Runs - Novelty 2003¶

ccsum2svdpqr¶

Run ID: ccsum2svdpqr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/8/2003
Task: task2
Run description: Did not use topic info. SVD followed by pivoted QR

ccsum3pqr¶

Run ID: ccsum3pqr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task3
Run description: HMMtrained on top 5 Pivoted QR Task3

ccsum3qr¶

Run ID: ccsum3qr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task3
Run description: Trained HMM with relevant sentences from first 5 docs and then used QR to extract novel sentences.

ccsum3svdpqr¶

Run ID: ccsum3svdpqr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Trained HMM on first 5 docs and used it to generate relevant sentences and then used SVD and pivoted QR for finding novel.

ccsum4spq001¶

Run ID: ccsum4spq001
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/16/2003
Task: task4
Run description: No use of topic fields SVD followed by QR with a threshold of .001

ccsum4svdpqr¶

Run ID: ccsum4svdpqr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/16/2003
Task: task4
Run description: No use of topic fields SVD followed by pivoted QR with a threshold of .001

ccsumlaqr¶

Run ID: ccsumlaqr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: HMM using LA Times relevant sentence for training and then followed a QR with partial pivoting. We do not use the topic field at all, your web form wouldn't let me leave it blank!

ccsummeoqr¶

Run ID: ccsummeoqr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: HMM using hand marked subset of 2003 novel sentences and then followed a QR with partial pivoting. We do not use the topic field at all, your web form wouldn't let me leave it blank!

ccsummeosvd¶

Run ID: ccsummeosvd
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: HMM using hand marked subset of 2003 novel sentences and then followed a SVD and a QR with partial pivoting. We do not use the topic field at all, your web form wouldn't let me leave it blank!

ccsumrelqr¶

Run ID: ccsumrelqr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: HMM using 2002 relevant sentences for training and then followed a QR with partial pivoting. We do not use the topic field at all, your web form wouldn't let me leave it blank!

ccsumrelsvd¶

Run ID: ccsumrelsvd
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: HMM using 2002 relevant sentences for training and then followed a SVD and a QR with partial pivoting. We do not use the topic field at all, your web form wouldn't let me leave it blank!

ccsumt2pqr¶

Run ID: ccsumt2pqr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/8/2003
Task: task2
Run description: Did not use topic info. Pivoted QR to select sentences

ccsumt2qr¶

Run ID: ccsumt2qr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/8/2003
Task: task2
Run description: Did not use topic info. QR to select sentences

ccsumt2svdqr¶

Run ID: ccsumt2svdqr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/8/2003
Task: task2
Run description: Did not use topic info. SVD followed by QR

ccsumt4pqr¶

Run ID: ccsumt4pqr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/16/2003
Task: task4
Run description: No use of topic fields QR with partial pivoting and a threshold of 0.01

ccsumt4qr¶

Run ID: ccsumt4qr
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/16/2003
Task: task4
Run description: No use of topic fields QR (no pivoting with a threshold of 0.7)

ccsumt4sqr01¶

Run ID: ccsumt4sqr01
Participant: ccs.conroy
Track: Novelty
Year: 2003
Submission: 9/16/2003
Task: task4
Run description: No use of topic fields SVD followed by pivoted QR with a threshold of .001

clr03n1d¶

Run ID: clr03n1d
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: CL Research parses and processes sentences into an XML representation of discourse structure, discourse entities, verbs, and prepositions, which is then used for matching up with a parsed representation of the topics.

clr03n1n2¶

Run ID: clr03n1n2
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: CL Research parses and processes sentences into an XML representation of discourse structure, discourse entities, verbs, and prepositions, which is then used for matching up with a parsed representation of the topics.

clr03n1n3¶

Run ID: clr03n1n3
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: CL Research parses and processes sentences into an XML representation of discourse structure, discourse entities, verbs, and prepositions, which is then used for matching up with a parsed representation of the topics.

clr03n1t¶

Run ID: clr03n1t
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: CL Research parses and processes sentences into an XML representation of discourse structure, discourse entities, verbs, and prepositions, which is then used for matching up with a parsed representation of the topics.

clr03n2¶

Run ID: clr03n2
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/8/2003
Task: task2
Run description: The CL Research system processes text into an XML representation which is then used for assessing relevance and novelty.

clr03n3f01¶

Run ID: clr03n3f01
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: The CL Research system parses and processes text into an XML representation, tagging the text with discourse, noun, verb, and preposition characteristics, which are then used in determining relevance and novelty.

clr03n3f02¶

Run ID: clr03n3f02
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: The CL Research system parses and processes text into an XML representation, tagging the text with discourse, noun, verb, and preposition characteristics, which are then used in determining relevance and novelty.

clr03n3f03¶

Run ID: clr03n3f03
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: The CL Research system parses and processes text into an XML representation, tagging the text with discourse, noun, verb, and preposition characteristics, which are then used in determining relevance and novelty.

clr03n3f04¶

Run ID: clr03n3f04
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: The CL Research system parses and processes text into an XML representation, tagging the text with discourse, noun, verb, and preposition characteristics, which are then used in determining relevance and novelty.

clr03n3f05¶

Run ID: clr03n3f05
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: The CL Research system parses and processes text into an XML representation, tagging the text with discourse, noun, verb, and preposition characteristics, which are then used in determining relevance and novelty.

clr03n4¶

Run ID: clr03n4
Participant: clresearch
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task4
Run description: The CL Research system parses and processes text into an XML representation, tagging the text with discourse, noun, verb, and preposition characteristics, which are then used in determining relevance and novelty.

ICT03NOV1BSL¶

Run ID: ICT03NOV1BSL
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: The similarity between topic and sentence is computed according to vector space model.

ICT03NOV1DTH¶

Run ID: ICT03NOV1DTH
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: The similarity between topic and sentence is computed according to vector space model. The threshold value for determination of relevance is automatically adapted for each document accoring to the time distribution of the 25 relevant documents.

ICT03NOV1NAR¶

Run ID: ICT03NOV1NAR
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: The similarity between topic and sentence is computed according to vector space model. The positive feature vector and negative feature vector are constructed according to title and narrative in the topic.

ICT03NOV1SQR¶

Run ID: ICT03NOV1SQR
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: The similarity between topic and sentence is computed according to vector space model, in which we utilize the chi-square statistic for the feature selection and feature weigthing.

ICT03NOV1XTD¶

Run ID: ICT03NOV1XTD
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: The similarity between topic and sentence is computed according to vector space model. The retrieved 75 documents and given 25 relevant documents are used for feature selection.

ICT03NOV2CUR¶

Run ID: ICT03NOV2CUR
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: We used the local cooccurence as the query expansion. Maximum Marginal Relevance was used to find the new sentence.

ICT03NOV2LPA¶

Run ID: ICT03NOV2LPA
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: We used the word overlap between two sentences to select the new sentence.

ICT03NOV2LPP¶

Run ID: ICT03NOV2LPP
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: We used the chi-square for the feature selection. The percentage of new sentences in relevant sentences is different accoring to the ranking of the current document in 25 documents.

ICT03NOV2PNK¶

Run ID: ICT03NOV2PNK
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: We used the relevant sentences and irrelerant sentences in the 25 documents to extract the features. Maximum Marginal Relevance was used to find the new sentence.

ICT03NOV2SQR¶

Run ID: ICT03NOV2SQR
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: We used the chi-square for the feature selection. Maximum Marginal Relevance is used to find the new sentence.

ICT03NOV3IKK¶

Run ID: ICT03NOV3IKK
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: KNN algorithm is used to select the relevant sentences.

ICT03NOV3KNN¶

Run ID: ICT03NOV3KNN
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: KNN algorithm is used to select the relevant sentences.

ICT03NOV3KNS¶

Run ID: ICT03NOV3KNS
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: KNN algorithm is used to select the relevant sentences.

ICT03NOV3WN3¶

Run ID: ICT03NOV3WN3
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Winnow algorithm is used to select the relevant sentences.

ICT03NOV3WND¶

Run ID: ICT03NOV3WND
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Winnow algorithm is used to select the relevant sentences.

ICT03NOV4ALL¶

Run ID: ICT03NOV4ALL
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Winnow algorithm is used to retrieve the new sentences given the relevant and new sentences based on words overlapping, sentence semantic distance, and head sentence tags.

ICT03NOV4LFF¶

Run ID: ICT03NOV4LFF
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Winnow algorithm is used to retrieve the new sentences given the relevant and new sentences based on words overlapping and sentence semantic distance.

ICT03NOV4OTP¶

Run ID: ICT03NOV4OTP
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Winnow algorithm is used to retrieve the new sentences given the relevant and new sentences based on words overlapping.

ICT03NOV4SQR¶

Run ID: ICT03NOV4SQR
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Maximum Marginal Relevance is used to retrieve the new sentence.

ICT03NOV4WNW¶

Run ID: ICT03NOV4WNW
Participant: cas-ict.bin
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Winnow algorithm is used to retrieve the new sentences given the relevant sentences.

IITBN1¶

Run ID: IITBN1
Participant: iitb.ramakrishnan
Track: Novelty
Year: 2003
Submission: 9/18/2003
Task: task4
Run description: principal component analysis with random walks on wordnet

Irit1¶

Run ID: Irit1
Participant: irit-sig.boughanem
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Coverage of 4 sentences, single terms used during analysis

Irit5q¶

Run ID: Irit5q
Participant: irit-sig.boughanem
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Coverage of 5 sentences, single terms used during analysis

IRITf2bis¶

Run ID: IRITf2bis
Participant: irit-sig.boughanem
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Term weighting (hightly relevant, lowly relevant, non relevant). Based on the full topic.

IRITfb1MtmIb¶

Run ID: IRITfb1MtmIb
Participant: irit-sig.boughanem
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Four types of terms are considered high, low, non relevant (same as IRITf2bis), in addition we consider irrelevant terms. Terms can be single words or phrases.

IRITfNegR2¶

Run ID: IRITfNegR2
Participant: irit-sig.boughanem
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Relevant sentences same as IRITf2bis. New sentences low filtering.

IritMtm4¶

Run ID: IritMtm4
Participant: irit-sig.boughanem
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Coverage of 4 sentences, phrases used text processing

IritMtm5¶

Run ID: IritMtm5
Participant: irit-sig.boughanem
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Coverage of 5 sentences, phrases used text processing

IRITnb1MtmI4¶

Run ID: IRITnb1MtmI4
Participant: irit-sig.boughanem
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: a modified version of run IRITfb1MtmI (parameter)

IRITnip2bis¶

Run ID: IRITnip2bis
Participant: irit-sig.boughanem
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Four types of terms are considered high, low, non relevant (like in IRITf2bis), in addition we consider irrelevant terms (narrative part).

Irito¶

Run ID: Irito
Participant: irit-sig.boughanem
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: nothing (last year many tries from different groups to 'filter' new sentences were less efficient than selecting all the relevant sentences).

ISIALL03¶

Results | Participants | Input | Summary | Appendix

Run ID: ISIALL03
Participant: usc-isi.hermjakob
Track: Novelty
Year: 2003
Submission: 8/30/2003
Task: task1
Run description: All sentences are assumed relevant.

ISIDSCm203¶

Results | Participants | Input | Summary | Appendix

Run ID: ISIDSCm203
Participant: usc-isi.hermjakob
Track: Novelty
Year: 2003
Submission: 8/30/2003
Task: task1
Run description: a presence of any subjective word in a sentence is the key of opinions.

ISIDSm203¶

Results | Participants | Input | Summary | Appendix

Run ID: ISIDSm203
Participant: usc-isi.hermjakob
Track: Novelty
Year: 2003
Submission: 8/30/2003
Task: task1
Run description: a presence of any subjective word in a sentence is the key of opinions.

ISINONE03¶

Results | Participants | Input | Summary | Appendix

Run ID: ISINONE03
Participant: usc-isi.hermjakob
Track: Novelty
Year: 2003
Submission: 8/30/2003
Task: task1
Run description: None of sentences are assumed relevant.

ISIRAND03¶

Results | Participants | Input | Summary | Appendix

Run ID: ISIRAND03
Participant: usc-isi.hermjakob
Track: Novelty
Year: 2003
Submission: 8/30/2003
Task: task1
Run description: Random selection

lexiclone03¶

Results | Participants | Input | Summary | Appendix

Run ID: lexiclone03
Participant: lexiclone.geller
Track: Novelty
Year: 2003
Submission: 8/29/2003
Task: task1
Run description: we used lexical cloning technology

MeijiHilF11¶

Run ID: MeijiHilF11
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Relevant filtering approach using conceptual fuzzy sets. Novelty calculating sentence score and redundancy score using conceptual fuzzy sets.

MeijiHilF12¶

Run ID: MeijiHilF12
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Relevant filtering approach using conceptual fuzzy sets. Novelty calculating sentence score and redundancy score.

MeijiHilF13¶

Run ID: MeijiHilF13
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Relevant filtering approach using conceptual fuzzy sets. Novelty calculating sentence score and redundancy score using conceptual fuzzy sets.

MeijiHilF14¶

Run ID: MeijiHilF14
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Relevant filtering approach using conceptual fuzzy sets. Novelty calculating sentence score and redundancy score

MeijiHilF15¶

Run ID: MeijiHilF15
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Relevant filtering approach using tf-idf word vector. Novelty calculating sentence score and redundancy score

MeijiHilF21¶

Run ID: MeijiHilF21
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: calculating sentence score using time window and redundancy score using conceptual fuzzy sets.

MeijiHilF22¶

Run ID: MeijiHilF22
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: calculating sentence score using time window and redundancy score using tf-idf word vector.

MeijiHilF23¶

Run ID: MeijiHilF23
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: calculating sentence score using time window.

MeijiHilF24¶

Run ID: MeijiHilF24
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: calculating basic sentence score only.

MeijiHilF31¶

Run ID: MeijiHilF31
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Relevant filtering approach using conceptual fuzzy sets. Novelty calculating sentence score and redundancy score using conceptual fuzzy sets.

MeijiHilF32¶

Run ID: MeijiHilF32
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Relevant filtering approach using conceptual fuzzy sets. Novelty calculating sentence score and redundancy score.

MeijiHilF33¶

Run ID: MeijiHilF33
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Relevant filtering approach using tf-idf word vector. Novelty calculating sentence score and redundancy score using conceptual fuzzy sets.

MeijiHilF34¶

Run ID: MeijiHilF34
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Relevant filtering approach using tf-idf word vector. Novelty calculating sentence score and redundancy score.

MeijiHilF41¶

Run ID: MeijiHilF41
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/12/2003
Task: task4
Run description: calculating sentence score using time window and redundancy score using conceptual fuzzy sets.

MeijiHilF42¶

Run ID: MeijiHilF42
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/12/2003
Task: task4
Run description: calculating sentence score using time window and redundancy score using tf-idf word vector.

MeijiHilF43¶

Run ID: MeijiHilF43
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/12/2003
Task: task4
Run description: calculating sentence score using time window.

MeijiHilF44¶

Run ID: MeijiHilF44
Participant: meijiu.takagi
Track: Novelty
Year: 2003
Submission: 9/12/2003
Task: task4
Run description: calculating basic sentence score only.

NLPR03n1f1¶

Run ID: NLPR03n1f1
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/2/2003
Task: task1
Run description: A Combination of following algorithms tf-idf, length normalization, relevant feedback and dynamic threshold, etc.

NLPR03n1f2¶

Run ID: NLPR03n1f2
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/2/2003
Task: task1
Run description: A Combination of following algorithms tf-idf, length normalization, relevant feedback and dynamic threshold, etc.

NLPR03n1w1¶

Run ID: NLPR03n1w1
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/2/2003
Task: task1
Run description: A Combination of following algorithms window-based weighting, relevant feedback and dynamic threshold, etc.

NLPR03n1w2¶

Run ID: NLPR03n1w2
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/2/2003
Task: task1
Run description: A Combination of following algorithms window-based weighting, relevant feedback and dynamic threshold, etc.

NLPR03n1w3¶

Run ID: NLPR03n1w3
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/2/2003
Task: task1
Run description: A Combination of following algorithms window-based weighting, relevant feedback and dynamic threshold, etc.

NLPR03n2d1¶

Run ID: NLPR03n2d1
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task2
Run description: We define a value called 'New Information Degree'(NID) based on 'idf' values to present whether a sentence includes new information related to the former sentences, and dynamic thresholds are used for different topics.

NLPR03n2d2¶

Run ID: NLPR03n2d2
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task2
Run description: We define a value called 'New Information Degree'(NID) based on 'bigram' to present whether a sentence includes new information related to the former sentences, and dynamic thresholds are used for different topics.

NLPR03n2d3¶

Run ID: NLPR03n2d3
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task2
Run description: We define a value called 'New Information Degree'(NID) based on 'idf' and 'bigram' to present whether a sentence includes new information related to the former sentences, and dynamic thresholds are used for different topics.

NLPR03n2s1¶

Run ID: NLPR03n2s1
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task2
Run description: We define a value called 'New Information Degree'(NID) based on 'idf' values to present whether a sentence includes new information related to the former sentences, and static threshold is used for different topics.

NLPR03n2s2¶

Run ID: NLPR03n2s2
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task2
Run description: We define a value called 'New Information Degree'(NID) based on 'bigram' to present whether a sentence includes new information related to the former sentences, and static threshold is used for different topics.

NLPR03n3d1¶

Run ID: NLPR03n3d1
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: We use the combination of the following algorithms core-window-based similarity degrees, relevant feedback, length normalizatoin, bigram-based new information retrieval and dynamic thresholds.

NLPR03n3d2¶

Run ID: NLPR03n3d2
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: We use the combination of the following algorithms tf-idf similarity degrees, query expansion, relevant feedback, length normalizatoin, bigram-based new information retrieval and dynamic thresholds.

NLPR03n3d3¶

Run ID: NLPR03n3d3
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: We use the combination of the following algorithms window-based similarity degrees, relevant feedback, length normalizatoin, idf-based new information retrieval and dynamic thresholds.

NLPR03n3s1¶

Run ID: NLPR03n3s1
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: We use the combination of the following algorithms core-window-based similarity degrees, relevant feedback, length normalizatoin, idf-based new information retrieval.

NLPR03n3s2¶

Run ID: NLPR03n3s2
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: We use the combination of the following algorithms tf-idf similarity degrees, query expansion, relevant feedback, length normalizatoin, idf-based new information retrieval.

NLPR03n4d1¶

Run ID: NLPR03n4d1
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/14/2003
Task: task4
Run description: We develop "New Information degree" based on 'idf' to represent whether a sentence includes novelty information, and use dynamic thresholds for different topics.

NLPR03n4d2¶

Run ID: NLPR03n4d2
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/14/2003
Task: task4
Run description: We develop "New Information degree" based on bigram to represent whether a sentence includes novelty information, and use dynamic thresholds (learned from training data) for different topics.

NLPR03n4s1¶

Run ID: NLPR03n4s1
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/14/2003
Task: task4
Run description: We develop "New Information degree" based on idf to represent whether a sentence includes novelty information, and use static threshold (learned from training data) for different topics.

NLPR03n4s2¶

Run ID: NLPR03n4s2
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/14/2003
Task: task4
Run description: We develop "New Information degree" based on bigram to represent whether a sentence includes novelty information, and use static threshold (learned from training data) for different topics.

NLPR03n4s3¶

Run ID: NLPR03n4s3
Participant: cas.nlpr
Track: Novelty
Year: 2003
Submission: 9/14/2003
Task: task4
Run description: We develop "New Information degree" based on 'tf' to represent whether a sentence includes novelty information, and use static threshold (learned from training data) for different topics.

NTU11¶

Run ID: NTU11
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/2/2003
Task: task1
Run description: Use IR System with Reference Corpus

NTU12¶

Run ID: NTU12
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/2/2003
Task: task1
Run description: Use IR System with Reference Corpus

NTU13¶

Run ID: NTU13
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/2/2003
Task: task1
Run description: Use IR System with Reference Corpus

NTU14¶

Run ID: NTU14
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/2/2003
Task: task1
Run description: Use IR System with Reference Corpus

NTU15¶

Run ID: NTU15
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/2/2003
Task: task1
Run description: Use IR System with Reference Corpus

NTU21¶

Run ID: NTU21
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task2
Run description: Use IR with reference corpus

NTU22¶

Run ID: NTU22
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task2
Run description: Use IR with reference corpus

NTU23¶

Run ID: NTU23
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task2
Run description: Use IR with reference corpus

NTU24¶

Run ID: NTU24
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task2
Run description: Use IR with reference corpus

NTU25¶

Run ID: NTU25
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task2
Run description: Use IR with reference corpus

NTU31¶

Run ID: NTU31
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: Use IR with reference corpus

NTU32¶

Run ID: NTU32
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: Use IR with reference corpus

NTU33¶

Run ID: NTU33
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: Use IR with reference corpus

NTU34¶

Run ID: NTU34
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: Use IR with reference corpus

NTU35¶

Run ID: NTU35
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/9/2003
Task: task3
Run description: Use IR with reference corpus

NTU41¶

Run ID: NTU41
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/16/2003
Task: task4
Run description: Use IR with reference corpus

NTU42¶

Run ID: NTU42
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/16/2003
Task: task4
Run description: Use IR with reference corpus

NTU43¶

Run ID: NTU43
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/16/2003
Task: task4
Run description: Use IR with reference corpus

NTU44¶

Run ID: NTU44
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/16/2003
Task: task4
Run description: Use IR with reference corpus

NTU45¶

Run ID: NTU45
Participant: ntu.chen
Track: Novelty
Year: 2003
Submission: 9/16/2003
Task: task4
Run description: Use IR with reference corpus

THUIRnv0311¶

Run ID: THUIRnv0311
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Local feedback to find relevant information and sentence to sentece overlap judgement.

THUIRnv0312¶

Run ID: THUIRnv0312
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Using WordNet synset and local co-occurrence MI to QE on relevant step, and sentence to sentece overlap judgement.

THUIRnv0313¶

Run ID: THUIRnv0313
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Retrieval using probabilistic model on relevant step (as baseline), and sentence to sentece overlap judgement.

THUIRnv0314¶

Run ID: THUIRnv0314
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Result filtering using different weighting scheme on relevant step, and sentence to sentece overlap judgement.

THUIRnv0315¶

Run ID: THUIRnv0315
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Sub-topic retrieval using long query on relevant step, and sentence to sentece overlap judgement.

THUIRnv0321¶

Run ID: THUIRnv0321
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task2
Run description: event . opinion

THUIRnv0322¶

Run ID: THUIRnv0322
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task2
Run description: nv4

THUIRnv0323¶

Run ID: THUIRnv0323
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task2
Run description: fixed threhold

THUIRnv0331¶

Run ID: THUIRnv0331
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task3
Run description: svm+nv4

THUIRnv0332¶

Run ID: THUIRnv0332
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task3
Run description: log+fb

THUIRnv0333¶

Run ID: THUIRnv0333
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task3
Run description: long+mi+qe

THUIRnv0334¶

Run ID: THUIRnv0334
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task3
Run description: long

THUIRnv0341¶

Run ID: THUIRnv0341
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Using triple overlap to finding new information with fixed redundancy threshold.

THUIRnv0342¶

Run ID: THUIRnv0342
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Using triple overlap to finding new information with fixed redundancy threshold.

THUIRnv0343¶

Run ID: THUIRnv0343
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Using documents cosine similarity to find new information with fixed redundancy threshold.

THUIRnv0344¶

Run ID: THUIRnv0344
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Using QE-based documents overlap to find new information, in which QE dictionary is generated by co-occurence in relevant sentences.

THUIRnv0345¶

Run ID: THUIRnv0345
Participant: tsinghuau.ma
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Using documents overlap and clustering of event and opinion topics to find new information.

UIowa03Nov01¶

Run ID: UIowa03Nov01
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: topic-sentence cosine similarity as relevance, with a minumum of one new entity as a criteria for 'new'

UIowa03Nov02¶

Run ID: UIowa03Nov02
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/4/2003
Task: task1
Run description: cosine similarity on topic-sentence as guard for relevance on a minimum of at least one new entity.

UIowa03Nov03¶

Run ID: UIowa03Nov03
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: given the relevant sentence for a topic, declare it novel...

UIowa03Nov04¶

Run ID: UIowa03Nov04
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: given the relevant sentence for a topic, declare it novel if there is at least one new entity or noun phrase

UIowa03Nov05¶

Run ID: UIowa03Nov05
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: given the relevant sentence for a topic, declare it novel if there is at least two new entities or noun phrases

UIowa03Nov06¶

Run ID: UIowa03Nov06
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 09/10/03
Task: task2
Run description: given the relevant sentence for a topic, declare it novel if there is at least three new entities or noun phrases

UIowa03Nov07¶

Run ID: UIowa03Nov07
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: given the relevant sentence for a topic, declare it novel if there is at least four new entities or noun phrases

UIowa03Nov08¶

Run ID: UIowa03Nov08
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: We expand the query term vector with relevant sentence terms, log the noun phrases and entities, and then use a guard similarity to judge relevance and occurrence of new NPs/entities for novelty.

UIowa03Nov09¶

Run ID: UIowa03Nov09
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: We expand the query term vector with relevant sentence terms, log the noun phrases and entities, and then use a guard similarity to judge relevance and occurrence of new NPs/e

UIowa03Nov10¶

Run ID: UIowa03Nov10
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task4
Run description: given the relevant sentence for a topic, declare it novel...

UIowa03Nov11¶

Run ID: UIowa03Nov11
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task4
Run description: given the relevant sentence for a topic, declare it novel if there is at least one new entity or noun phrase

UIowa03Nov12¶

Run ID: UIowa03Nov12
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task4
Run description: given the relevant sentence for a topic, declare it novel if there is at least two new entities or noun phrases

UIowa03Nov13¶

Run ID: UIowa03Nov13
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task4
Run description: given the relevant sentence for a topic, declare it novel if there is at least three new entities or noun phrases

UIowa03Nov14¶

Run ID: UIowa03Nov14
Participant: uiowa.eichmann
Track: Novelty
Year: 2003
Submission: 9/11/2003
Task: task4
Run description: given the relevant sentence for a topic, declare it novel if there is at least four new entities or noun phrases

umbcnew1¶

Run ID: umbcnew1
Participant: umarylandbc.kallurkar
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Cluster relevant sentences and return one sentence per cluster as novel sentences.

umbcnew2¶

Run ID: umbcnew2
Participant: umarylandbc.kallurkar
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Compute sentence-sentence similarity and return dissimilar sentences as novel sentences.

umbcnew3¶

Run ID: umbcnew3
Participant: umarylandbc.kallurkar
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Compute sentence-sentence similarity and return dissimilar sentences as novel sentences, using the reduced dimension SVD computation.

umbcrun1¶

Run ID: umbcrun1
Participant: umarylandbc.kallurkar
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Query expansion using terms from the description for finding correlated terms, clustering sentences, and querying to find the best clusters

umbcrun2¶

Run ID: umbcrun2
Participant: umarylandbc.kallurkar
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Query expansion by finding usefull terms such as nouns, verbs adjectives and then, clustering sentences and querying to find the best clusters

umbcrun3¶

Run ID: umbcrun3
Participant: umarylandbc.kallurkar
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: Query expansion by selecting the terms with largest term-term similarity scores using SVD and then, clustering sentences and querying to find the best clusters

umich1¶

Run ID: umich1
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: We used a maximum entropy classifier with sentence features extracted using the MEAD summarizer in choosing novel and relevant sentences.

umich2¶

Run ID: umich2
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: We used a maximum entropy classifier with sentence features extracted using the MEAD summarizer in choosing novel and relevant sentences.

umich21¶

Run ID: umich21
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Used 6 sentence features calcuated using the MEAD software to train a maximum entropy model to predict novel sentences.

umich22¶

Run ID: umich22
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Used 6 sentence features calcuated using the MEAD software to train a maximum entropy model to predict novel sentences.

umich23¶

Run ID: umich23
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Used 6 sentence features calcuated using the MEAD software to train a maximum entropy model to predict novel sentences.

umich24¶

Run ID: umich24
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Used 6 sentence features calcuated using the MEAD software to train a maximum entropy model to predict novel sentences.

umich25¶

Run ID: umich25
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task2
Run description: Used 6 sentence features calcuated using the MEAD software to train a maximum entropy model to predict novel sentences.

umich3¶

Run ID: umich3
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: We used a maximum entropy classifier with sentence features extracted using the MEAD summarizer in choosing novel and relevant sentences.

umich31¶

Run ID: umich31
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Used 6 sentence features calcuated using the MEAD software to train a maximum entropy model to predict novel sentences.

umich32¶

Run ID: umich32
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Used 6 sentence features calcuated using the MEAD software to train a maximum entropy model to predict novel sentences.

umich33¶

Run ID: umich33
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Used 6 sentence features calcuated using the MEAD software to train a maximum entropy model to predict novel sentences.

umich34¶

Run ID: umich34
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Used 6 sentence features calcuated using the MEAD software to train a maximum entropy model to predict novel sentences.

umich35¶

Run ID: umich35
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/10/2003
Task: task3
Run description: Used 6 sentence features calcuated using the MEAD software to train a maximum entropy model to predict novel sentences.

umich4¶

Run ID: umich4
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: We used a maximum entropy classifier with sentence features extracted using the MEAD summarizer in choosing novel and relevant sentences.

umich41¶

Run ID: umich41
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Trained a maximum entropy model using 6 MEAD-based sentence features to distinguish novel sentences.

umich42¶

Run ID: umich42
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Trained a maximum entropy model using 6 MEAD-based sentence features to distinguish novel sentences.

umich43¶

Run ID: umich43
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Trained a maximum entropy model using 6 MEAD-based sentence features to distinguish novel sentences.

umich44¶

Run ID: umich44
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Trained a maximum entropy model using 6 MEAD-based sentence features to distinguish novel sentences.

umich45¶

Run ID: umich45
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/17/2003
Task: task4
Run description: Trained a maximum entropy model using 6 MEAD-based sentence features to distinguish novel sentences.

umich5¶

Run ID: umich5
Participant: umich.radev
Track: Novelty
Year: 2003
Submission: 9/3/2003
Task: task1
Run description: We used a maximum entropy classifier with sentence features extracted using the MEAD summarizer in choosing novel and relevant sentences.