Platform to experiment with machine understanding

chevron-icon
Back
project-presentation-img
Presentation
Denis Shabratov
Project Owner

Platform to experiment with machine understanding

Funding Requested

$141,000 USD

Expert Review
Star Filled Image Star Filled Image Star Filled Image Star Filled Image Star Filled Image 0
Community
Star Filled Image Star Filled Image Star Filled Image Star Filled Image Star Filled Image 3.7 (7)

Overview

The new combinatorial ML approach is able to build models using limited data to generate solutions for complex technical and scientific problems and to train a machine understanding of solved problems and learned methods. The project is to create an environment in which everyone can train machine understanding stage by stage on examples of problems from simple to more and more complex, develop a natural language with abstract terms to indicate the features of problems, methods and solutions, and supervise the problem solving by the trained AI using the language. Based on ML agent from combinatorial.ai.

Proposal Description

How Our Project Will Contribute To The Growth Of The Decentralized AI Platform

The new ML approach can build models using limited data to generate solutions for complex technical and scientific problems and train a machine understanding of the problems and the solutions. It opens ways to use ML with machine understanding for broad user adoption by researchers and innovators seeking new solutions
The open platform will provide tools for studying, testing, and using ML with machine understanding for everyone and will give new paths to new business models and markets

Our Team

Our team brings together developers of fundamental machine learning and training algorithms, programmers with experience building their own ML learning agents, and programmers with business experience applying machine learning techniques to products for markets.

View Team

AI services (New or Existing)

Development ML model with machine understanding

Type

New AI service

Purpose

Development of ML models with machine understanding. This entails a human-understandable description of the problem-solving method making the model understandable for both developers and users. The architecture of these models allows for their reuse as a complete solution or as part of other constituent solutions.

AI inputs

Training program for a specific problem. Training bots.

AI outputs

ML model with machine understanding for specific problem or problem domains.

Company Name (if applicable)

APPARATUS SCIENTIA d.o.o.

The core problem we are aiming to solve

Substantial requirements for computational resources and training data volume, the underlying property to generate "hallucinations", as well as an inability to abstract and algorithmic reasoning, are crucial limits for state-of-the-art ML which poses a serious challenge to its applications in engineering, R&D, and hard science. 
These methodological challenges of neural networks prevent the implementation of machine understanding, hence human understanding and control of the reasons for machine decision-making.
The inability to train on a limited amount of data results in significant barriers to applying machine learning in most specific problem domains, which form the core of human knowledge and technologies.

Our specific solution to this problem

We designed the architecture and algorithms that set us apart from the trend of neural network development. This distinction arose from our understanding of the deep limitations inherent in neural network technology, particularly in getting closer to natural intelligence by functionality, not by arrangement, as in the neural network. 
Our approach is based on the implementation of stage-by-stage training from simple to complex, starting with mastering the basic terms of a problem domain and progressing to solving increasingly complex problems that require the ability to combine simple solutions. This culminates in machine understanding, which is the ability to choose the right methods for solving problems and explain the method choices.
Our approach and experience in implementing ML from simple to complex by stage-by-stage training, which also addresses the challenge of algorithmic reasoning in learning agents, is disclosed in patent application PCT/EP2024/059055.

Project details

The project aims to build a publicly available platform that allows everyone to use a new machine learning approach based on stage-by-stage training from simple to complex and create and test their models for solving specific problems and problem domains based on the machine understanding of the problem domain that forms as a result of training. 
Other platform users can share and reuse the created models as a complete solution or as part of creating more complex solutions and models. Thus, the market for models based on machine understanding, as well as the services market, can start to form by solving the full range of scientific, technical, engineering, and production problems using these models. 

Stage-by-stage training from simple to complex
In many applications, including programming, the solutions to complex problems are combinations of elements found in solving simple problems.
While the neural network aims to find a solution for the whole problem, the stage-by-stage training approach decomposes complex problems with long solutions into sub-tasks with shorter solutions. This results in an exponential reduction of training time and required data volume. 
Combining solutions to complex problems can require transforming the elements of simple problem solutions. The ability to modify allows the transformed elements to be reused as solution elements in solving new problems. In this way, the stage-by-stage training of problem-solving in new domains is organized.
The stage-by-stage training method can be used to train signs the AI agent learns to use to make assumptions about ways of finding a solution to a complex problem—ways of combining elements and their transformation for successful combining. When the agent is trained to express the elements of a solution and how they can be combined in an external sign system, it starts exhibiting machine understanding as the ability to choose the right methods for solving problems and explain the choices.
Based on the explanation received, the developer or user either accepts it or modifies the problem definition or the training program to change the machine's decision-making and reasoning. In this way, a means of human control for machine decision-making and reasoning can be implemented.

Combinatorial AI agent
To learn from simple to complex, a learning agent must give higher priority to combining elements of simple solutions over other mechanisms of producing solutions for complex problems and must have mechanisms for transforming the elements when combining them.
Our approach and experience in implementing these mechanisms, which also addresses the challenge of algorithmic reasoning in learning agents, is disclosed in patent application PCT/EP2024/059055

Training 
To achieve machine understanding, a supervisor needs a comprehensive representation of the state of the training process. The basic objects in the representation are training bots that communicate with the learning agent and each other to make rewards for the agent’s responses to their messages. The training environment provides a means for the supervisor to develop a training program for a particular problem domain as a tangling system of training bots, divide it into stages, introduce hints at each stage, and evaluate the training progress. In the evolutionary approach to training, the training environment can maintain a population of learning agents and perform selection by their ability to get rewards. Additional tools may be required to configure bots and agents, test their interaction scenarios, support versioning and revision control for training programs and trained models, and benchmark the development of machine understanding in the problem domain.

Milestone 1 - Project Initiation and Architecture Design.
6 weeks
Deliverables: 
•    Set up development environments, necessary toolchains, and databases. 
•    Completed solution architecture schema and a data flow chart

Milestone 2 –Training environment specifications.
8 weeks

    Training bot API
Training bots are objects that feed the learning agents with training data and make rewards for their responses. The API is needed to provide diagnostic information or make the training function manageable by other bots or the training environment.
    Deliverable – training bot API specification

    Learning agent API
Learning agents can be considered as bots but with additional functionality. The environment may need to signal whether to upload a pre-trained model to save the trained model and the exploration/exploitation value for the training session.
    Deliverable - agent API specification

Milestone 3 – Training language and Training bot language processor
8 weeks

    Training bot language
Training bots can be programmed using traditional algorithmic or functional languages, but the experience of creating training programs shows that these programs become complicated too quickly because of training specifics. A different approach based on regular expressions can make training programs more compact and easy to read.
    Deliverable – training bot language specification

    Training bot language processor
A means of processing code in the language for the training bots must be realized either as a compiler, translator, or interpreter.
    Deliverables – training bot language processor executable, documentation.

Milestone 4 – Training bot development environment.
8 weeks

    Training bot development environment
Training programs code used to grow and branch fast. Additional means may be required to navigate through the system of training bots that constitutes the training program, represent the training program in a compact and straightforward manner, and edit it visually besides using standard development tools like SVN.
    Deliverable – training bot IDE, documentation.

Training environment scripting language
Although any training program can be coded in bots or even using traditional programming languages, there is a specifics about training machine understanding, e.g. stage-by-stage training with additional functions built on the stage representation. It’s proposed to implement the specifics separately using a script language of the environment. 
    Deliverable – training program scripting language specification.

    Training environment script language processor
Means to process code in the script language used to configure the training environment must be realized.
    Deliverable - training environment script language processor source code and executable, documentation.

Milestone 5 –Training environment user interface.
8 weeks

    Training environment user interface
User interface of the training environment provides tools to manage training bots and training data, test communication scenarios between the bots and the agents, configure tracing, combine SVN for training programs and trained models, save trained models on disk and restore models, create learning agent populations and automate the selection of models as in evolutionary approach, configure exploration/exploitation parameter and benchmark the training progress.
    Deliverables – training environment user interface source code, executable and documentation.

    Tracing interfaces
Tracing module can be considered as necessary part of the training environment. Just as regular tracing techniques in programming help to detect unexpected behavior in communication between the agents and the training system, tracing also helps to see training in real-time before managing or reporting bots are implemented for the training program.
    Deliverables – API specification for tracing from bots and agents, the user interface to manage to trace output from different sources and its backend, documentation.

Milestone 6 –  Use cases
8 weeks

    Training program
A straightforward example of a training program that leads to development of machine understanding, expressed in terms of stages.
    Deliverables – training program, training bots, trained agents for each stage of the training program, supervision guidance.

     Training bots and their interoperation scenarios
The training environment is a medium for training bots and training agents that may communicate with each other in order to complete the training program. The variations of bots and their interoperation scenarios, which may deviate from the training program mainstream, are needed to demonstrate the flexibility of the training system and the generalizing ability of the approach.
    Deliverables – examples of training bots and interoperation scenarios with descriptions of expected effects.

    Tests of bots and interoperation scenarios
Tests are mostly bots, which are complementary to training bots as they communicate to the tested bots or take part in interoperation scenarios just to diagnose and report to human supervisor, or other bot, or training environment if the system works as expected.
    Deliverables – examples of bots, interoperation scenarios and their tests, and specification of the reporting interfaces.

     Default training bot for communication and supervision
One or more training bots may provide a specific interface to human users or supervisors that can be used to “touch” machine understanding by interrupting the training programs with problems not from the training dataset to see the immediate response(s) from the agent(s) and evaluate them. 
    Deliverable – the default training bot, a guideline on how to see the development of machine understanding in the agent using the bot.

    Tracing scripts examples
Tracing is a tool for developing effective training programs. It provides data for a quick view of the communication that takes place in the environment between the bots and the agents.
    Deliverable – tracing script examples with the description of their value for the development of the training program.

Milestone 7 – Onboading. Beta release
4 weeks
Deliverable – Beta release

Competition and USPs

We have a unique value proposition (UVP). We provide the ability for the user to build the new type of ML models with machine understanding by stage by stage training on examples of problems from simple to more and more complex, developing the language with abstract terms to indicate the features of problems, methods, and solutions, and supervise the problem solving by the trained AI using the language.

This entails a human-understandable description of the problem-solving method by machine, making the ML model understandable for both developers and users. The architecture of these ML models allows for their reuse as a complete solution or as part of other constituent solutions, setting us revolutionary apart from our competitors.

Needed resources

We plan to attract resources for the development of interface design and platform testing.

Links and references

www.combinatorial.ai

Revenue Sharing Model

Custom Model

Custom Description:

The platform's unique proposition lies in its development of ML models with machine understanding. This entails a human-understandable description of the problem-solving method, making the model understandable for both developers and users. The architecture of these models allows for their reuse as a complete solution or as part of other constituent solutions, setting us revolutionary apart from our competitors.
The platform's business model implies that developers sell their own solutions to their models, both within the platform and as a service for problem-solving for external users.
The platform's revenue part will follow the model of AppStore, Google Play, SingularityNET AI marketplace, and others.

Proposal Video

DF Spotlight Day - DFR4 - Pavel Malyshkin - Platform To Experiment With Machine Understanding

4 June 2024
  • Total Milestones

    7

  • Total Budget

    $141,000 USD

  • Last Updated

    4 Jun 2024

Milestone 1 - API Calls & Hostings

Description

This milestone represents the required reservation of 25% of your total requested budget for API calls or hosting costs. Because it is required we have prefilled it for you and it cannot be removed or adapted.

Deliverables

You can use this amount for payment of API calls on our platform. Use it to call other services or use it as a marketing instrument to have other parties try out your service. Alternatively you can use it to pay for hosting and computing costs.

Budget

$35,250 USD

Milestone 2 - Project Initiation and Architecture Design

Description

This phase focuses on establishing the project infrastructure setting up development environments and updating the platform's plan and architecture.

Deliverables

• Set up development environments necessary toolchains and databases. • Completed solution architecture schema and a data flow chart

Budget

$7,750 USD

Milestone 3 - Development of Training environment specifications

Description

Training bot API Training bots are objects that feed the learning agents with training data and make rewards for their responses. The API is needed to provide diagnostic information or make the training function manageable by other bots or the training environment. Learning agent API Learning agents can be considered as bots but with additional functionality. The environment may need to signal whether to upload a pre-trained model to save the trained model and what is the exploration/exploitation value for the training session.

Deliverables

• Training bot API specification • Agent API specification

Budget

$18,000 USD

Milestone 4 - Training bot development environment

Description

Development of Training bot development environment Training bot development environment Training programs code used to grow and branch fast. Additional means may be required to navigate through the system of training bots that constitutes the training program represent the training program in a compact and straightforward manner and edit it visually besides using standard development tools like SVN. Training environment scripting language Although any training program can be coded in bots or even using traditional programming languages there is a specifics about training machine understanding e.g. stage-by-stage training with additional functions built on the stage representation. It’s proposed to implement the specifics separately using a script language of the environment. Training environment script language processor Means to process code in the script language used to configure the training environment must be realized.

Deliverables

• Training bot IDE documentation. • Training program scripting language specification. • Training environment script language processor source code and executable documentation.

Budget

$29,000 USD

Milestone 5 - Development of Training environment user interface

Description

Training environment user interface User interface of the training environment provides tools to manage training bots and training data test communication scenarios between the bots and the agents configure tracing combine SVN for training programs and trained models save trained models on disk and restore models create learning agent populations and automate the selection of models as in evolutionary approach configure exploration/exploitation parameter and benchmark the training progress. Deliverables – training environment user interface source code executable and documentation. Tracing interfaces Tracing module can be considered as necessary part of the training environment. Just as regular tracing techniques in programming help to detect unexpected behavior in communication between the agents and the training system tracing also helps to see training in real-time before managing or reporting bots are implemented for the training program. Deliverables – API specification for tracing from bots and agents the user interface to manage to trace output from different sources and its backend documentation.

Deliverables

• Training environment user interface source code executable and documentation. • API specification for tracing from bots and agents the user interface to manage to trace output from different sources and its backend documentation.

Budget

$23,500 USD

Milestone 6 - Use cases

Description

Development use cases Training program A straightforward example of a training program that leads to development of machine understanding expressed in terms of stages. Training bots and their interoperation scenarios The training environment is a medium for training bots and training agents that may communicate with each other in order to complete the training program. The variations of bots and their interoperation scenarios which may deviate from the training program mainstream are needed to demonstrate the flexibility of the training system and the generalizing ability of the approach. Tests of bots and interoperation scenarios Tests are mostly bots which complement training bots as they communicate with the tested bots or take part in interoperation scenarios to diagnose and report to a human supervisor another bot or the training environment if the system works as expected. Default training bot for communication and supervision One or more training bots may provide a specific interface to human users or supervisors that can be used to “touch” machine understanding by interrupting the training programs with problems not from the training dataset to see the immediate response(s) from the agent(s) and evaluate them. Tracing scripts examples Tracing is a tool for developing effective training programs. It provides data for a quick view of the communication that takes place in the environment between the bots and the agents.

Deliverables

• Training program training bots trained agents for each stage of the training program supervision guidance. • Examples of training bots and interoperation scenarios with descriptions of expected effects. • Examples of bots interoperation scenarios and their tests and specification of the reporting interfaces. • The default training bot a guideline on how to see the development of machine understanding in the agent using the bot. • Tracing script examples with the description of their value for the development of the training program.

Budget

$21,500 USD

Milestone 7 - Onboarding

Description

Onboarding. Beta release

Deliverables

Beta release of platform

Budget

$6,000 USD

Join the Discussion (2)

Sort by

2 Comments
  • 0
    commentator-avatar
    Gombilla
    Jun 3, 2024 | 1:34 PM

    Hi there. Great job putting this up. I also understand that as this platform aims to accommodate users of varying levels of expertise and understanding, ensuring that the training process remains accessible and comprehensible to all participants could be challenging. Additionally, the team should also be keen on managing the progression from simple to complex problems and ensuring that users receive adequate guidance and feedback throughout the training process may require significant resources and expertise. Good luck !

    • 0
      commentator-avatar
      Denis Shabratov
      Jun 4, 2024 | 12:58 PM

      Thank you for review. We're growing, and one part of our team are mathematicians who are focused on creating a useful process for creating of training program. And the aim of the project is to create tools to do that for everyone.We're interested in helping every user in this area.

Reviews & Rating

Sort by

7 ratings
  • 1
    user-icon
    Joseph Gastoni
    May 20, 2024 | 3:28 AM

    Overall

    4

    • Feasibility 4
    • Viability 4
    • Desirabilty 4
    • Usefulness 5
    APPARATUS SCIENTIA\'s Combinatorial ML platform

    This project proposes a platform for training AI models with "machine understanding" through a stage-by-stage approach. Here's a breakdown of its strengths and weaknesses:

    Feasibility:

    • Moderate-to-High: The core concept of stage-by-stage training seems feasible. However, building a user-friendly platform with all the proposed functionalities requires significant development effort.
    • Strengths: The modular design with training bots and a scripting language suggests potential for efficient development.
    • Weaknesses: The complexity of the proposed architecture and user interface might require more development time and resources than anticipated.

    Viability:

    • Moderate: Success depends on the effectiveness of the training approach, platform usability, and user adoption within the chosen target market.
    • Strengths: The focus on interpretable AI with "machine understanding" addresses a growing concern in the field.
    • Weaknesses: The project needs to demonstrate the effectiveness of the approach compared to existing methods and convince users of its value proposition.

    Desirability:

    • Moderate: For researchers, developers, and users interested in interpretable and reusable AI models, this can be desirable.
    • Strengths: The ability to build and understand AI models with clear reasoning holds significant appeal.
    • Weaknesses: The complexity of the platform and the nascent stage of development might deter some users.

    Usefulness:

    • Moderate (in proposal stage): The proposal outlines a novel approach to AI development with potential benefits for various fields.
    • Strengths: This project has the potential to improve the transparency and explainability of AI models, leading to broader trust and adoption.
    • Weaknesses: The long-term impact on user adoption, model effectiveness, and overall impact on the AI field needs evaluation.

    Additional Points:

    • Independent evaluation and comparison with existing interpretable AI solutions can strengthen the project's value proposition.
    • Developing a clear target audience and focusing on use cases that demonstrate the platform's advantages is crucial for user adoption.
    • Prioritizing core functionalities for an initial release and iteratively improving based on user feedback can be a good development strategy.

    APPARATUS SCIENTIA's Combinatorial ML platform has the potential to be valuable for the AI field. However, careful consideration of the development roadmap, target audience, and clear demonstration of the approach's effectiveness is required to achieve widespread adoption.

    Here are some strengths of this project:

    • Focuses on a critical challenge in AI: interpretability and reasoning behind model decisions.
    • Proposes a unique stage-by-stage training approach with the potential for modular and reusable models.
    • Highlights the platform's potential for broad user adoption by researchers, developers, and end-users.

    Here are some challenges to address:

    • Demonstrating the effectiveness of the approach compared to existing techniques through rigorous testing and benchmarks.
    • Balancing the platform's functionality with user-friendliness to avoid overwhelming potential users.
    • Convincing the target audience of the platform's unique value proposition and its contribution to their specific needs.

    user-icon
    Denis Shabratov
    Jun 4, 2024 | 12:44 PM
    Project Owner

    Thank you very much for a very informative and thorough review.
    Our general aim of the project is to make open and easy access for everyone for real technology to train AI with machine understanding and communication language development under developer supervision.
    Our technology completely focuses on industrial, engineering, and R&D problems and processes.
    We are committed to facilitating more real experiments with our AI agent and training method. We are ready to invest as much effort as needed to ensure everyone can effectively use our system, demonstrating our dedication to everyone's support.
    We work for several industrial partners, but we want to attract the community to build open technology with machine understanding.

  • 0
    user-icon
    CLEMENT
    Jun 3, 2024 | 1:40 PM

    Overall

    4

    • Feasibility 4
    • Viability 4
    • Desirabilty 4
    • Usefulness 4
    Democratizes access to AI experimentation/training

    Firstly, I would say that this project has the potential to make a significant impact on the development and democratization of AI understanding. Providing a platform where users can experiment with machine understanding and train AI models stage by stage, this project empowers individuals to contribute to the advancement of AI technology and understanding. In a long run, this will foster a collaborative environment where users can learn from each other, share insights, and collectively tackle complex technical and scientific problems.

    Kudos to the team !

    user-icon
    Denis Shabratov
    Jun 4, 2024 | 12:53 PM
    Project Owner

    Thank you!
    We are very open and interested in working with everyone to solve their challenges and develop their technology using intelligence under the control of developers in industry and science. 

  • 0
    user-icon
    Nicolad2008
    Jun 7, 2024 | 6:45 AM

    Overall

    4

    • Feasibility 5
    • Viability 4
    • Desirabilty 4
    • Usefulness 4
    Understanding of machinery and application

    Perhaps the permission to build a model from limited data to solve complex technical and scientific problems of the project is appropriate. This open platform facilitates the research, inspection and use of ML with understanding of machines, opening new business models and markets. However, this project cannot avoid the limitations of developing a complete natural language to describe the characteristics of the problem, methods and solutions that require a lot of time and effort, which can slow down progress and increase costs. Moreover, the monitoring of the problem solving process because AI has been trained to be highly accurate, this poses a great technical and safety challenge. Applying ML with understanding of machines in reality also needs to be tested and carefully evaluated, requiring a costly and complicated verification process. Therefore, although the development potential of the project is very large, these challenges need to be resolved to ensure the feasibility and long -term effectiveness.

    user-icon
    Denis Shabratov
    Jun 10, 2024 | 9:21 AM
    Project Owner

    Thank you for your comment!
The problems of human use of AI are indeed a crucial challenge both for developers of new ML methods and those who apply them in their technologies and businesses.
    At the presentation meeting, we were asked only one question - do you pretend to create General AI?

    Our answer to this question is no. Because our aim is to enhance human capabilities, not to replace humans.
Understanding is the basis of what people call intelligence.
    We propose methods for human control and driven creation of new solutions based on automation of the using of known knowledge, methods, and solutions.
Any real controlled creation is impossible without understanding. This is a challenge for any team of people, not just with a machine.
    We aim to create and master methods of forming understanding and knowledge that appears as a result of understanding.
    It seems to us that this is the future of AI, not in imitation of man in texts, drawings, and music, something in which there are geniuses and creators among us people.

  • 0
    user-icon
    BlackCoffee
    Jun 10, 2024 | 12:34 AM

    Overall

    4

    • Feasibility 4
    • Viability 3
    • Desirabilty 3
    • Usefulness 4
    The impact comes from democratization

    Democratizing knowledge & understanding of AI is what this proposal can do through impact. In fact, they are doing this well. Those who pay attention will realize that democracy is shown quite clearly. Democracy will help teams and communities develop together, most typically through completely natural interaction and support. I appreciate what the proposal does for the community in direct and indirect ways.

  • 1
    user-icon
    Tu Nguyen
    May 23, 2024 | 8:49 AM

    Overall

    4

    • Feasibility 5
    • Viability 4
    • Desirabilty 4
    • Usefulness 4
    Platform To Experiment With Machine Understanding

    This proposal aims to address the barriers to applying machine learning in most specific problem areas. They will build a publicly available platform that allows anyone to use a new machine learning approach based on training in stages from simple to complex, and create and test their models to solve problems. solve specific problems and problem domains based on the machine's understanding of the problem domain formed as a result of the training process. This is a positive solution. Another positive point is that they have determined the expected time to complete the milestones.
    Other opinion: information about members should be clearer. They should share members' social media links.

    user-icon
    Denis Shabratov
    Jun 4, 2024 | 12:50 PM
    Project Owner

    Thank you for the positive review. We are very keen to create an open platform for a new method that provides a new learning technique for technical problems on limited amounts of data and with the ability to generate insights
    Unfortunately, we really don't describe ourselves enough
    But are completely open to any communication! Write to us d.shabratov@combinatorial.ai

  • 0
    user-icon
    TrucTrixie
    Jun 9, 2024 | 1:52 PM

    Overall

    3

    • Feasibility 3
    • Viability 2
    • Desirabilty 2
    • Usefulness 3
    The usefulness is not really clear

    We see that easily recognizable usefulness is an approach to developing AI (this is a novel approach). From this approach, it will bring benefits to the community later and have the ability to expand the scope (this is my speculation but the team has not presented it clearly). In short, the usefulness is undeniable although the team needs more clarification.

    user-icon
    Denis Shabratov
    Jun 10, 2024 | 9:39 AM
    Project Owner

    Thanks for the feedback. Yes, the description of our team is currently somewhat complicated by the fact that the main members of the team, as described in the project, specialize primarily in mathematical methods used in machine learning. For various business reasons, unfortunately, we do not have a wide representation of our work.
However, we are expanding, and I think in the coming months, we will be able to look more presentable and attractive for project presentations and submissions. 
    From my point of view, it is not very clear how an incomplete description of the team can lower the feasibility and usefulness scores so much :)  In any case, we are well on our way to implementing our version of machine understanding technology and are keen to get it tested and applied with a wide range of developers who will be interested as soon as possible. 

  • 0
    user-icon
    Max1524
    Jun 8, 2024 | 1:38 PM

    Overall

    3

    • Feasibility 2
    • Viability 3
    • Desirabilty 3
    • Usefulness 3
    Please disclose the identities of all members

    Through an ML approach, building models uses limited data to create solutions to scientific and technical problems. This is a fairly new approach and what it creates is also a novelty for many people. I see it as a high-tech innovation. This is easy to explain because it is the product of basic machine learning and training algorithm developers & programmers. Therefore, the human resources team must truly be of high quality and full of intelligence. Through the presentation of the members, I have only seen the profiles of 2 members, the remaining 2 members are anonymous - not showing transparency in feasibility. I request that my full identity be made public.

    user-icon
    Denis Shabratov
    Jun 10, 2024 | 9:36 AM
    Project Owner

    Thanks for the feedback. Yes, the description of our team is currently somewhat complicated by the fact that the main members of the team, as described in the project, specialize primarily in mathematical methods used in machine learning. For various business reasons, unfortunately, we do not have a wide representation of our work.
However, we are expanding, and I think in the coming months, we will be able to look more presentable and attractive for project presentations and submissions. 
    From my point of view, it is not very clear how an incomplete description of the team can lower the feasibility and usefulness scores so much :)  In any case, we are well on our way to implementing our version of machine understanding technology and are keen to get it tested and applied with a wide range of developers who will be interested as soon as possible. 

Summary

Overall Community

3.7

from 7 reviews
  • 5
    0
  • 4
    5
  • 3
    2
  • 2
    0
  • 1
    0

Feasibility

3.9

from 7 reviews

Viability

3.4

from 7 reviews

Desirabilty

3.4

from 7 reviews

Usefulness

3.9

from 7 reviews

Get Involved

Contribute your talents by joining your dream team and project. Visit the job board at Freelance DAO for opportunites today!

View Job Board