UNL Platform

Releases

	Status: release Date: 15 Jul 2014 Version: 1.2
The dEep-to-sUrface natural language GENErator. EUGENE is fully automatic. It takes a UNL input and delivers an output in natural language without any human intervention. Similarly to the UNLization tools, it is language-independent and has to be parametrized to the natural language input through a dictionary and a grammar, provided as separate interpretable files.
Go to EUGENE

	Status: release Date: 15 Jul 2014 Version: 1.2
IAN is the acronym for Interactive ANalyser. Differently from the UNLeditor, it includes a grammar for natural language analysis and operates semi-automatically. The word sense disambiguation is still carried out by the language specialist, but the system can filter the candidates using an optional set of disambiguation rules. The syntactic processing is done automatically through the natural language analysis grammar, but syntactic ambiguities are signalled to the user, who may backtrack and choose different syntactic paths. In any case, human interaction is always optional, and is used to improve the results. In case of no human intervention, the system simply outputs the most likely alternative, which is the one corresponding to the highest priority in the lexicon and in the grammar.
Go to IAN

NORMA is the UNL normalizer. It is expected to normalize knowledge bases, i.e., sets of statements expressed in the UNL KB format. Normalization, in this case, means suppression of redundancies (relations with the same source and target nodes); suppression of tautologies (relations where the source node is the same as the target node); suppression of contradictions (opposite relations between the same nodes, or the same relations between opposite nodes); generalization (replacing a node by a hyper-node, or a relation by a hyper-relation); specification (replacing a hyper-node by a node or a set of nodes, or a hyper-relation by a relation); merge (replacing two nodes by a single node); and division (replacing one node by two or more nodes). The main goal of the system is to organize and consolidate raw knowledge bases. NORMA has been developed by the Library of Alexandria.

Go to Norma

SEAN is the acronym for Shallow Enhanced ANalyser. Differently from IAN, it is fully automatic, and does not allow for any human intervention. The main differences to the other UNLization technologies are the following: 1) SEAN is a multi-document analyser: the input may be not only a single document (as in UNLeditor and IAN) but a whole collection of documents; 2) SEAN is a word-driven analyser: the unit of analysis is a word (and not a sentence as in UNLeditor and IAN), to be provided by the user; and 3) SEAN is a shallow analyzer: the analysis targets the surface structure of natural language sentences (and not the deep structure, as in UNLeditor and IAN). The main consequences of such choices are that the results of SEAN are not appropriate for translation, but for information retrieval and extraction only, because it provides a rather rough and partial analysis of the natural language input. SEAN has been developed by the Library of Alexandria.

Go to SEAN

	Status: release Date: 23 Jul 2013 Version: 1.1
UNL Editor is a graph-based UNL authoring tool.The whole decision-making process is human: the language specialist uploads the text to be analysed; selects the corresponding UWs (i.e., the nodes in the graph); creates semantic relations between nodes; and assigns attributes to nodes. All is done through a graphic interface that allows users to manipulate high-level graphs instead of low-level UNL statements. The system has been recently improved with a base of past UNLizations that is expected to accelerate this process, but the core decisions remain fully human. The UNLeditor is specially suitable for documents demanding high-quality UNLization, as the ones used in translation.
Go to UNL Editor

	Status: release Date: 19 Jul 2013 Version: 1.1
Dictionary Editor is a visual editor for creating and editing UNL Dictionaries both for IAN and EUGENE.
Go to Dictionary Editor

	Status: release Date: 19 Jul 2013 Version: 1.1
Rule Editor is a visual editor for creating and editing UNL Rules both for IAN and EUGENE.
Go to Rule Editor


UTk is a multi-platform web-based system. It is a leading platform for extracting and manipulating natural language data for natural language processing.
Go to UTK


AIDAN is part of a broader initiative, TUT (Text to Text through UNL), which is intended to prove the efficacy of UNL as a language-independent knowledge representation language, to be used in automatic summarization and text simplification.
Go to AIDAN

Welcome to {unl.dev}.

Welcome to {unl.dev}, a UNLWeb satellite project. Part of our content is for registered users only. In order to get full access please follow the sign in link on top. Please check UNL specs by UNDL Foundation (http://www.unlweb.net/wiki/).

Tools

What's new ?

EUGENE 1.2 and IAN 1.2
- Bug fixes.
- Performance optimization.
- UI and encoding support improvements.
- New Features:
   - Includes the option UNDO for the deletion of files and entries.
   - One click loading of files.
   - The range interval is user-defined.
   - Backtracking (top-down approach).
EUGENE 1.1
- Bug fixes.
- Performance optimization.
- UI and encoding support improvements.
- New Features:
   - In the list rules a node can be matched one or more times by adding "+" at the end of a node condition.
   - The trace level can be higher than "None" when processing sentences in range.
   - The "Compare" tool can work in "differences only" mode.
   - Mathematical operations within rules (for numeric nodes):
        (%x):=(%x-1); (i.e., reduce the value of %x in 1)
        (%x):=(%x+1); (i.e., add 1 to %x)
        (%x):=(%x*2); (i.e., multiply %x by 2)
        (%x):=(%x/2); (i.e., divide % by 2)
   - Assigning nodes UW value, or string value to attributes by reference:
        (%x,ATTRIBUTE=[%y])(%y):=ACTION; the system checks whether the value of the attribute ATTRIBUTE is the HEADWORD of %y
        (%x,ATTRIBUTE=[[%y]])(%y):=ACTION; the system checks whether the value of the attribute ATTRIBUTE is the UW of %y
        (%x,ATTRIBUTE="%y")(%y):=ACTION; the system checks whether the value of the attribute ATTRIBUTE is the STRING of %y
        CONDITION:=(%x,ATTRIBUTE=[%y])(%y): the system assigns the attribute ATTRIBUTE to %x with the value of the HEADWORD of %y
        CONDITION:=(%x,ATTRIBUTE=[[%y]])(%y): the system assigns the attribute ATTRIBUTE to %x with the value of the UW of %y
        CONDITION:=(%x,ATTRIBUTE="%y")(%y): the system assigns the attribute ATTRIBUTE to %x with the value of the STRING of %y

IAN 1.1
- Bug fixes
- Performance optimization
- UI and encoding support improvements
- New Features:
   - In the list rules a node can be matched one or more times by adding "+" at the end of a node condition.
   - The input text can be pre-processed for segmentation and normalization.
   - The result of pre-processing can be saved. The pre-processing can be skipped for such documents.
   - The trace level can be higher than "None" when processing sentences in range.
   - The "Compare" tool can work in "differences only" mode.
   - Mathematical operations within rules (for numeric nodes):
        (%x):=(%x-1); (i.e., reduce the value of %x in 1)
        (%x):=(%x+1); (i.e., add 1 to %x)
        (%x):=(%x*2); (i.e., multiply %x by 2)
        (%x):=(%x/2); (i.e., divide % by 2)
   - Assigning nodes UW value, or string value to attributes by reference:
        (%x,ATTRIBUTE=[%y])(%y):=ACTION; the system checks whether the value of the attribute ATTRIBUTE is the HEADWORD of %y
        (%x,ATTRIBUTE=[[%y]])(%y):=ACTION; the system checks whether the value of the attribute ATTRIBUTE is the UW of %y
        (%x,ATTRIBUTE="%y")(%y):=ACTION; the system checks whether the value of the attribute ATTRIBUTE is the STRING of %y
        CONDITION:=(%x,ATTRIBUTE=[%y])(%y): the system assigns the attribute ATTRIBUTE to %x with the value of the HEADWORD of %y
        CONDITION:=(%x,ATTRIBUTE=[[%y]])(%y): the system assigns the attribute ATTRIBUTE to %x with the value of the UW of %y
        CONDITION:=(%x,ATTRIBUTE="%y")(%y): the system assigns the attribute ATTRIBUTE to %x with the value of the STRING of %y

Dictionary Editor 1.1
- Bug fixes
- Performance optimization
- UI and encoding support improvements
- New Features:
   - Entries of a dictionary can be reordered by drag-n-drop.
   - Single dictionary can be loaded by clicking on its name.
   - Dictionary groups are collapsible/expandable.
   - Same dictionary can participate in more than one group.
   - The "New Dictionary" wizard now allows to update the content of an existing dictionary.

Rule Editor 1.1
- Bug fixes
- Performance optimization
- UI and encoding support improvements
- New Features:
   - Rules of a ruleset can be reordered by drag-n-drop.
   - Single ruleset can be loaded by clicking on its name.
   - Ruleset groups are collapsible/expandable.
   - Same ruleset can participate in more than one group.
   - The "New Ruleset" wizard now allows to update the content of an existing ruleset.
   - Better rule validation. Rules leading to impossible graphs are reported during the parsing.
   - New "N-rule" editor for pre-processing in IAN. Allows to manage List-to-List type rules.

UNL Editor 1.1
- Bug fixes
- New Features:
   - Grouping.
   - Better sharing functionality

EUGENE 1.0.1
IAN 1.0.1
- Bug fixes

EUGENE 1.0
IAN 1.0
UNL Editor 1.0
Rule Editor 1.0
Dictionary Editor 1.0
- Bug fixes
- Performance optimization
- New Features
- UI Improvements.

EUGENE 0.9.5
IAN 0.9.5
UNL Editor 0.9.5
Rule Editor 0.9.5
Dictionary Editor 0.9.5
- Bug fixes
- Performance optimization
- New Features
- UI Improvements.

EUGENE 0.9.1 beta
- Tabbed View
- JQuery integration.
- Bug Fixes.
- UI Improvements.
- Overall code optimization and cleanup.

UNL Editor 0.9.1
- Improved memory lookup.
- Expand/Collapse toggle button for all {unl} tags.
- Zipped download for documents.
- Arrowed relations in graph editor.
- Faster UNL/NL/Memory search.
- JQuery integration.
- Bug Fixes.
- Minor UI fixes.
- Overall code optimization and cleanup.
- Video Tutorials.

Dictionary Editor 0.9.1
- Zipped download for dictionaries.
- Improved search.
- JQuery integration.
- Bug Fixes.
- Minor UI fixes.
- Overall code optimization and cleanup.
- Video Tutorials.

Rule Editor 0.9.1
- Zipped download for rulesets.
- Improved search.
- JQuery integration.
- Bug Fixes.
- Minor UI fixes.
- Overall code optimization and cleanup.
- Video Tutorials.