How does Aspecta's NLP solution transform public procurement?

Our collaboration with KInIT
KInIT is an independent and non-profit institute focused on research in intelligent technologies. It brings together and educates experts covering artificial intelligence and multiple fields of computer science, also connecting to other disciplines. Aspecta collaborates with KInIT in efforts aimed at automating text classification in public procurement applications. Through a series of experiments, we will explore various approaches to multilabel classification, including linguistic and statistical methods and the integration of large language models, such as SlovakBERT, which are known for their effectiveness in various NLP tasks. Based on the research results, the existing software application for public procurement requirements can be adapted to significantly increase its efficiency.

Challenges in the field of public procurement
Public procurement plays a key role in acquiring the goods and services necessary for the effective provision of public services. However, the traditional public procurement process is often accompanied by issues that create inefficiency and consume time for both government agencies and suppliers. Consequently, it is essential to find innovative solutions to these challenges.
Software support is used at various stages to streamline the procurement process. Adopting software tools with a user-friendly interface and intuitive use can significantly reduce time and minimize errors. Recommendation, suggestion, or validation components within these software applications can greatly assist users in providing consistent and accurate information, ultimately improving data quality.

The performance of natural language processing and language models
Natural Language Processing (NLP) and large language models (LLMs) are key technologies in the field of artificial intelligence. NLP is a branch of AI that involves the analysis, understanding, and processing of human language by a computer. On the other hand, LLMs are advanced machine learning models trained on massive amounts of text data, enabling them to generate coherent and meaningful text. The most well-known large language model today is undoubtedly ChatGPT by OpenAI.
Applications of NLP and LLMs in Public Procurement
Natural language processing (NLP) and large language models (LLMs) find broad applications across various fields, including automatic text classification, information extraction, machine translation, and text generation. In healthcare, these technologies can analyze medical records and identify patterns that aid in disease prediction. In finance, they can analyze financial reports and help forecast market trends. In the legal sector, NLP and LLM technologies can analyze legal documents and efficiently extract relevant information.
In the context of public procurement, these technologies offer the potential for automating and streamlining processes such as data categorization and analysis. By leveraging NLP and LLMs, governments can efficiently cluster and analyze public procurement data, leading to improved decision-making and resource allocation. Practical applications of NLP in public procurement can be applied to:

- Extracting relevant information, identifying key terms, and understanding context can assist in the automation of the project review and evaluation process.
- Analyzing supplier information (their qualifications, past performance, and references) can assist in the automation of supplier capability assessment and comparison of supplier proposals.
- Analyzing public procurement contracts, extracting essential provisions, terms, and conditions can assist in contract management and monitoring contract compliance.
- Analyzing large volumes of data, such as financial reports and public records, can help identify issues and refine the assessment of supplier integrity and financial stability.
- Advanced search features enable more efficient retrieval of relevant information, the development of recommendation systems based on historical data and user preferences, suggesting suitable suppliers, contract clauses, or procurement strategies.

Aspecta's NLP Solution: Enhancing Procurement Efficiency
Aspecta’s solution leverages NLP technology to categorize, cluster, and analyze public procurement data with greater efficiency. By harnessing the power of NLP, the solution can automate and streamline various aspects of the procurement process, including tasks such as data categorization, supplier evaluation, contract analysis, and risk assessment. By reducing the time spent on these tasks and minimizing errors, the solution enables government agencies and suppliers to work significantly more efficiently and effectively.
The joint project of Aspecta and KInIT demonstrates that modern technologies and innovations can bring significant advantages across various processes. The project is an example of the practical implementation of theoretical advancements and highlights the opportunities that artificial intelligence and innovation offer in public procurement.

Conclusion
Aspecta is a company driven by the power of innovation and digital transformation. Implementing NLP technology in the field of public procurement and leveraging our expertise in artificial intelligence align with our company's commitment—at Aspecta, we recognize the value of staying at the forefront of technological advancement and using it to achieve tangible results for our clients and society. By utilizing AI and NLP, we aim to optimize public procurement processes, achieve cost savings, and improve service delivery.
The collaboration between Aspecta and KInIT brings a groundbreaking solution to public procurement. By harnessing the power of natural language processing (Natural Language Processing – NLP) and large language models (Large Language Models – LLMs), governments can revolutionize the way public procurement data is categorized, analyzed, and utilized.