DAT255 Course Project - Exploration of summarization techniques using machine learning

Summarization techniques using Large Language Models

llm

This blog post talks about our course project in DAT255 Deep Learning Engineering.

Nora Kristiansen og Torbjørn Vatnelid

Document Stuffing

The simplest form of summarization, where you simply “stuff” the document into your prompt. Works well for small amounts of text that fit into the model’s context window.

Nora Kristiansen, Torbjørn Vatnelid

Map Reduction

A good way to work around token limitations. If you are summarizing text that is too long for your chosen model’s context window, Map Reduction can be used.

Nora Kristiansen, Torjørn Vatnelid

KMeans Clustering

llm

When you reach a certain length of text to summarize, the other methods become too expensive or are not able to summarize well enough. We need to find a way to extract all…

Nora Kristiansen og Torbjørn Vatnelid

Presentation

code

analysis

llm

This is a post with executable code.

Nora Kristiansen, Torjørn Vatnelid