Semantic Memory

Understanding Semantic Memory

Semantic memory is one of the types of memory an agent has, corresponding to non-salient, fact based memory. This can be any sort of information-based knowledge, such as PDFs, documents, news articles, or just plain text.

Using Semantic Memory

You can add any kind of information, whether plain text, local file path, or remote URL, using the SemanticMemory.ingest method.

The below examples assume the pre-requisite setup for a SwiftAgent

Setup

from swiftagent import SwiftAgent

agent = SwiftAgent()

1

Initialize Memory

python

from swiftagent.memory import SemanticMemory

memory = SemanticMemory()

2

Add Semantic Memory

python

agent.add_semantic_memory_section(memory)

3

Add information

python

memory.ingest("Sample text")

memory.ingest("https://arxiv.org/pdf/2304.04370"