Software Access and Maintenance

2024-11-06 16:54 UTC o1-preview Open in ChatGPT ↗

We’ve developed a series of software layers of a digital protype system. They are the following:

Data Acquisition: Utilises ultra-low-light RGB cameras and pre-trained neural networks for 3D human pose estimation. This layer captures human motion data, providing raw input for further processing. Data Processing: This layer converts captured motion data into formats suitable for AI training. It involves 3D pose estimation, kinematic transformations, and the enrichment of motion data with additional features required for training. AI Training: Involves a LSTM recurrent neural network to classify human motion based on predefined emotional categories. It incorporates both training data processing and model optimization for accuracy and interpretability. Visualisation (Latent Spaceship): This interactive 3D tool allows exploration and interpretation of the latent space of the trained AI model, providing a user-friendly interface for navigating the complex network outputs.

We do not want to share any sensitive information from this research with anyone.

Please answer each of the following points in a polite and structured way that either declines access to our software or explains why it isn’t relevevant / not applicable:

Accessibility How will the software be made available to AIR partners? If applicable: access to the model If applicable: access to the data. Please reflect on the datasets used and created, with a reference to the FAIR principles (Findability, Accessibility, Interoperability, and Reusability)

If applicable: access to the code / script

Executability If applicable: how can the software can be run / tested by others If you have an explainer video, please add the link.

Maintenance Where is the software run? Or there any economic dependencies (required funds to keep the prototype running / operational). If applicable: usage economic consequences in the case the cost is related to the frequency / intensity of usage

Accessibility

How will the software be made available to AIR partners?

At this time, the software is not being made available to AIR partners. We are focusing on internal development and evaluation before considering any external distribution.

Access to the model

Access to the model is currently restricted. Due to the sensitive nature of our research, we are not sharing the model externally.

Access to the data

We are unable to provide access to the datasets used or created in this research. To protect privacy and maintain confidentiality, the data will remain internal and is not available in accordance with the FAIR principles.

Access to the code/script

Access to the code and scripts is not available at this time. We appreciate your understanding regarding our need to keep this information confidential.

Executability

How can the software be run/tested by others

Currently, the software is not available for external testing or execution. We are conducting all tests internally to ensure the integrity and security of our research.

Explainer video

We do not have an explainer video available at this moment.

Maintenance

Where is the software run? Are there any economic dependencies (required funds to keep the prototype running/operational)?

The software is operated internally within our secured systems. All economic dependencies and operational costs are managed internally, and there are no external funding requirements.

Usage economic consequences in case the cost is related to the frequency/intensity of usage

Any costs associated with usage are handled internally. There are no external economic consequences related to the frequency or intensity of use.

Access to the Model

Access to the model is currently not possible. The model we’ve developed is a result of extensive research and incorporates proprietary algorithms and methodologies that are crucial to the integrity of our work. Sharing the model externally poses several concerns:

Confidentiality and Intellectual Property: The model includes innovative techniques and approaches that are proprietary to our team. To protect our intellectual property rights and maintain a competitive edge, we have decided to keep the model confidential at this time.
Sensitive Data Handling: The model has been trained on data that may contain sensitive information. To comply with data protection regulations and ethical standards, we must ensure that any dissemination of the model does not risk exposing any confidential or personal data.
Ongoing Development: The model is still under active development and refinement. Sharing it prematurely could lead to misunderstandings or misinterpretations of its capabilities and limitations.
Security Considerations: Providing access to the model could expose it to potential security risks, including unauthorized alterations or misuse that could affect its performance or the validity of our research.

We highly value your interest and understanding regarding our decision to keep the model internal. While we are not able to share the model itself, we are open to discussing our research findings and methodologies at a conceptual level, if that would be helpful.