The FIBEP Tech Day 2023 conference brought together experts from the media intelligence industry to discuss the latest trends, technologies, and innovations impacting the sector. The event covered a wide range of topics, reflecting the diverse interests and expertise of the speakers and attendees.

The conference emphasized the importance of collaboration and innovation in the media intelligence industry, with a particular focus on the potential of emerging technologies such as ChatGPT, natural language generation (NLG), and large language models (LLMs) to transform the field.

NLG in the light of ChatGPT

Nesin Veli, Product Manager at Identrics, delivered a presentation on the topic of NLG in the light of ChatGPT.

NLG is a technology that uses natural language processing (NLP) to create content from data. It can be used to generate reports and other types of documents, as well as automate tasks like answering questions about your business or creating email templates for customer service agents. The use of NLG in media intelligence has been around since the early 2000s, when search engines first started indexing news sites on the web. 

Learn more about the best sources of training data for large language models.

Nesin Veli’s presentation delved into the evolution of NLG, highlighting its transformation from simple rule-based systems to advanced deep learning models thanks to neural networks and transfer learning techniques. This progress has led to cutting-edge language models like GPT-3 and ChatGPT, impacting industries such as media intelligence.

NLG’s applications span journalism, marketing, customer support, and content creation, and its use in media intelligence includes automated summarisation, sentiment analysis, topic classification, and insights generation. There are numerous benefits of using NLG, including  increased efficiency, cost savings, and real-time insights.

However, relying on technologies like ChatGPT comes with risks and drawbacks, such as data privacy concerns, biases, and the potential for misleading information.

Therefore, in his presentation, Nesin provided an overview of the training process for language and conversational models, focusing on the importance of pre-training and fine-tuning.

Custom language models tailored to specific tasks and incorporating domain knowledge can provide more accurate insights in media intelligence. These models can also address data privacy challenges and mitigate biases in pre-trained models. Nesin Veli emphasised the critical role of high-quality, diverse, and well-labeled data in building custom language models.

Explore the training data for language model solution by Identrics.

In conclusion

The FIBEP Tech Day 2023 conference provided valuable insights into the latest trends and technologies in the media intelligence industry. As the industry continues to evolve, NLG and other emerging technologies will play an increasingly important role in driving innovation and delivering value to clients.

Contact us today to discuss how our AI solutions can support your business projects.