september 1, 2022
@
8:00 am
–
september 30, 2023
@
8:00 am
CEST
(PROJECT)
The VentureNet data exchange project was a research project completed in 2023 in collaboration with SINTEF and the Norwegian Centre for Research and Innovation in Cybersecurity in Critical Sectors (SFI NORCICS) based at the Norwegian University of Science and Technology (NTNU). The project received co-financing from the Research Council of Norway and Innlandet County through the FORREGION program.
Background
A major EU industrial research project on data harmonization, euBusinessGraph, identified major challenges with access to business data, citing that “it is extremely expensive, time consuming and error prone to find, interpret and reconcile”, especially across industries, geographic regions and languages.
Machine learning (ML) plays an important role in automating data processing tasks such as natural language processing (NLP), but human input is needed to train ML algorithms and thereby improve the accuracy of ML models. One of the challenges identified by data scientists is a lack of machine learning models with humans-in-the-loop (HITL). 80% of the cost of Big Data and Artificial Intelligence (AI) projects is time spent by data scientists finding, interpreting and harmonizing data. A lack of HITL limits the effectiveness of ML models and their potential to reduce cost in Big Data and AI projects.
VentureNet aims to create a virtual business accelerator hub powered by its proprietary Information Management System (IMS), which centralizes access to multiple digital ecosystems, facilitating data exchange and reducing manual data entry. Developed in collaboration with SINTEF and NTNU, the project focuses on creating a secure data exchange framework with Human-in-the-Loop (HITL) capabilities that can be scaled up in European data markets.
Project Results
The project delivered a report introducing a novel framework and new technical solution for data exchange that will facilitate dynamic harmonization and enrichment of business data with HITL ML models. The envisioned solution is a digital B2B workspace and marketplace that will automate data management tasks and processes and facilitate involvement of non-technical data providers (small businesses) in data enrichment with HITL models.
SINTEF compared VentureNet’s IMS to leading low-code platforms and AI solutions, such as Oracle APEX, AWS Amplify, Open AI GPT, Sheet GPT, and Google Natural Language API and concluded the following:
We compared VentureNet against a few low-code/no-code platforms. Low-code/no-code platforms significantly accelerate software development by simplifying and automating code generation, making it possible for non-developers to create applications, thereby democratizing development and reducing the technical skill barrier. However, these platforms can limit customization capabilities, restrict advanced functionality, and lead to vendor lock-in situations, potentially impeding scalability, and long-term growth for more complex or unique business requirements.
Potential for higher accuracy and filtering of hallucinations, using HITL, especially in sensitive domains, such as healthcare.
The principal advantage of the VentureNet platform lies in its comprehensive integration capabilities. It can fulfill all project requirements, including form processing, API management, low-code graphical user interface elements, and data enrichment. Unlike other generic platforms, VentureNet enables the development of highly customized solutions tailored to specific project needs.
Although it is possible to assemble similar platforms using excellent off-the-shelf tools, coordinating, and integrating these disparate components requires significant time and effort. Additionally, the complexities associated with billing processes can often pose substantial challenges. VentureNet solves these concerns by offering a unique solution that simplifies project management and reduces operational overhead. This approach ensures a streamlined workflow and efficient project execution.
Contact VentureNet to learn more about the project results and collaboration opportunities.
VentureNet
contact@venturenet.no
Norwegian 
English