Episodic and Semantic Agent Memory Quiz

Answers and explanations appear as you take the quiz.

1. What is the difference between episodic and semantic memory in an AI agent?
   • Episodic stores specific past events; semantic stores general knowledge
   • Episodic stores general facts; semantic stores specific events
   • They are two names for the same thing
   • Episodic is for images; semantic is for text
2. What limitation applies to a language model without an external memory or retrieval system?
   • It cannot process any conversation text
   • It can use information in its current context but does not retain arbitrary past interactions across sessions
   • It always remembers every previous user message
   • It can retrieve private data without a memory store
3. How are memories typically stored for fast retrieval?
   • As printed paper logs
   • Only inside the model's weights
   • As embeddings (vectors) in a vector database
   • As compressed video files
4. At query time, how does an agent find relevant memories?
   • It reads every memory in order every time
   • It picks memories at random
   • It only ever uses the newest memory
   • It embeds the current situation and finds nearest neighbors by similarity
5. What does 'consolidation' do for an agent's memory?
   • It rewrites clusters of repeated episodes into concise semantic facts
   • It deletes all memories nightly
   • It encrypts the database
   • It copies memory to a backup server
6. Which is an example of episodic rather than semantic memory?
   • The user is a vegetarian
   • On Tuesday the user said they prefer morning meetings
   • The user prefers TypeScript
   • The user lives in Canada
7. A stored note says “this user is a vegetarian.” Which memory type does that represent?
   • Episodic memory: a single time-stamped event
   • A model parameter update
   • Semantic memory: a generalized fact distilled from experience
   • A vector-search query
8. What information helps an episodic memory retain the context of a past interaction?
   • Only a similarity score
   • Only a model temperature
   • A list of unrelated future predictions
   • A timestamp and source or event context