---
title: "Frame vs semantic net"
subject: "Artificial Intelligence"
chapter: 5
key: "semester_6/artificial_intelligence:frame-vs-semantic-net"
source: https://www.bctengineer.com/q/semester_6/artificial_intelligence/frame-vs-semantic-net
site: BCT Engineering
---

# Frame vs semantic net

Artificial Intelligence · Chapter 5: Structured Knowledge Representation

- Asked in: 2074 Bhadra, 7 marks
- Video: [Semantic nets and Frames in AI](https://www.youtube.com/watch?v=LZQa5puHwmE) — Dr. Kavita Oza
- Answer checked by Claude Opus 5.5 (high effort) on 2026-10-14
- Page: https://www.bctengineer.com/q/semester_6/artificial_intelligence/frame-vs-semantic-net
- Whole chapter: https://www.bctengineer.com/md/semester_6/artificial_intelligence/chapter-5

## Question

What is a Frame? How is it different from semantic net in knowledge representation?

## Short answer

A frame is a record-like structure with slots, defaults, constraints, and attached procedures, organized in is-a hierarchy. A semantic net is a pure graph of nodes and arcs. Frames have internal node structure, defaults, and demons; semantic nets lack these. Both support inheritance.

## Answer

A **frame** is a structured representation of an object or situation: a named collection of slots, each with a value, default, restriction or attached procedure, arranged in an `is-a` / `instance-of` hierarchy with inheritance (Minsky).

Example:
```text
Frame: Car                          Frame: MyCar
  is-a: Vehicle                       instance-of: Car
  wheels: 4 (default)                 colour: Red
  fuel: [petrol or diesel]            owner: Hari
  mileage: [if-needed: km / litres]
```

A **semantic net** is a graph of nodes and labelled arcs, e.g. MyCar --instance--> Car --is-a--> Vehicle, MyCar --colour--> Red.

#### Difference between frames and semantic nets
| Point | Frame | Semantic net |
|---|---|---|
| Structure | Record-like: object with slots and fillers | Graph: nodes joined by labelled arcs |
| Node content | Node has internal structure (slots, facets) | Node is just a label |
| Defaults | Explicit default values for slots | No built-in defaults |
| Procedures | Attached procedures (if-needed, if-added demons) | None; purely declarative |
| Constraints | Type and range restrictions on slots | Not supported directly |
| Visual form | Table-like | Diagram of links |
| Best for | Complex objects with many attributes | Simple relations and taxonomies |
| Inference | Inheritance plus procedures and expectations | Inheritance by following links |
| Complexity | Heavier to design | Simple and easy to draw |

Both support inheritance, and a frame system can be seen as a semantic net whose nodes are given internal structure.

## Check yourself

The answer's key points as multiple-choice questions, with the correct option and why.

1. What is the main structural difference between a frame and a semantic-net node?
   A. A frame is a graph; a net is a table
   B. A frame node has internal structure (slots and facets); a net node is only a label
   C. A net node holds procedures; a frame does not
   D. There is no difference
   Answer: B. A frame system is like a semantic net whose nodes are given internal structure. (Key point: Node content)

2. Which feature do frames have that semantic nets lack?
   A. Attached procedures (if-needed, if-added demons) and defaults
   B. Inheritance along is-a links
   C. Representing objects and relations
   D. Graph visual form
   Answer: A. Nets are purely declarative with no built-in defaults or demons. (Key point: Procedures)

3. Which is each representation best for?
   A. Frames: simple taxonomies; nets: objects with many attributes
   B. Both only for numerical calculation
   C. Frames: probabilistic reasoning; nets: logic proofs
   D. Frames: complex objects with many attributes; nets: simple relations and taxonomies
   Answer: D. Nets are simple to draw; frames are heavier to design but richer. (Key point: When to use)

## Related in Artificial Intelligence

- [Frames, semantic net; sakti gauchan](https://www.bctengineer.com/question.md?key=semester_6%2Fartificial_intelligence%3Aframes-semantic-net-convert-sentences) — Artificial Intelligence, chapter 5
- [KR models; semantic nets](https://www.bctengineer.com/question.md?key=semester_6%2Fartificial_intelligence%3Akr-models-semantic-nets) — Artificial Intelligence, chapter 5
- [Semantic nets with example](https://www.bctengineer.com/question.md?key=semester_6%2Fartificial_intelligence%3Asemantic-net-with-example) — Artificial Intelligence, chapter 5, asked 3×
- [Declarative vs procedural knowledge](https://www.bctengineer.com/question.md?key=semester_6%2Fartificial_intelligence%3Adeclarative-vs-procedural-knowledge) — Artificial Intelligence, chapter 4, asked 4×
- [Knowledge vs learning, acting humanly](https://www.bctengineer.com/question.md?key=semester_6%2Fartificial_intelligence%3Aknowledge-vs-learning-acting-humanly) — Artificial Intelligence, chapter 1
- [Knowledge, reasoning; forward chaining](https://www.bctengineer.com/question.md?key=semester_6%2Fartificial_intelligence%3Aforward-chaining-kr-reasoning) — Artificial Intelligence, chapter 4

## Related in other subjects

- [Knowledge management vs expert systems](https://www.bctengineer.com/question.md?key=semester_8%2Finformation_system%3Akm-vs-expert-systems) — Information System, chapter 4
- [Array representation vs dynamic list](https://www.bctengineer.com/question.md?key=semester_4%2Fdata_structure_and_algorithm%3Aarray-vs-dynamic-list) — Data Structure and Algorithm, chapter 3
- [Data representation; analog vs digital system](https://www.bctengineer.com/question.md?key=semester_5%2Fdata_communication%3Adata-representation-analog-digital) — Data Communication, chapter 1
- [TDM vs statistical TDM and STDM frame](https://www.bctengineer.com/question.md?key=semester_8%2Ftelecommunications%3Atdm-statistical-tdm) — Telecommunications, chapter 3, asked 2×