JiangSu Provincial Key Laboratory for Computer Information Processing Technology

Suzhou, China

JiangSu Provincial Key Laboratory for Computer Information Processing Technology

Suzhou, China
SEARCH FILTERS
Time filter
Source Type

Huang X.,Soochow University of China | Lv Q.,Soochow University of China | Lv Q.,Jiangsu Provincial Key Laboratory for Computer Information Processing Technology | Qian P.-D.,Soochow University of China | Qian P.-D.,Jiangsu Provincial Key Laboratory for Computer Information Processing Technology
Zidonghua Xuebao/Acta Automatica Sinica | Year: 2011

This paper proposes an exemplar selection algorithm (ESA) for protein structures clustering, which is a necessary post-processing step for protein structure prediction. The widely-used quality threshold (QT) algorithm in protein structure prediction depends on clustering radius derived from experience, which also affects clustering distribution in other widely-used clustering algorithms such as affinity propagation (AP). The proposed exemplar selection algorithm can analyze clustering results, choose the best exemplar, and confirm clustering parameter such as clustering radius. Experimental results on real protein structure predictions confirm the effectiveness of our exemplar selection algorithm, which can choose the best exemplar with no experience parameter, and can find the best parameter fitting for data set. Copyright © 2011 Acta Automatica Sinica. All rights reserved.


Kong F.,JiangSu Provincial Key Laboratory for Computer Information Processing Technology | Kong F.,Soochow University of China | Zhou G.,JiangSu Provincial Key Laboratory for Computer Information Processing Technology | Zhou G.,Soochow University of China
PACLIC 25 - Proceedings of the 25th Pacific Asia Conference on Language, Information and Computation | Year: 2011

This paper systematically explores the effectiveness of dependency and constituent-based syntactic information for anaphoricity determination. In particular, this paper proposes two ways to combine dependency and constituent-based syntactic information to explore their complementary advantage. One is a dependency-driven constituent-based structured representation, and the other uses a composite kernel. Evaluation on the Automatic Content Extraction (ACE) 2003 corpus shows that dependency and constituent-based syntactic information are quite complementary and proper combination can much improve the performance of anaphoricity determination, and further improve the performance of coreference resolution.


Li L.-Z.,Soochow University of China | Zhu Y.-Q.,Soochow University of China | Zhu Y.-Q.,Jiangsu Provincial Key Laboratory for Computer Information Processing Technology | Yang Z.,Soochow University of China
Tongxin Xuebao/Journal on Communications | Year: 2010

Scalability is the mainly problem that impedes the extensive deployment of multicast in MPLS networks. An aggregation algorithm for multicast flow based on bidirectional shared tree was proposed in order to resolve the problem. The aggregated degrees of multicast flows were computed according to the relation among nodes. The label edge routers were clustered into the sets of leaf nodes on shared trees, and tree manager server computed the topology of bidirectional shared trees. The multicast flows, whose aggregated degrees were greater than the specific threshold, were finally converged into the trees. The test result indicates that it can greatly reduce the occupation of MPLS labels. The processes of medium nodes are simplified and the forwarding states of routers are reduced in the algorithm. It will greatly enhance the scalability of MPLS multicast.


Xu F.,Soochow University of China | Xu F.,Jiangsu Provincial Key Laboratory for Computer Information Processing Technology | Zhu Q.-M.,Soochow University of China | Zhu Q.-M.,Jiangsu Provincial Key Laboratory for Computer Information Processing Technology | And 2 more authors.
Ruan Jian Xue Bao/Journal of Software | Year: 2013

As a critical sub-task in discourse structure analysis, implicit discourse relation recognition (iDRR) is a challenging natural language processing task. Traditional approaches focus on exploring concepts and sense in discourse, which result in poor performance. This paper first systematically explores the efficiency of shallow semantic and attitude prosody-driven sentence-level sentiment information in discourse. Next, the paper proposes a simple but effective tree structure and finally investigates the efficiency of a composite kernel. Evaluation on Penn Discourse Treebank (PDTB) 2.0 shows the importance of shallow semantic and sentiment information across the discourse, and the appropriateness of the composite kernel in iDRR. It also shows that this system significantly outperforms other ones currently in the research field. © 2013 ISCAS.


Lv Q.,Soochow University of China | Lv Q.,Jiangsu Provincial Key Laboratory for Computer Information Processing Technology | Xia X.-Y.,Soochow University of China | Xia X.-Y.,Jiangsu Provincial Key Laboratory for Computer Information Processing Technology | And 2 more authors.
International Journal of Automation and Computing | Year: 2012

Frequent counting is a very so often required operation in machine learning algorithms. A typical machine learning task, learning the structure of Bayesian network (BN) based on metric scoring, is introduced as an example that heavily relies on frequent counting. A fast calculation method for frequent counting enhanced with two cache layers is then presented for learning BN. The main contribution of our approach is to eliminate comparison operations for frequent counting by introducing a multi-radix number system calculation. Both mathematical analysis and empirical comparison between our method and state-of-the-art solution are conducted. The results show that our method is dominantly superior to state-of-the-art solution in solving the problem of learning BN. © 2012 Institute of Automation, Chinese Academy of Sciences and Springer-Verlag Berlin Heidelberg.


Wen W.,Soochow University of China | Lu Q.,Soochow University of China | Lu Q.,Jiangsu Provincial Key Laboratory for Computer Information Processing Technology | Yang P.,Soochow University of China | And 3 more authors.
2010 4th International Conference on Bioinformatics and Biomedical Engineering, iCBBE 2010 | Year: 2010

Sidechain prediction is an important subproblem of protein design and structure prediction. Construction of rotamer library is the basis for protein sidechain prediction because it provides the basic searching space for prediction. However, the state-of-the-art rotamer libraries focus on the statistical information of individual amino acids, ignoring the direct affection of its adjacent amino acids. This article presents a sequence- and backbone-dependent rotamer library. Both the conformation information of adjacent amino acids and torsion angle of the current residue are taken into account to construct a sequence- and backbone-dependent library by HMM. Evaluation on all 13 free modeling targets of CASP8 based on our rotamer library is conducted. Comparing with side-chain prediction based on the state-of-the-art rotamer library, our library outperforms the sidechain prediction accuracy on all the test targets to a certain extent. © 2010 IEEE.


Li J.-H.,Soochow University of China | Li J.-H.,Jiangsu Provincial Key Laboratory for Computer Information Processing Technology | Zhou G.-D.,Soochow University of China | Zhou G.-D.,Jiangsu Provincial Key Laboratory for Computer Information Processing Technology | And 4 more authors.
Ruan Jian Xue Bao/Journal of Software | Year: 2011

This paper explores semantic role labeling (SRL) in the Chinese language for nominal predicates. In addition to the widely adopted features of verbal SRL, various nominal predicate-specific features are also explored. Moreover, the nominal SRL performance has been improved by properly integrating features that were derived from a state-of-the-art verbal SRL system. Finally, the paper explains in detail the nominal predicate recognition, which is essential in a fully automatic nominal SRL system. Evaluations on Chinese NomBank show that proper integration of a verbal SRL system significantly improves the performance of a nominal SRL. It also shows that this nominal SRL system achieves the performance of 72.67 in F1-measure on golden parse trees and golden predicates, and outperforms the state-of-the-art nominal SRL systems in the Chinese language; however, the performance drops to 55.14 in F1-measure on automatic parse trees and automatic predicates. © 2011 ISCAS.

Loading JiangSu Provincial Key Laboratory for Computer Information Processing Technology collaborators
Loading JiangSu Provincial Key Laboratory for Computer Information Processing Technology collaborators