Here you get a chance to try out the tools we've developed. All of them analyze sentences which you input, so be creative!

Question Answering Systems [ News Articles ] [ Web based ]

As the amount of information grows on the web, it becomes harder to find information. Research on Question Answering Systems aims at making the task of finding information easier. The goal is to replace current search technologies, which are based solely on key-word search, with the ability to process questions and find explicit answers for them.

Question Classification

Responding correctly to a free form question requires the computer to have an awareness of what is the question about, and to the constraints that the question imposes on a possible answer. For instance, the answer to a question like: "Who is the president of France" needs to be a name of a person. Accurately classifying potential answers sets the stage for later selecting the correct answer from among several candidates. See how it's done.

Shallow Parsing

Enabling a machine to respond to natural language input demands that the machine is equipped with the capacity to identify syntactical phrases in sentences. It is virtually impossible to manually write a comprehensive set of rules that accurately defines the appropriate solution to every task of this nature. However, the availability of annotated corpora (collections of text) and robust machine learning techniques make it possible to employ machines to learn this task from training examples.

Named Entity Recognition

Understanding text to the level that we can extract information from it in an intelligent way and answer questions with respect to it requires the ability to identify different types of entities and categories in text. E.g., this phrase represents a name of a person, an organization, a location and other semantic categories. This is a context sensitive problem ("Washington" is a location in one context and a person in another) and machine learning techniques are used to resolve this and determine the appropriate semantic category of entities.

Name Identification and Tracing

JFK, John F. Kennedy, John, Kennedy ?
U of I, Illinois, UIUC, UI,... ?
A given entity - representing a person, a location, or an organization - may be mentioned in text in multiple, ambiguous ways. Understanding natural language and supporting intelligent access to textual information require identifying whether different mentions of a name, within and across documents, represents the same entity. We demonstrate a browsing tool that incorporates some of our newly developed Machine Learning based technologies in this area. I enables users to trace different mentions of the same entity, presented in different textual forms, across documents.

Information Extraction

Information extraction is the technology that transfers human-readable documents into machine-readable data. Useful and important information can be extracted from lots of unorganized documents such as news articles and emails, and stored in databases. Then, it is relatively easy to get answers to the type of structured queries that ordinary search engines do not support. We demonstrate the technology by showing its ability to extracts specific phrases of interest in two types of documents --- seminar announcements and job postings.

Context Sensitive Spelling Correction [ Spelling Demo ] [ Preposition Demo ]

Spelling errors that result in valid words (e.g., replacing "peace" by "piece" "feel" by " fill", "desert by dessert", "buy by "by") cannot be caught by a conventional spell checkers, and account for some 25% of all spelling errors (and more for non-native speakers of English).

Context sensitive spelling correction is a machine learning based technology that we developed, which has been shown to be extremely effective in learning to correct these errors, performing with an accuracy level greater than 95%.

We present two demos of this technology. The first is a simple HTML-based spell checker. There is a form in which the user inputs text. The script will then suggest corrections for any errors it finds (it only detects errors resulting in valid words, which is the purpose of this demo; conventional errors are not detected).

The second demo concentrates on the problem of context sensitive mistakes in preposition use ("on", "over", "in") and is intended to help ESL (English as a Second Language) students.

Part of Speech Tagger:

The importance of assigning each word in a sentence the part of speech (POS) tag that it assumes in that sentence stems from the fact that identifying POS is one of the early stages in the process performed by various natural language related processes such as speech recognition, translation, and information retrieval and extraction. See how it's done!

Semantic Role Labeling

Beyond the syntactical analysis of natural language sentences is the extraction of its semantic information. Semantic role labeling is one of such task which identifies the verb and argument structure in natural language sentences, and is an important task toward natural language understanding.

Are the servers running?