What is Open Mind Common Sense?

by Stephen M. Walker II, Co-Founder / CEO

What is Open Mind Common Sense?

Open Mind Common Sense (OMCS) is an artificial intelligence project that was based at the Massachusetts Institute of Technology (MIT) Media Lab. The project was active from 1999 to 2016 and aimed to build and utilize a large commonsense knowledge base from the contributions of many thousands of people.

The project was initiated by Marvin Minsky, Push Singh, Catherine Havasi, and others, with the goal of harnessing the distributed human computing power of the Internet to collect common sense knowledge. The project has accumulated more than a million English facts from over 15,000 contributors.

OMCS's software is built on three interconnected representations: a natural language corpus that people interact with directly, a semantic network built from this corpus called ConceptNet, and a matrix-based representation of ConceptNet called AnalogySpace that can infer new knowledge.

The project also had a Brazilian initiative, named Open Mind Common Sense in Brazil (OMCS-Br), led by the Advanced Interaction Lab at Federal University of São Carlos (LIA-UFSCar). This project started in 2005, in collaboration with the Software Agents Group at the MIT Media Lab, with the main goal to collect common sense stated in Brazilian Portuguese.

The knowledge collected by Open Mind Common Sense has enabled research projects at MIT and elsewhere. The project was a first attempt at realizing the idea that the problem of constructing a system with common sense could be distributed. It was a kind of second-generation common sense database, learning much from Cyc, but aiming to do things in a different way.

The project faced the challenge of how to represent the knowledge it collected. It used a variety of methods to acquire commonsense knowledge from the general public over the web, including templates, word-sense disambiguation, methods of clarifying entered knowledge, and analogical inference to provide feedback.

The project's main goal was to teach computers enough about the everyday world so that they could reason about it like humans do. The project's website was designed to make it easy and fun for people to work together to give computers this knowledge.

How does Open Mind Common Sense (OMCS) work?

Open Mind Common Sense (OMCS) is an artificial intelligence project initiated at the Massachusetts Institute of Technology (MIT) Media Lab. The project's primary goal is to help computers understand the meanings of words that people use, essentially teaching them about the everyday world.

OMCS, also known as ConceptNet, was launched in 1999 and was the first crowd-sourced project used to train an AI. It has since grown to include knowledge from other crowd-sourced resources, expert-created resources, and games with a purpose.

The project works by collecting facts, rules, stories, and descriptions from the general public over the web. It uses natural language templates and employs word-sense disambiguation and methods of clarifying entered knowledge. It also engages in analogical inference to provide feedback and allows participants to validate knowledge.

The knowledge collected is transformed into a semantic network called ConceptNet, which is a graph whose edges, or assertions, express common sense relationships between two short phrases, known as concepts. The edges are labeled from a defined set of relations, such as IsA, HasA, or UsedFor, expressing what relationship holds between the concepts.

Another important component of OMCS is AnalogySpace, which uses data from the OMCS project and represents knowledge as a matrix of objects or concepts along one axis, and features of those objects along another, yielding a sparse matrix of very high dimension. This technique helps in reducing the dimensionality of common sense knowledge and can infer new knowledge.

The project has been successful in accumulating more than a million English facts from over 15,000 contributors. The knowledge collected by Open Mind Common Sense has enabled research projects at MIT and elsewhere.

What is the difference between ConceptNet and Open Mind Common Sense (OMCS)?

Open Mind Common Sense (OMCS) and ConceptNet are closely related, but they serve different purposes within the realm of artificial intelligence.

OMCS is an artificial intelligence project that was initiated at the Massachusetts Institute of Technology (MIT) Media Lab. The project's primary goal was to help computers understand the meanings of words that people use, essentially teaching them about the everyday world. OMCS worked by collecting facts, rules, stories, and descriptions from the general public over the web. The knowledge collected was transformed into a semantic network, which is where ConceptNet comes into play.

ConceptNet, on the other hand, is a semantic network that originated from the crowdsourcing project OMCS. It is designed to help computers understand the meanings of words that people use. ConceptNet is a graph whose edges, or assertions, express common sense relationships between two short phrases, known as concepts. The edges are labeled from a defined set of relations, such as IsA, HasA, or UsedFor, expressing what relationship holds between the concepts.

In essence, OMCS is the project that collected the data, and ConceptNet is the semantic network that was built from this data. ConceptNet has since grown to include knowledge from other crowdsourced resources, expert-created resources, and games with a purpose. It is used to create word embeddings -- representations of word meanings as vectors, similar to word2vec, GloVe, or fastText, but better.

So, while OMCS and ConceptNet are part of the same initiative to imbue artificial intelligence with a better understanding of human language and common sense, they represent different stages and components of this process. OMCS is the data collection project, and ConceptNet is the semantic network built from the data collected by OMCS.

More terms

Scaling Laws for Large Language Models

Scaling laws for Large Language Models (LLMs) refer to the relationship between the model's performance and the amount of resources used during training, such as the size of the model, the amount of data, and the amount of computation.

Read more

What is machine learning?

Machine learning is a subset of artificial intelligence (AI) that deals with the design and development of algorithms that can learn from and make predictions on data. These algorithms are able to automatically improve given more data.

Read more

It's time to build

Collaborate with your team on reliable Generative AI features.
Want expert guidance? Book a 1:1 onboarding session from your dashboard.

Start for free