Summarising using Prompt Engineering

aansh savlaaansh savla
2 min read

Summarising means giving a concise overview of a lengthy topic. There is a lot of text available today and we don't have enough time to read everything. Hence summarising is a very hot research topic in the field of Artificial Intelligence. There are various applications of summarizing in R&D, Social Media Marketing, Legal Contract Analysis etc. A Large Language Model(LLM) like ChatGPT can be used for summarising huge text.

Let's see an application of summarising in financial research. Financial Analysts have to read market reports and annual reports of companies including balance sheets, cash flow statements, profit and loss statements, director's reports etc to make an investment decision. However, reading everything is merely impossible due to time constraints. Hence summarising can help analysts quickly derive the market signals from the content.

The example below shows a piece of text from the Director's report of Reliance Industries Ltd. for the year 2022. This piece of text has 2 sections:

1) Digital Services

2) Oil and Gas(Exploration and Production)

The text is just a part of the report but it still has much more content to be read by an analyst. So we try to summarise it.

The summary is much shorter and it's difficult for an analyst to make a decision based on just this small paragraph. So we just take one section and try to extract relevant information from it. Summarising can involve irrelevant information but extracting helps to get the information we are concerned with.

Now we have text only related to Digital Service but it is still lengthy. A financial analyst is much more concerned about numbers. Numerical data helps in making financial decisions fast. So we refine our prompt to extract numerical facts.

This gives a point-wise description of the 'Digital Service' section with numerical data at each point. A few points might not have numerical data but the model is intelligent enough to mention those points since it can be important for a financial analyst to know those points.

Similarly, we can extract numerical facts about the 'Oil and Gas' section. Thus summarising using prompt engineering can help financial analysts to convert big annual reports into a precise point-wise description of the company's performance.

0
Subscribe to my newsletter

Read articles from aansh savla directly inside your inbox. Subscribe to the newsletter, and don't miss out.

Written by

aansh savla
aansh savla