Multidisciplinary Neuroscientific Technologies Ltd / Harry Muzart
H: Home Page - BioNeuroTech
DCT: Developer Coding Tools, Programming Platforms, Base Code Info
RP: Research & Pre-Publication Works
A: About BNT, and Products & Services
L: Lab Experiment Methods (VR-Simu/DIY-Tech, Applications)
​
​
​
Computational Modelling & Implementation :
C-bCM-S: Computational (Bio)Chemical & Molecular Systems [⋆⋆]
C-NhaNp-S: Comp. NeuroHistoAnatomical & NeuroPhysiological Sys. [⋆⋆⋆⋆⋆⋆⋆]
N: Note on Neural Networks
C-OBNR-S: Comp. Organismal BioMedical & Neuro-Robotic Sys. [⋆⋆⋆⋆]
C-Cogn-S: Computational Cognitive Systems (non-A.I. & A.I.) [⋆⋆⋆⋆⋆⋆⋆]
C-mE-S: Computational Macro-Economic (Populations) Systems [⋆]
Gen: Generalist & Misc Projects [⋆⋆]
​
​
​
iML: Interactive Deep Machine Learning Data Apps (webpage-based) [⋆⋆]
​ R: Repository of External Projects
W: Website API/IDE {TBC}
​ D: Disclaimer (Legal Policies) / Intellectual Property / Commercial / AdSense
C-Cogn-S: Computational Cognitive Systems
(non-A.I. & A.I.) [⋆⋆⋆⋆⋆⋆⋆]
​
​
Animal/Human + non-A.I. computer/machine + A.I. computer/machine
​
​
​
The term 'Cognitive' here refers to either the human user's own cognition interacting with a computer program and learning from it; or it can refer to an 'intelligent program' which may or may not use A.I. (Artificial Intelligence, strong general-purpose AI --- ie. DML, deep machine learning).
Android App #1 (Java) – 'Flying Neural Net' Basic non-AI 2D Game
The link to my Google Play Account where you can download my published Android Apps:
https://play.google.com/store/apps/developer?id=BNT+AppLab+%28Multidisc+Neurosci+Tech+Ltd%29
​
​
​
​
https://developer.android.com/distribute/console/index.html
Developer Console
Project Fields:​ 2D Games Development & Design
Description: xxxxx
My Work & Contributions: Coded all from sctratch – 5000 lines of code +
Collaborations/Feedback: xxxxx
Location Activities: Home-based
Status/Timeline:
Started: Sept 2016
Most Research Done: Jan 2017
Self-testing done: April 2017
Published to Beta testing: May 2017
Late 2017: Full Production
2018+: Monetization revenue (TBC)
Other Links to external sources:
--- JDK 8.0 JRESE https://java.com/en/download/win10.jsp
--- https://developer.android.com/studio/index.html
--- https://android.stackexchange.com/
--- https://github.com/tensorflow/tensorboard
--- www.cocos2d-x.org/ Cocos2d-x Open-Source Game Development Platform
--- https://docs.unity3d.com/Manual/index.html
--- https://github.com/Harry-Muzart
​
Hardware: Intel Quad Core i7 & Celeron Computers; Android Acer Tablet; Phone
Operating Programs & Software Tools: MS Win 10 OS; Android OS; Android Studio 2.3, Eclipse.
Programming Languages: Java & XML (but C++ could be used for native)
Coding Elements: OnClickL; ShPrefs (AdMob); Java 7.0 and AS (.jva, .xml, gradle build)
Computational Techniques: xxxxxxxxxx
Link to Programming Code Files:
Public (full open access) (Google Drive Cloud and GitHub): https://drive.google.com/drive/folders/0B0qSKFqszohLTTZIZkg1c09xaVU?usp=sharing
Private (shared password-protected) (Google Drive Cloud): XXXXX
File types: xxxxx
​License: xxxxx
Android Studio Project (Blank)
--- MyApplication14oct2017 >
1595 files (665 subfolders) 39.4 MB
app > src > main > res > layout
activity_main_activity1a7b.xml 1 KB
app > src > main > java > com > example > harrymuzart > myapplication14oct2017 >
MainActivity1a7b.java
app > src > main
AndroidManifest.xml
build.gradle
​
Base Math Coding & Program Computations & System (Diagrammatic):
See Image below
Images below: Content Files by Me, Screenshots by Me.
Video Embeds below: see my YouTube channel.
​
For other Pics/Vids, see:
https://photos.app.goo.gl/UK6yX8SSnARWdJjw6
​
...
Actual App:
TBC:
 |  |  |
|---|---|---|
 |  |  |
 |  |  |
 |
Other Android Apps – In development:
​
--- Flying Neural Net (Basic Game) 1.0 (see above)
--- Neuroscience (Interdisciplinary) eLearning Quizzes [TBC]
--- Digital Health Data social app (like 'Streams' by Google DeepMind Health)
--- A real-time image object classier (using ML like https://github.com/llSourcell/YOLO_Object_Detection) using Android camera, which can then be used for other purposes.
--- Psychiatry patients CBT App (Like SloMo by KCL IoPPN, also others by UCL ICN and Imperial)
--- A Mediation/Mindfulness App (based on literature in audio-visual-tactile cogn neurosci), similar to much already on the market.
--- A brain-training IQ-test-like App
--- Interactive Brain Connectomics & 3D Rotations (like by CSH Labs and Allen Institute)
--- RSS Reader & News Feed
--- Social Feed
--- UCL Navigational Map App
[Images and videos available for each of the above.]
​--- (*) Cognitive Sciences, Cognitive Behaviour, eLearning, Games, Game Theory Mechanics, Communications, HCI, Educ., Graphical 3D models {Cognition & Interactive Comms}
​
Computational & Cognitive Neuroscience relevancy:
​
About CN In terms of content
---CN Comput
---N Cognitive
---N
How built - CN In terms of applications
---CN Computational Neuroscience: Base Coding for Intelligent Systems
---Neurobiol Labs - hands-on end-user applications in ucl/nhs
---Cognitive Neuroscience: Design and Management
End user CN
---CN
---N Physical Integration
---N
Windows Phone App
​
*** NeuroCogFeeds (App by HM)
​
(some accessibility discontinued)
App Developed (Windows 8 Phone) (not yet published in Windows App Store)
--- This was done using User-Interface App Development in Visual Studio, not actual programming (which would’ve been C++/C#/.NET/VisBASIC/Obj-C in this case)
--- Human-Computer Interaction leading to eLearning (learning by interaction and visualization).
--- This app is different from other platforms because:
(a) you do not need to access each feed manually, it would save people a lot of time (and the next app version will feature about 100 more feed lists).
and (b) the app does not depend on an internet connection, nor do the archived feeds, only the newest data depends on an internet connection; also, if one website has a server problem, it will not affect other parts of the app.
https://drive.google.com/drive/folders/0B0qSKFqszohLTTZIZkg1c09xaVU?usp=sharing
​
https://github.com/Harry-Muzart > //hm_proj_002#//
​
https://developer.microsoft.com/en-us/ > https://appstudio.windows.com/en-us/
video: youtube TBC
https://github.com/Harry-Muzart/hm-project-2/blob/master/NeuroCogFeeds.Shared.shproj
Build Variable Module
Other Links to external sources:
 |
|---|
 |
Machine Learning (with Python & Extensions)
​
*** Python ML Mini-Project 'AlphaMem'
Project still being worked on - see GitHub for my (very slow!) progress towards decoding all this
My articles:
https://sciarticles.wordpress.com/article-2/ > //latter_part// (in Python v3.0+)
​
https://github.com/Harry-Muzart > //hm_proj_001#// > Jupyter code
​
My links to external resources:
https://scientifically.org.uk > //-2e-comp-neuro-ml// > TensorFlow neural network simulation.
​
​https://www.youtube.com/playlist?list=PL3-vVGjXPjaIFCik7cD-6AYmvZNWbNJfh
https://www.youtube.com/playlist?list=PL3-vVGjXPjaKhlqCcMeZMoIsK8RD00bNk
​
Images: Content by me, Screenshots by me.
Project Fields:​ xxxxx
Description: xxxxx
My Work & Contributions: xxxxx
Collaborations/Feedback: xxxxx
Location Activities: xxxxx
Status/Timeline: xxxxxx
Other Links to external sources:
https://github.com/Harry-Muzart/tensorflow
https://www.python.org/downloads/
https://pypi.python.org/pypi/numpy
https://sebastianraschka.com/deep-learning-resources.html
Hardware: Intel Quad Core i7 & Celeron Computers;
Operating Programs & Software Tools: MS Win 10 OS; Python 3.6, Shell, IDLE; Java packages; Tensorflow 1.3
Programming Languages: Python
Coding Elements: xxxxxx
Computational Techniques: xxxxxxxxxx
Link to Programming Code Files:
Public (full open access) (Google Drive Cloud and GitHub): https://drive.google.com/drive/folders/0B0qSKFqszohLTTZIZkg1c09xaVU?usp=sharing
Private (shared password-protected) (Google Drive Cloud): XXXXX
File types: .py
​License: xxxxx
--- lib site packages numpy matlib.py (2 KB)
--- 19 py files - testings (20.7KB)
--- MASTER_NN_012.py (16KB)
--- about_MASTER_NN_012_file.txt (1KB)
Base Coding & Program Computations & System (Diagrammatic):
XXXXXX
Images:
Content Files by Me, Screenshots by Me
Video Embeds: (video will be made and uploaded soon)
|
|---|
|
Deep Convolution Neural Networks and Object Recognition
Also see : https://www.bioneurotech.com/iwa-dml
For the links to all the code, see:
https://github.com/Harry-Muzart/harry-muzart.github.io
https://harry-muzart.github.io/
This includes work on the video below.
For each frame, the data output looks like this:
[{"label": "person", "confidence": 0.92, "topleft": {"x": 82, "y": 30}, "bottomright": {"x": 733, "y": 525}}, {"label": "tie", "confidence": 0.91, "topleft": {"x": 331, "y": 411}, "bottomright": {"x": 421, "y": 525}}] [{"label": "person", "confidence": 0.92, "topleft": {"x": 82, "y": 30}, "bottomright": {"x": 733, "y": 525}}, {"label": "tie", "confidence": 0.91, "topleft": {"x": 331, "y": 411}, "bottomright": {"x": 421, "y": 525}}] [{"label": "person", "confidence": 0.92, "topleft": {"x": 82, "y": 30}, "bottomright": {"x": 733, "y": 525}}, {"label": "tie", "confidence": 0.91, "topleft": {"x": 331, "y": 411}, "bottomright": {"x": 421, "y": 525}}] [{"label": "person", "confidence": 0.92, "topleft": {"x": 82, "y": 30}, "bottomright": {"x": 733, "y": 525}}, {"label": "tie", "confidence": 0.91, "topleft": {"x": 331, "y": 411}, "bottomright": {"x": 421, "y": 525}}] [{"label": "person", "confidence": 0.92, "topleft": {"x": 82, "y": 30}, "bottomright": {"x": 733, "y": 525}}, {"label": "tie", "confidence": 0.91, "topleft": {"x": 331, "y": 411}, "bottomright": {"x": 421, "y": 525}}] [{"label": "person", "confidence": 0.92, "topleft": {"x": 82, "y": 30}, "bottomright": {"x": 733, "y": 525}}, {"label": "tie", "confidence": 0.91, "topleft": {"x": 331, "y": 411}, "bottomright": {"x": 421, "y": 525}}] [{"label": "person", "confidence": 0.92, "topleft": {"x": 82, "y": 30}, "bottomright": {"x": 733, "y": 525}}, {"label": "tie", "confidence": 0.91, "topleft": {"x": 331, "y": 411}, "bottomright": {"x": 421, "y": 525}}] [{"label": "person", "confidence": 0.92, "topleft": {"x": 82, "y": 30}, "bottomright": {"x": 733, "y": 525}}, {"label": "tie", "confidence": 0.91, "topleft": {"x": 331, "y": 411}, "bottomright": {"x": 421, "y": 525}}] [{"label": "person", "confidence": 0.92, "topleft": {"x": 82, "y": 30}, "bottomright": {"x": 733, "y": 525}}, {"label": "tie", "confidence": 0.91, "topleft": {"x": 331, "y": 411}, "bottomright": {"x": 421, "y": 525}}]
​
​
​
For a different cloud-supported Matlab version, see:
https://photos.app.goo.gl/UWCuHimwxddnNFZw5
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
Reinforcement Learning
Deep Reinforcement Learning AI (Agent POV) (Backwards Replay of Finding Reward-based Navigation Solutions to 3D Maze)
Video adaptation by Harry Muzart
* THIS VIDEO CONTAINS NO AUDIO
Original Work - Code & Research:
https://deepmind.com/research/
https://github.com/deepmind/lab
https://arxiv.org/abs/1612.03801
https://www.bioneurotech.com
https://github.com/Harry-Muzart
Google DeepMind videos:
Channel: https://www.youtube.com/channel/UCP7j...
Training Environments: https://www.youtube.com/playlist?list...
Lab Nav Maze Lvl 1: https://www.youtube.com/watch?v=M40rN...
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
​
MDP (Markov Decision Processes) and RQL (Reinforcement Q Learning) (in Matlab)
​
https://github.com/Harry-Muzart/MDP_RL_HM_model_1x
​
ANN agents in grid env with actions A (via transition matrices), and states S, and rewards R; trained over many epochs.
​
(not free energy Bayesian model optimization or active inference)
​
see:
​
https://uk.mathworks.com/products.html?s_tid=gn_ps
https://uk.mathworks.com/help/referencelist.html?type=function&s_tid=CRUX_gn_function
https://uk.mathworks.com/matlabcentral/profile/authors/9089185-harry-muzart
​
​
​
 |  |
|---|---|
 |  |
 |  |
Game in Python/Other, using GUI Visual Display (see SourceForge Apps)
https://sourceforge.net/u/harry-muzart-01/download-apps/
https://sourceforge.net/u/harry-muzart-01/download-apps/2019/05/datax45654k/
​
A 2D game. A basic agent navigates a 2D grid world and need to find rewards pellets and avoid negative items. (Some elements not yet finished)
My file:
--- Pygame.py (2KB)
External by others:
|
|---|
|
TBC
Chrome Extension for ML Analytics linked to User-based Online Databases
​
BNT.ext for Chrome
​
"BNT Extension" (deployed but doesn't do anything yet, full version TBC).
---1--- G Web App Scripts of Chrome Developer Environment on Windows 10 PC (using JavaScript, HTML5, CSS) ---> Chrome (by Google) Internet Browser *Extensions*, for any OS (G Chrome OS, MS Windows 7+, iOS/Mac, Linux, Other --- for any desktop PC / laptop / tablet / phone). {Internet connection needed}
--- Apps will be about Computational Neuroscience / Machine Learning , and they will aim to be directly use Computational Neuroscience / Machine Learning.
--- App Link - Science Papers Preferences Suggestions Feeds
--- Extensions - Sentiment Analysis of Text and Images
--- Website Tools - Database SQL PHP
Other Links to external sources:
https://developers.google.com/apps-script/
https://developer.chrome.com/home
My files:
​
--- icon.png 34KB
--- manifest,json 1KB
--- popupBNT.html 2KB
--- popup.js 5KB
 |  |
|---|
3D Virtual Reality Simulation (in Unity Engine), for Human & AI Spatial Navigation ; + 360 3D Augmented Reality Applications Dev
​
Currently being developed:
A 3D simulation similar to the GTA5/Doom game engine framework mods, which is done is Unity 2.0, using Visual Studio on MS Windows, and that will be trainable on machine learning algorithms. This will be used as a test-bed for spatial navigation in first and third person.
 |  |  |
|---|---|---|
 |  |
​
For other Interactive 360 3D Virtual Reality, 360 3D Video-Audio Navigation, 360 3D Augmented Reality:
#1 - Testing Existing Programs (e.g. from Google AR).
#2 - Contribute feedback and/or code to those in beta mode.
#3 - Inspirations to building my own.
​​
https://photos.app.goo.gl/Sf1jFH32AfVGmPD59
​
​
​
​
​
​
Other
My Other & Progressing & Future Work, and external, etc.
​
​
#### My Future Projects (currently being researched) ####
--- Web ML with Python and SQLite
--- ML/AI Systems for Varied Applications
--- Cognitive Analytics for Visual Recognition, Language Comprehension, Spatial Predictions
--- Cognitive analytics for visual recognition, language comprehension, spatial predictions
--- Intelligent internet networking social forum
--- Unity3D/Unreal Game, GTA Mod for AI
--- Unity3D/Unreal Mod Game Simulator
--- Visual Anim Graphics Video Programming
--- DeepMind Open-Source lab
--- Prolog for AI
--- Go, Lua and Ruby for AI
--- iOS app using Swift
--- Google Assistant App
--- HCI
Other future works will be related to these other external independent projects by other people/organisations listed below.
​
Here is a new list :
​
â–º https://sites.google.com/site/bioneurotech/-repo-2
â–º https://docs.google.com/document/d/1Nv6BimGLFgvbSYeNQZO5Cjna_iWcquwoF47MMsll5fA/edit?usp=sharing
Here is a previous list...
Also See:
--- StackExchange/GitHub on Google Ai Expt
--- Google AI experiments
https://experiments.withgoogle.com/ai
Also see:
All Prior Art - An amazing project that uses deep machine learning algorithms in order to attempt to automatically generate innovate ideas and publicly publish all possible new 'prior art' (see https://www.epo.org/learning-events/materials/inventors-handbook/novelty/prior-art.html), thereby making the published concepts not patent-able. The purpose is to democratize ideas, provide an incentive for reform in the US and international patent systems, and to pre-empt patent trolls.
|
|---|
|
#