• BF: Native streams (dynamic point clouds 2D/3D) by @detlefarend in #703
• BF-Various: Scientific Referencing by @steveyuwono in #755
• BF-Adaptive Systems by @laxmikantbaheti in #699
• BF - Bug Fix in Howto for Normalizers by @laxmikantbaheti in #858

3rd Party Support / Wrappers

• WR: Add Offscreen Viewer MuJoCo and Camera Functionality by @rizkydiprasetya in #713
• WR: Add Gymnasium Wrapper by @rizkydiprasetya in #714
• WR: MuJoCo include body position and orientation as state, init MuJoCo state based on MLPro state space by @rizkydiprasetya in #725
• WR: Add MuJoCo camera fovy to state by @rizkydiprasetya in #727
• WR: Gym removal, SB3 update to 2.0 alpha by @rizkydiprasetya in #741
• WR: MuJoCo add depth data to the state by @rizkydiprasetya in #729
• WR: Stable Baseline 3 by @rizkydiprasetya in #792

Documentation

• RTD: Typos, Class Diagram SL by @detlefarend in #695
• RTD-OA: Activation and initial structure by @detlefarend in #706
• RTD-BF-Streams: Documentation for 3rd Party Support by @laxmikantbaheti in #722
• RTD: Refactoring API References by @steveyuwono in #742
• RTD cleanup by @detlefarend in #844
• MLPro Extension Hub by @detlefarend in #857
• RTD: Refactoring of Howtos by @detlefarend in #859

Other Changes

• Bug: RTD - PDF generation fails #851 by @detlefarend in #860
• RL: Debugging Reward.C_TYPE_EVERY_AGENT by @steveyuwono in #862
• Howtos: Refactoring by @detlefarend in #864

Full Changelog: v1.1.0...v.1.2.0

Improved Persistence

What's Changed

MLPro-BF - Basic Functions

• BF: Repositioning xyz arrow in MuJoCo by @rizkydiprasetya in #686
• BF: Persistence by @detlefarend in #669

Documentation

• RTD: Bugfix of Control Element for Version Selection by @detlefarend in #681
• Add anaconda download badges on Readme by @rizkydiprasetya in #684

Full Changelog: v1.0.2...v1.1.0

Corrections

What's Changed

MLPro-SL - Supervised Learning

• SL: Added missing init in MLPro-SL by @steveyuwono in #667

3rd Party Support / Wrappers

• WR : Added MuJoCo XML in MANIFEST.in by @rizkydiprasetya in #668

Full Changelog: v1.0.0...v1.0.2

Hello World!

What's Changed

MLPro-SL - Supervised Learning

• SL: Initial design by @detlefarend in #435
• SL: Supervised Learning1 by @steveyuwono in #640

MLPro-RL - Reinforcement Learning

• RL: Refactoring by @detlefarend in #515
• RL: Solution for Double Pendulum Env. by @steveyuwono in #398
• RL: Solution of Double Pendulum by @detlefarend in #535
• RL: Add multi-processing on MPC by @steveyuwono in #571

MLPro-OA - Online Adaptivity

• OA: Online adaptivity (preparation) by @detlefarend in #409
• OA: Online Adaptivity by @detlefarend in #504
• OA: Online adaptivity by @detlefarend in #521
• OA: Online adaptivity by @detlefarend in #560
• OA: Online adaptivity by @detlefarend in #593

MLPro-BF - Basic Functions

• BF: Multitasking by @detlefarend in #432
• BF-ML: New classes MLTask, MLWorkflow by @detlefarend in #441
• BF: Normalizer Classes by @laxmikantbaheti in #421
• BF-UI: Fixed dimension access by @detlefarend in #447
• BF-ML: Class MLTask - Event-based adaptation by @detlefarend in #451
• BF-MT: Fixes and enhancements by @detlefarend in #445
• BF-Math: Z-transform update - parameter update based on single new data instance by @laxmikantbaheti in #467
• BF: Extensions of class Dimension by @detlefarend in #480
• BF: Plot extensions by @detlefarend in #477
• BF-Streams: Enhancements by @detlefarend in #482
• BF: New module bf.ops by @detlefarend in #491
• BF-ML/-Plot: Refactoring by @detlefarend in #493
• BF-ML: Refactoring by @detlefarend in #502
• BF-Ops: Refactoring by @detlefarend in #510
• BF-Streams by @detlefarend in #513
• BF-Math-Normalizers: Update parameters of Z-transofrmer with a replaced element by @laxmikantbaheti in #505
• BF-ML: New abstract class AdaptiveFunction #522 by @detlefarend in #523
• BF-Streams by @detlefarend in #514
• BF: New base class FctSTrans by @detlefarend in #550
• BF: Streams by @detlefarend in #547
• BF-Systems: New classes Controller, Actuator, Sensor by @detlefarend in #554
• BF-Streams: Extensions by @detlefarend in #558
• BF: Debugging on TransferFunction and UnitConverter by @steveyuwono in #588

3rd Party Support / Wrappers

• SB3: HER Implementation by @steveyuwono in #414
• Minor improvement: _extract_observation method (Agent class) + PettingZoo by @steveyuwono in #437
• Refactor compatibility for both gym version 0.21.0 and 0.25.0 by @rizkydiprasetya in #580

Documentation

• BF-UI: Fixed dimension access(2) by @detlefarend in #448
• HowTo files: MPC on Grid World Problem by @steveyuwono in #428
• BF: Refactoring Howtos by @detlefarend in #454
• Refactoring HowTo RL/GT by @steveyuwono in #463
• RTD: Basic Functions by @detlefarend in #469
• Bug fixing howto_rl_wp by @steveyuwono in #503
• RTD: Basic functions by @detlefarend in #471
• Fix howto_rl_wp_003 by @steveyuwono in #563
• New class TransferFunction by @steveyuwono in #583
• RTD: Final Documentation v1.0.0 by @detlefarend in #585
• RTD: Final Documentation v1.0.0-2 by @steveyuwono in #624
• RTD: Final Documentation v1.0.0-3 by @detlefarend in #629

Quality Assurance

• Quality Assurance v1.0.0 by @detlefarend in #586
• Quality Assurance v1.0.0-2 by @steveyuwono in #625
• Quality Assurance v1.0.0-3 by @laxmikantbaheti in #644

Other Changes

• Autodoc Generation CI by @rizkydiprasetya in #410
• Move ROS related environment to another repo by @rizkydiprasetya in #545
• Simulation of MPPS by @steveyuwono in #386
• Temp fix for breaking changes from Numpy by @rizkydiprasetya in #565
• WR: MuJoCo Integration by @rizkydiprasetya in #610

Full Changelog: v0.9.2...v1.0.0

Pre-Release to try out

What's Changed

• BF: Root class Wrapper by @detlefarend in #371
• BF: Event handling by @detlefarend in #379
• RL: Action Planner using MPC by @steveyuwono in #366
• BF-ML-Model: Event Handling by @detlefarend in #381
• OA: Online adaptivity2 by @detlefarend in #375
• RL: Updating BGLP by @steveyuwono in #385
• BF-ML: Renaming maturity to accuracy by @steveyuwono in #394
• RL: Double pendulum environment by @steveyuwono in #323
• ADMIN: Conda build by @rizkydiprasetya in #397
• BF: Multiprocessing by @detlefarend in #393
• OA: Online adaptivity 3 by @detlefarend in #383
• OA: Initial class design by @detlefarend in #401
• BF: Extensions by @detlefarend in #403
• BF: Extensions(2) by @detlefarend in #404
• RTD: Documentation (Hackathon - 26/08/2022 and 02/09/2022) by @steveyuwono in #392

Full Changelog: v0.9.1...v0.9.2

Pre-Release to try out

What's Changed

• MLPro: OA by @laxmikantbaheti in #358
• UR5 Update by @rizkydiprasetya in #364
• Update due to the latest introduction of Gym 0.25.0 by @steveyuwono in #369

Full Changelog: v0.9.0...v0.9.1

Pre-Release to try out

What's Changed

• Release of first version of random generator policy by @steveyuwono in #329
• Automate Build UR5 ROS Workspace by @rizkydiprasetya in #339
• Add gazebo visualization GUI toggle for UR5 by @rizkydiprasetya in #340
• Bglp update by @steveyuwono in #341
• Homogeneous Transformation Matrix by @rizkydiprasetya in #336
• Read the docs by @steveyuwono in #342
• Bglp by @steveyuwono in #350
• SB3 Wrapper: Enable the possibility to process reward type C_TYPE_EVERY_AGENT by @steveyuwono in #351
• Data Stream Mining by @detlefarend in #244
• UR5 Real ROS and Socket based by @rizkydiprasetya in #356
• Add termination condition for batch production scenario by @steveyuwono in #363

Full Changelog: v0.8.6...v0.9.0

Pre-Release to try out

Full Changelog: v0.8.5...v0.8.6

Pre-Release to try out

What's Changed

• RL-POOL: Renaming and Refactoring MLP Envmodel for Robotinhtm by @rizkydiprasetya in #238
• Model Improvements 8 by @detlefarend in #239
• RTD Updates by @detlefarend in #235
• Data stream mining by @detlefarend in #243
• SB3 off policy algorithm by @rizkydiprasetya in #256
• Ur5 fix and SB3 fixed using 1.3.0 by @rizkydiprasetya in #259
• fix sb3 by @rizkydiprasetya in #263
• PettingZoo Wrapper Update by @steveyuwono in #262
• OpenAi gym wrapper, SB3 test script, robotinhtm env by @rizkydiprasetya in #264
• Model improvements #9 by @detlefarend in #266
• RL-Train: turned off training data logging on eval by @detlefarend in #270
• Reformat Code by @rizkydiprasetya in #272
• Use deterministic for evaluation by @rizkydiprasetya in #273
• Hyperopt wrapper + bglp update by @steveyuwono in #268
• Hyperopt by @steveyuwono in #274
• Model Improvements 11 by @detlefarend in #280
• Hyperopt DataStoring by @steveyuwono in #281
• Update _adapt method in MultiAgent class by @steveyuwono in #282
• Double Pendulum Pool Env by @budiatmadjajaWill in #286
• Rtd updates by @budiatmadjajaWill in #257
• RL-Train: new param p_stagnation_entry by @detlefarend in #289
• RL-Train: Opt. continue after stagnation detection by @detlefarend in #290
• Model improvements15 by @detlefarend in #292
• Minor Improvement on Class MultiAgent by @steveyuwono in #295
• RL-Env: Include gripper function in UR5 environment by @budiatmadjajaWill in #299
• Refactoring due to PettingZoo version 1.17.0 by @steveyuwono in #302
• Optuna Wrapper Class by @steveyuwono in #303
• Optuna Wrapper: Change methods names by @steveyuwono in #305
• Optuna Wrapper: Add tuning recap visualization by @steveyuwono in #307
• Doublependulum env improvement by @detlefarend in #311
• Model improvements20 by @laxmikantbaheti in #314
• Rtd updates3 by @rizkydiprasetya in #308
• Howto 20 and auto detect ros workspace by @rizkydiprasetya in #317
• Wrapper Gym Enhancement by @steveyuwono in #328

Full Changelog: v0.8.1...v0.8.5

Pre-Release to try out

What's Changed

• RTD Edit by @rizkydiprasetya in #233

Full Changelog: v0.8.0...v0.8.1