Idiom Savant You can find a colorful description for almost anything in the texts of the world-wide-web. Of course, you can also find a great deal of nonsense and irrelevance too.Idiom Savant is a linguistic magnet for finding the sharpest needles in the haystacks of the internet. Enter a term of interest, such as priest or politician or critic or movie and Idiom Savant will show you two lists of resonant descriptions, divided acording to the perceived affect of the words employed: one list for positive descriptions, and another for negative descriptions. A sensible measure of pragmatic comparability (not just semantic similarity) is used to find the most comparable terms for your input, and to show you the positive and negative descriptions pertaining to those other terms also. For instance, enter the term critic and you will find apt descriptions for judge and monster also.
Dorian: Analogical Portraiture
Dorian is a knowledge-base that explores this multiplicity of categorization when dealing with proper-named entities. Dorian’s knowledge-base of proper-named entities is harvested from the Google n-grams, and associates entities with the categories that speakers most commonly attribute to them. Dorian’s knowledge-base is supplemented by the category-system in Wikipedia, which adopts a less subjective, curated approach to categorization.
The Lex-Ecologist Ecologists study the natural environment of plants and animals. Our plants and animals are words and concepts, and our environments are large text corpora. The Lex-Ecologist allows you to explore the rich textual environment for words provided by the text of the on-line encyclopaedia Wikipedia. Observe the behaviour of concepts in this environment: observe what they do, what is done to them, what they act upon, and how they congregrate into groups.
Jigsaw BardThe Jigsaw Bard is an online application that allows you to find resonant phrases for a large range of simple properties, like quiet, or for an even larger range of complex blended properties, like quiet and calm. The Bard has already scoured vast amounts of web text to identify phrases that have a resonant quality, and has automatically indexed these phrases on the properties they most poetically suggest. Most phrases are found art in this respect -- they are well-formed fragments of English that the Bard thinks have both a poetic quality and a useful communicative function. But some phrases (shown in blue) have been directly composed by the Bard itself. Have a look and see what you think of the Bard’s compositional abilities.
Aristotle: An Interactive Metaphor Finder
Let Aristotle help you find appropriate metaphors to describe a given person or thing. Simply enter the target for your metaphor (called the tenor in metaphor research), choose a property you would like to accentuate, and Aristotle will select a range of possible vehicles to carry this meaning. Click on any of these vehicles to understand the full import of the metaphor you are about to use.
SardonicusSardonicus is a simile-finder that knows the exemplary properties of different objects in the real world. It has acquired this knowledge by sifting the contents of the web in search of meaningful comparisons. It knows that ninjas are stealthy and that bowling balls are heavy and smooth enough to be called bald. It also has a healthy sense of irony, so it knows that roller-coasters are not exactly a model of consistency, and that turtles are not generally prized for their speed. The similes in Sardonicus are divided into straight-faced "factual" similes and tongue-in-cheek "ironic" similes, and are organized hierarchically using a taxonomy of adjectives. Try it now, it might put an ironic smile on your face.
Mondrian: Mapping of Names, Descriptions and Roles in Analogy
Mondrian is a knowledge-base of commonplace associations that have been mined from the Google n-grams database of frequent web-content.
Mondrian views the world as a collection of triples, of the form Subject-Relation-Object.
You can query Mondrian to see what triples have a given Subject, Relation or Object (or any combination of these).
For instance, put Rabbi in the Subject field and Mondrian will give you all its triples in which Rabbi is the subject.
When you click on a relation, Mondrian shows you analogies for this relation. Mondrian uses the squaring rule to detect analogies: S1-R-O1 is analogical to S2-R-O2 if Mondrian thinks that S1-like-S2 and O1-like-O2. Hence, Mondrian builds squaring relations between parallel triples.
Click on a column of the relations table to see the table re-sorted with that column as a key.
Mondrian will also show you the common compounds that a subject engages in.
Click here to see Mondrian in action.
DimSum
Different languages tend to represent different cultural and conceptual perspectives on the world. To the originating culture, such lexicalized perspectives may seem entirely conventional and stale, but to another they may well provide fresh and even innovative insights into the meaning and creative uses of words. DimSum aims to mine these insights from the lexical structure of Chinese, a logomorphemic language that exhibits its semantic structure quite openly in its orthographic realization.
ZeitGeistLanguage is a dynamic landscape in which words are not fixed landmarks, but unstable signposts that switch directions as archaic senses are lost and new, more topical senses, are gained. Frequently, entirely new lexical signposts are added as newly minted word-forms enter the language. One can experience the variety and inventiveness of the most creative new words in English with ZeitGeist, a creative neologism generator.
The Analogical Thesaurus The Analogical Thesaurus is an attempt to create a semantic index of words and ideas that can be accessed using a variety of conceptual perspectives. Besides the traditional alphabetic index (as in conventional dictionaries) and a taxonomic index (e.g., as found in WordNet), the Analogical Thesaurus provides an analogical and metonymic means of browsing through words and ideas.
As presented here, the Analogical Thesaurus has been distilled from a marriage of two resources. The first is HowNet lexical ontology, using analogical techniques described in our group publications. The second is Wikipedia, an on-line open-source encyclopedia, whose rich inter-topic reference structure allows us to extract implicit relationships from HowNet entries and their Chinese orthography.
You can access the Analogical Thesaurus on-line. Be sure to read the help documentation first.
Thesaurus RexConcept taxonomies offer a powerful means for organizing knowledge, but this organization must allow for many overlapping and fine-grained perspectives if a general-purpose taxonomy is to reflect concepts as they are actually employed and reasoned about in everyday usage. Think about a term like "tofu" or "robot" -- besides the basic meanings of Food or Robot, these terms evoke different perspectives and categorizations in different contexts and for different people. For robust natural language understanding (NLU), we require our lexicons/ontologies to appreciate and exploit all of these perspectives. Thesaurus Rex acquires finely-discriminating taxonomies from the a variety of different starting points, or seeds, via a process of bootstrapping from three different sources: WordNet, ConceptNet and the web at large.
KNOW-BEST: Knowledge-Based Entertainment and Scholastic Testing
The KNOW-BEST Project is predicated on the belief that lexical resources like WordNet and HowNet are sufficiently rich to provide a knowledge-bsed creative language system for a new generation of computer games. Great strides have been made in producing computer games that are visually realistic and compelling, and an increasing demand for this graphical richness in turn has lead to the development of graphics-engines or physics-engines to make games even more physically immersive.
However, comparatively little academic and commercial interest has been expended on the notion of a language-engine for computer games, a creative knowledge-based layer that can provide linguistic intelligence across a variety of games.
We are looking to WordNet and HowNet for the generative means to imbue games with dynamic linguistic content, in the form of language-based puzzles and on-the-fly textual content.
This dynamic content has, additionally, a scholastic application. If games can be imbued with challenging language puzzles, this content can also be used to test and evaluate students.
Play a demo version of this game
KNOW-BEST is generously funded by an Enterprise Ireland Commercialization-Fund grant. All game designs and implementations are copyright Tony Veale and the Creative Language Systems Group.
CWFB Puzzle Questionnaire
The CWFB Puzzle Questionnaire investigates and evaluates the correctness of solutions to automatically-generated puzzles. These puzzles are generated using the CIA World Factbook (or CWFB), but the perceived difficulty of these puzzles will of course depend on the subject’s prior knowledge of geography. Any interested person can extend his/her horizon in the geopolitical arena by solving a variety of interesting puzzles provided here. In order to model personal puzzle difficulty, we ask you to describe your familiarity in different geographical domains (e.g., Europe, Asia) first.