Share this article
09. 10. 2023

Unlocking AI Insights: Key Takeaways from Working with Large Language Models

Author

Robert Eckstein, Machine Learning Engineer, and Martin Ludvík, Head of AI

The advancements in artificial intelligence, particularly with large language models (LLMs), have had a significant impact on various sectors. As Semantic Visions is in the process of exploring LLMs capabilities and integrating these models into our workflows, we have encountered challenges, surprises, and many learning opportunities. In this article, we will share key takeaways from our experiences, offering insights for both technical enthusiasts and businesses looking to leverage these AI tools. In our exploration, Semantic Vision’s machine learning team focused on five core areas using LLMs’ capabilities: Topic Classification, Named Entity Recognition (NER), Text Paraphrasing, Knowledge Graph Population, and Victim/Aggressor Role Labeling.

Key Takeaways

  • Topic Classification:
    • Large Language Model (LLM) fine-tuning is now an almost trivial task with OpenAI tools.
  • Name Entity Recognition:
    • While Large Language Models excel in many tasks, specialized models designed and optimized for specific applications can outperform them. Named Entity Recognition is one such example, where specialized models are, for the time being, better.
  • Text Paraphrasing Findings:
    • Text paraphrasing and summarization with GPT models is very impressive and already suitable for production. GPT 3.5 Davinci is slightly better than ChatGPT 3.5.
  • Knowledge Graph Population Findings:
    • Using LLMs, transforming unstructured data into knowledge graphs has become more feasible than ever.
  • Victim/Aggressor Role Labeling:
    • Fine-tuned Llama2 model with 13 billion parameters provided on average 10% better accuracy on the Victim/Aggressor task than the much larger ChatGPT 3.5 model, with 175 billion parameters.

——————————————————————————————————————

You can find the full version of the whitepaper here:

Ready to dive deeper into the insights and knowledge offered in our whitepaper? Discover the full version today for a comprehensive look at working with Large Language Models and expert analysis. Click the link below to access the complete whitepaper and stay ahead of the curve.

Download

Send download link to:

Read next

Leveraging Open Source Intelligence for Improved Supply Chain Risk Management
14. 02. 2024

Leveraging Open Source Intelligence for Improved Supply Chain Risk Management

Unlock the secrets to fortified supply chain resilience with our latest white paper, "Leveraging Open Source Intelligence for Improved Supply Chain Risk Management." Delve into the dynamic landscape of global supply chains, addressing regulatory compliance, Brexit, COVID-19, and geopolitical conflicts.

Read more
The Dark Underbelly of Audiovisual AI
12. 02. 2024

The Dark Underbelly of Audiovisual AI

The World Economic Forum’s Global Risks Report highlighted a significant immediate risk: the proliferation of false information through artificial intelligence (AI). Released before the annual Davos gathering, the report emphasized the challenges posed by AI-generated...

Read more
Analysis: The number of cyberattacks on Czech banks increased 5 times in the last year, with 131 incidents recorded last year
05. 02. 2024

Analysis: The number of cyberattacks on Czech banks increased 5 times in the last year, with 131 incidents recorded last year

The number of cyberattacks or serious threats directed at Czech banking institutions reached 131 last year. This is the highest number in history, compared to 26 in 2022, and it increased fivefold. This is mainly...

Read more
TOP 5 Media Highlights from Davos 2024
01. 02. 2024

TOP 5 Media Highlights from Davos 2024

The predominant focus in discussions at DAVOS revolved around the vulnerabilities within the global supply chain, a subject underscored prominently by all involved. In his address, the Chinese Premier highlighted trade barriers, emphasizing their pivotal...

Read more
Future-Proofing Financial Institutions: The Power of Early Warning System Solutions in Risk Management
24. 01. 2024

Future-Proofing Financial Institutions: The Power of Early Warning System Solutions in Risk Management

Early Warning Systems (EWS), as a cutting-edge technological solution poised to revolutionise traditional credit monitoring practices, Early Warning Systems serve as a crucial tool for identifying risks associated with biased or subjective decision-making, offering a...

Read more
Overview of the 3 leading publicly traded automakers by revenue in 2023
19. 01. 2024

Overview of the 3 leading publicly traded automakers by revenue in 2023

In this analysis, Semantic Visions provide a deep dive into these challenges in relation to the top three publicly traded automakers by revenue - Volkswagen, Toyota, and Ford, who faced considerable disruptions in 2023.

Read more
S&P 500 ESG Violations Spotlight on Alphabet, Amazon, and Meta in 2023
02. 01. 2024

S&P 500 ESG Violations Spotlight on Alphabet, Amazon, and Meta in 2023

In the current dynamic business world embedding Environmental, Social, and Governance (ESG) principles into company policies is increasingly critical. This shift reflects a growing awareness among investors, consumers, and regulatory bodies about the impact of...

Read more
AI in Healthcare: Successes, Pitfalls, and Lessons
18. 12. 2023

AI in Healthcare: Successes, Pitfalls, and Lessons

The COVID-19 pandemic has undeniably acted as a catalyst for the advancement and widespread adoption of digital technology and AI-based health solutions. The imperatives of remote care, swift diagnostics, and effective treatment strategies have created...

Read more

Ready to see it in action?

Request a demo