Kategorier: Alle - application - machine - processing - learning

av AiSara Mindmap 2 år siden

228

Vision AiSara 2021

An AI Video Trimming Engine is being developed with the goal of providing a mobile application capable of simple yet effective tasks, including gesture recognition, video recording, and editing.

Vision AiSara 2021

Vision AiSara 2021

Marketing and Design

The question that investors will ask is How will you turn your forecasts into reality?
Having identified your product or service, and the demand for it in the market, how will you reach that market?

UI/UX and Design

Scenario

Nodes

Constraint & Events

Model

To understand terms and flow of the web application based on petroleum engineer perspective

To provide UI/UX and flow of the application

Petroleum Engineer

Front-end Developer

Interpretation feature

Data processing feature

Brainstorming app

Some animation in the prototype might not be able to be developed during the development of the application

Cannot export whole artboard into javascript file

New name for the application

To improve the UI/UX and flow of the application

Adobe illustrator

Back-end developer

Front-end developer

Session

Group chat

Chatbot

Social Media Marketing

Create some post with voiceover/persona

Create some engaging content or something in trend

Deliver what AI and ML is to the audience

To give exposure to the audience about AiSara, let people know AiSara and its product

Canva

Grammarly

Microsoft Powerpoint

Adobe Photoshop

Adobe Premiere Pro

Meetedgar

Voiceover Person

Team for content ideas

Content creator

Update social media on the website

News, article, and blog

Quotation post

Tips and Tricks

9am posting/daily

Low engagement

No better reach

Less followers

Delfi

How will the operations of your organization need to change, to support this plan?
Investors will need to know that you will also have the capabilities to deliver it, sustain it and scale it up.

Streamlit into Delfi
Data Driven Production Forecast

Data Modelling for high number of target but limited training variable

2D rate vs time Scenario

Assisted History Matching

Server (local/cloud)

Help to reduce time in generating History Matching scenario and find lowest error model

Performance Issue for big dataset

Improved or targeted UI for Oil and Gas

Visualization

Variable Ranking

Blindtest

3D

AiSara Modelling

Forecast Manager

PRSB

If you win the investment or support that you need, where will you start?

competition with what is already available in the market
client restriction regarding the data usage
Aisara Engine Modelling with limited function for client
Different set of Data used during preparation and client's data
Preparation until end 2021
Mimic PE software but with advance AI/ML modules & AiSara
for PRSB research usage
Create easy interface for engineer to fix PDG data and do analysis with available data

1 Algorithm/ML Engineer

1 Petroleum Engineer

1 Project Lead

General

Finalizing graphic user interface of the web application

ML/AI

Test & Deploy several pre-process modules in web app

Data Related

Testing the dynamic of web app plotting

Finalizing on web application plotting

Petroleum Engineering Physics

Discussion & Research on PE modules

Decision on the modules inside the project (Phase 1)

Stock Market App

It must be clear to investors what you are asking for, when you need it and when & how it will be repaid - in short, what value they will get for their investment.

1 Machine Learning Engineer

1 Data Analyst

Provide Dynamic Analysis and Screening for Existing Market/Stocks

Aivie

To develop a fully functional AI Video Trimming Engine
To develop our own gesture recognition machine
Mobile application with simple task

Chaquopy

ffmpeg for video editing

machine learning module

haarcascade

deep learning module

rembg

opencv

Android Studio

Prototype

Zeplin

Adobe Illustrator

Adobe XD

Machine learning engineer

Calculate the videos to ensure whether it reached 10seconds
Saving videos
Background matting
Splitting the video into frames
Detect stop or stopSubVideo (machine learning classfier)
Recording of video (Android studio)
Detecting gesture in the mobile application (machine learning classifier)
Engine

The gesture recognition does not work as intended

Design Prototype
Back-end

Chaquopy does not support video processing.

Front-end

New Product opportunity

Describe the vision and the opportunity that the plan is based on.
The investor's question you answer here is Where are you headed, and why?

Product 2
Product 1

To provide easiness ......

End time or attached schedule

Main topic

Brainstorming App

Investors will ask: how well do you know your customers and your market? Your business plan should show that you are targeting your market by well-defined segments and are focused on creating value for customers.

Investment schedule
Alpha: 2022

Free Version

Supabase/MongoDB

Firebase & analytics etc

App Hosting

Google Play Store

App Store

Mobile device

IOS

Require MacOS for development/VM

Android

Windows/Ubuntu workstation

Front End Mobile Developer

Backend API Developer

Gamification researcher & implementer

NLP Developer

Provide Easy to Use App for Discussion and Reflection for new idea, problem or scattered idea
Subtopic
Uncertainty: natural bot conversation

Opt 2: GPT3/ DialogFlow (integration and cost)

Require a lot of data to train

Idea management
Chat room for idea collaboration
NLP API

Dev Ops

Local Server:

Networking:



Although the executive summary appears first in the document, it is easier to complete it last, when you can summarise and prioritize the key points in your plan.

Local Server

Provide Internal Developer to host existing server in local server

Adjustable router for external access

Local Server in KL

1 DEvOps Engineer

Setup

NFS-server

Database

App Deployment

Purchase

Networking

Virtual Port setup for hypervisor

Remote Work Router

Hypervisor

Docker

Kubernetes

Allow remote team to access local internet using VPN

Host nfs-server for kubernetes

Host AiSara Core Julia locally at office and accessible anywhere

Host AiSara Core locally at office and accessible anywhere

Host Stock Market App Data Server

AiSara Core Algorithm

Your business plan is a commercially sensitive document and you may wish to add a confidentiality statement (or non-disclosure agreement) at the front of the document.
Type in or attach your statement of confidentiality

Research

Stripping of AiSara Algorithm for PRSB

Ensembling Techniques

API development

Hidden Variable

Accuracy Improvement

Cloud

For production capabilities

2 Developer

Develop enhancement tools for the aisara core algorithm

Proposing and creating new products for AiSara

Improving the accuracy

Enhancement to the aisara core algorithm

Iteration in testing and implementing new tools and techniques

Aisara Web

Internal Server

Marketing tool for Aisara core engine that only utilize AiSara Engine

Internal Engineer testing

Testing with Aisara engine in local Server

Testing with AiSara Core in Julia

Evaluation of all AiSara web

Core Algorithm Julia

Integration of new Julia algorithm with different products of our company

Research on size reduction of docker image

To reduce the load and price of cloud hosting through reducing the hardware allocation

Core Algorithm API


Research:


Benchmarking: (2022)


Purpose and Target

Give Client Copy of AiSara Engine with limited freedom and specific usage only

Add a sum-up of how this investment will be spent. Include:

Schedule

December 2021

API for AiSara 2.0

November 2021

Add a summary of how much is required and when.

API for ensemble algorithm

Resource

Equipment

Server (cloud or local)

Man power

Add highlights of the return or repayment plan for investors. Consider:

1 Developer

Uncertainty/Risk

Minor bug in moving all AiSara core to local server

Technical feature/details

Adjustable AiSara Module for different Client

Preprocess layer which only accept specific type of data or limited number of variables

Limit number of variable

Remove one of the module which are not useable for all-purpose data