forked from andre-martins/TurboParser
-
Notifications
You must be signed in to change notification settings - Fork 0
/
ChangeLog
48 lines (42 loc) · 2.21 KB
/
ChangeLog
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
2009-11-26 TurboParser 0.1 [email protected]
Initial implementation, based on a C++ implementation of MSTParser and
using the commercial solver ILOG CPLEX.
2012-09-18 TurboParser 2.0 [email protected]
Major refactoring:
* Implementation of TurboTagger, a sequence model for POS tagging.
* TurboParser now outputs labeled dependencies.
* Dictionary and distance pruning to achieve fast decoding.
* Replaced ILOG CPLEX by the free library AD3.
* Now use free libraries glog and gflags for easy logging and commandline
flag processing.
2012-10-10 TurboParser 2.0.1 [email protected]
Fixed initialization of STL vectors, which was causing problems in some
compilers. Replaced the AD3 package by a newer version that also addresses
this issue (thanks to Atro Voutilainen and Denis Maua for reporting this).
2012-12-06 TurboParser 2.0.2 [email protected]
Fixed a bug that where causing some of the predicted parses not to be
valid trees. Now there is a final projection step that ensures that all
output parses are trees.
2013-05-23 TurboParser 2.1.0 [email protected]
New release with some additional features:
* Introduced third-order parts (as described in our ACL 2013 short paper).
* Windows-compatible (tested under MSVC; thanks to Afonso Mendes).
* TurboTagger is now faster.
2014-06-26 TurboParser 2.2.0 [email protected]
New release with some additional features:
* Implemented a Python wrapper using Cython.
* Added a semantic parser, TurboSemanticParser (as described in our SemEval
2014 paper)
* Added a tokenizer for Portuguese.
2015-11-06 TurboParser 2.3.0 [email protected]
New release with some additional features:
* Added a named entity recognizer, TurboEntityRecognizer, based on the
Illinois Entity Tagger.
* Added a coreference resolver, TurboCoreferenceResolver, based on the
Berkeley Coreference Resolution System.
* Implemented a constituent parser by reducing that problem to dependency
parser followed by unary recovery (as described in our ACL 2015 paper).
* As a by-product, implemented an external dependency labeler,
TurboDependencyLabeler, to be applied optionally after the dependency parser.
* Compatible with C++0x.
* Windows-compatible (tested under MSVC; thanks to David Nogueira).