Digital Intuition
How to give computers human-like intuition so they can better understand us.

Language is central to how people reason about and understand their world. As computers increasingly pervade human lives and decision-making processes, they must learn to understand and mediate human-to-human interaction. People use their intuitive knowledge of the world and the experiences they've had in the past to react intelligently to the world around them. If we were to give machines these capabilities, they could help us make better-informed decisions, conquer mountains of data, and expand the reach of our creativity and intelligence.

Research Projects

  • CharmMe

    Catherine Havasi, Brett Samuel Lazarus and Victor J Wang

    CharmMe is a mobile social discovery application that helps people meet each other during events. The application blends physical and digital proximity to help you connect with with other like-minded individuals. Armed with RFID sensors and a model of how the Lab works, CharmMe determines who you should talk to using information including checking in to conference talks or “liking” projects using QR codes. In addition, possible opening topics of conversation are suggested based on users' expressed similar interests.

    view site

  • ConceptNet

    Catherine Havasi, Robert Speer, Henry Lieberman and Marvin Minsky
    Imparting commonsense knowledge to computers enables a new class of intelligent applications better equipped to make sense of the everyday world and assist people with everyday tasks. Our approach to this problem is ConceptNet, a freely available commonsense knowledge base that possesses a great breadth of general knowledge that computers should already know, ready to be incorporated into applications. ConceptNet 5 is a semantic network with millions of nodes and edges, built from a variety of interlinked resources, both crowd-sourced and expert-created, including the Open Mind Common Sense corpus, WordNet, Wikipedia, and OpenCyc. It contains information in many languages including English, Chinese, Japanese, Dutch, and Portuguese, resulting from a collaboration of research projects around the world. In this newest version of ConceptNet, we aim to automatically assess the reliability of its data when it is collected from variously reliable sources and processes.
    view site

  • Corona

    Rob Speer and Catherine Havasi

    How can a knowledge base learn from the Internet, when you shouldn't trust everything you read on the Internet? CORONA is a system for building a knowledge base from a combination of reliable and unreliable sources, including crowd-sourced contributions, expert knowledge, Games with a Purpose, automatic machine reading, and even knowledge that is imperfectly derived from other knowledge in the system. It confirms knowledge as reliable as more sources confirm it or unreliable when sources disagree, and then by running the system in reverse it can discover which knowledge sources are the most trustworthy.

  • Divisi: Reasoning Over Semantic Relationships

    Robert Speer, Catherine Havasi, Kenneth Arnold, and Jason Alonso
    We have developed technology that enables easy analysis of semantic data, blended in various ways with common-sense world knowledge. The results support reasoning by analogy and association. A packaged library of code is being made available to all sponsors.

  • GI Mobile

    Catherine Havasi and Brett Lazarus

    GI Mobile is a mobile companion to the Media Lab Glass Infrastructure system. It incorporates the MessageMe messaging system to deliver a suite of location-aware features that complement the Glass Infrastructure. These include locating others in the Lab, browsing projects physically near you, and sending location-based messages. In addition, GI Mobile will alert you when you pass by projects you may be interested in based on what projects you have "liked."

  • MessageMe

    Brett Lazarus

    MessageMe is a location-based messaging infrastructure. It consists of a messaging server that delivers messages to recipients as they enter a desired physical space in the Lab. MessageMe builds on the Glass Infrastructure system, utilizing the RFID readers at each screen to determine users' locations.

  • Narratarium

    V. Michael Bove Jr., Catherine Havasi, Katherine (Kasia) Hayden, Daniel Novy, Jie Qi and Robert H. Speer

    Remember telling scary stories in the dark with flashlights? Narratarium is an immersive storytelling environment to augment creative play using texture, color, and image. We are using natural language processing to listen to and understand stories being told, and thematically augment the environment using color and images. As a child tells stories about a jungle, the room is filled with greens and browns and foliage comes into view. A traveling parent can tell a story to a child and fill to room with images, color, and presence.

    view site

  • Open Mind Common Sense

    Marvin Minsky, Robert Speer, Catherine Havasi
    The biggest problem facing artificial intelligence today is how to teach computers enough about the everyday world so that they can reason about it like we do—so that they can develop "common sense." We think this problem may be solved by harnessing the knowledge of people on the Internet, and we have built a Web site to make it easy and fun for people to work together to give computers the millions of little pieces of ordinary knowledge that constitute "common sense." Teaching computers how to describe and reason about the world will give us exactly the technology we need to take the Internet to the next level, from a giant repository of Web pages to a new state where it can think about all the knowledge it contains; in essence, to make it a living entity.
    view site

  • Red Fish, Blue Fish

    Robert Speer and Catherine Havasi

    With commonsense computing, we can discover trends in the topics that people are talking about right now. Red Fish Blue Fish takes input in real time from lots of political blogs, and creates a visualization of what topics are being discussed by the left and the right.

  • Second-Language Learning Using Games with a Purpose

    Catherine Havasi and Kasia Hayden

    An online language learning tool and game with a purpose (GWAP) designed to simultaneously gather annotated speech and text data useful for improving natural language processing (NLP) applications and serve as an English-language learning resource.

  • Semantic Synesthesia

    Catherine Havasi, Jason Alonso and Robert H. Speer

    Semantic Synesthesia is a program that guesses a color to represent a given input word or sentence, taking into account both physical descriptions of objects and emotional connotations. This novel application of artificial intelligence uses knowledge about the world to build a model of how people think about objects, emotions, and colors, and uses this model to guess an appropriate color for a word. Colorizer works over static text and real-time input, such as a speech recognition stream. It has applications in games, arts, and Web page design.

  • Story Space

    Catherine Havasi and Michael Luis Puncel

    Analogy Space, a previous project under the Digital Intuition group, developed a technique of plotting concepts in a many-dimensional semantic space in order to identify clusters of concepts that are similar to each other. Story Space will apply this technique to human narrative in order to provide a measure of similarity between different stories. It has had preliminary success using datasets that are easily broken up into discrete events, such as "how-to" articles from the internet. The next steps involve using automatic event taggers to determine the progression of a story.

  • The Glass Infrastructure

    Henry Holtzman, Andy Lippman, Matthew Blackshaw, Jon Ferguson, Catherine Havasi, Julia Ma, Daniel Schultz and Polychronis Ypodimatopoulos

    This project builds a social, place-based information window into the Media Lab using 30 touch-sensitive screens strategically placed throughout the physical complex and at sponsor sites. The idea is get people to talk among themselves about the work that they jointly explore in a public place. We present Lab projects as dynamically connected sets of "charms" that visitors can save, trade, and explore. The GI demonstrates a framework for an open, integrated IT system and shows new uses for it.

    view site

  • Understanding Dialogue

    Catherine Havasi, Anjali Muralidhar and Personal Robots Group

    In order to extend the Digital Intuition group's ability to understand human language, a module that fills in the gaps of current technology must be developed to understand dialogue. This module will be based on data from the Switchboard human-human dialogue corpus, as well as a dataset of recorded dialogues between parents and children while reading an interactive e-book created by the Lab's Personal Robots group. The goal is for the module to be able to identify the emotion and mood of the dialogue in order to make inferences about what parents and children generally talk about when reading the book, and to make suggestions about additional conversation topics. Conversations between an adult and child while reading a book can greatly contribute to the learning and development of young children.