Runs - Chemical 2011¶

BiTeMboth¶

Run ID: BiTeMboth
Participant: BiTeM
Track: Chemical
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: techsurv
MD5: 380b803b7adc94a7b8884e9a984380be
Run description: Previous text and MeSH runs combined.

BiTeMmesh¶

Run ID: BiTeMmesh
Participant: BiTeM
Track: Chemical
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: techsurv
MD5: 74933a1e5d2364c17be2758142e477d4
Run description: MeSH concepts oriented Document Representation.

BiTeMPAcitFB¶

Run ID: BiTeMPAcitFB
Participant: BiTeM
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart
MD5: 83e705ec985696e99dcbd21bca281b8f
Run description: Run BiTeMPAcomb with citations feedback strategy.

BiTeMPAcomb¶

Run ID: BiTeMPAcomb
Participant: BiTeM
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart
MD5: 088fb74fb1c4d5dddec12403a5af1caf
Run description: Fusion of BiTeMPAtext and BiTeMPAmesh.

BiTeMPAmesh¶

Run ID: BiTeMPAmesh
Participant: BiTeM
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart
MD5: daf72575c4ad315d21dfaebd0a4805ba
Run description: Document representation in MeSH concepts (mapped in documents and topics). Application and publication dates filtering.

BiTeMPAtext¶

Run ID: BiTeMPAtext
Participant: BiTeM
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart
MD5: 8b289ad43756a3c6734308a509f08b8c
Run description: Simple document representation with free-text fields. Application and publication dates filtering.

BiTeMtext¶

Run ID: BiTeMtext
Participant: BiTeM
Track: Chemical
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: techsurv
MD5: 6bab1313d4f492e5c394be23be9344b5
Run description: simple text representation.

chemoCR_v93_p10¶

Run ID: chemoCR_v93_p10
Participant: chemoCR
Track: Chemical
Year: 2011
Submission: 7/29/2011
Task: image2struct
Run description: this run has been produced by running chemoCR v0.93 in batch mode, using parameter set Houben-Weyl, switched on validation of the result and cleaning of small fragments, no image preprocessing used, some letters have been trained for the OCR

ChemReader_Run1¶

Run ID: ChemReader_Run1
Participant: ChemReader
Track: Chemical
Year: 2011
Submission: 8/7/2011
Task: image2struct
Run description: TREC-CHEM 2011 Image2Structure Task ChemReader's result (Set #1)

ChemReader_Run2¶

Run ID: ChemReader_Run2
Participant: ChemReader
Track: Chemical
Year: 2011
Submission: 8/7/2011
Task: image2struct
Run description: TREC-CHEM 2011 Image2Structure Task, ChemReader's result (Set #2)

CMUTSmans¶

Run ID: CMUTSmans
Participant: CMU.LIRA
Track: Chemical
Year: 2011
Submission: 8/5/2011
Type: manual
Task: techsurv
MD5: bd4efbd591a93d18ecf27bd13e4599c5
Run description: Manually created Boolean CNF style queries. Used whole topic descriptions. Consulted PubMed, PubChem, Thesaurus.com and Wikipedia. Retrieval results from these queries were manually examined and the queries were adjusted when needed. On average 30mins was spent on each topic. Standard stopwords were removed from the queries. Language model retrieval model with Dirichlet mu=7000.

CMUTStncs¶

Run ID: CMUTStncs
Participant: CMU.LIRA
Track: Chemical
Year: 2011
Submission: 8/5/2011
Type: automatic
Task: techsurv
MD5: 091fa667e16da3644dac2efaa4995cb6
Run description: Automatically created queries, aggregating all words from title, narrative and details fields. Standard stopwords were removed from the queries. Language model retrieval model with Dirichlet mu=7000.

CMUTStncws¶

Run ID: CMUTStncws
Participant: CMU.LIRA
Track: Chemical
Year: 2011
Submission: 8/7/2011
Type: automatic
Task: techsurv
MD5: af6fee8eb36395e3c0ddd758f7607703
Run description: Automatically created queries, weighting words from title by 0.3, words from narrative by 0.6 and words from details by 0.1. Standard stopwords were removed from the queries. Language model retrieval model with Dirichlet mu=7000.

DUT11PARun1¶

Run ID: DUT11PARun1
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart_sm
MD5: 43d108f9653f5c24e77bcb1d3ab53518
Run description: In DUT11PARun1, we only use query-likelihood model as our retrieval model, and use title field as query field, this run is our baseline.

DUT11PARun2¶

Run ID: DUT11PARun2
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart_sm
MD5: 5f8109bae933b7a2017ae2a1e7f9d98a
Run description: The DUT11PARun2 which is based on DUT11PARun1, is also adopt the query-likelihood model as retrieval model, but it uses the phrase expansion(PE) construct method to construct query. The PE method considers the relative sequency and relative distance of the query terms.

DUT11PARun3¶

Run ID: DUT11PARun3
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart_sm
MD5: 16a7cc2ca9cebbb38280939497d6db67
Run description: The DUT11PARun3 which is based on DUT11PARun2. However, it also use the query expansion method to get the final result.

DUT11PARun4¶

Run ID: DUT11PARun4
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart_sm
MD5: a67c99d82895c0dfdea3c8d944500e25
Run description: In DUT11PARun4, we use title field, description field and claims field of the query to contrust our new query, we select the terms that have high frequency as our new query terms. In addition, we also use date filter to obtain our final result.

DUT11PARun5¶

Run ID: DUT11PARun5
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart_sm
MD5: b2b09aaa0cb2806253db92652388e912
Run description: The DUT11PARun5, just likes the DUT11PARun4, but in the process of select new query terms, we use a different weighting method with respect to DUT11PARun4. We use the tf-idf weight model to select the most important terms to obtain our final query terms and use indri search engine to get our result.

DUT11PARun6¶

Run ID: DUT11PARun6
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart_sm
MD5: 896bf91616e44e8c549e88ba9d7445ec
Run description: In DUT11PARun6, we only use title field of query, and use In_expB2 model as our retrieval model.

DUT11PARun7¶

Run ID: DUT11PARun7
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart_sm
MD5: e4b4dae0500981cb080c5dd99879775f
Run description: In this run, we also use tf-idf weighting model to select our query terms and use title,description and claims field in query. After that, we use In_expB2 retrieval model to get our result.

DUT11PARun8¶

Run ID: DUT11PARun8
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart_sm
MD5: d5e20e4cd4d477f9ee68cbaa406fcee3
Run description: Based on DUT11PARun6, this run also uses In_expB2 retrievl model. In addition, it also use the query expansion method to get further result.

DUT11PARun9¶

Run ID: DUT11PARun9
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 8/31/2011
Task: priorart_sm
MD5: 92331fc3eb83a1a471d04f22f944ff49
Run description: The DUT11PARun9, is a reranking process to DUT11PARun5. That is, we use the citation relationship of each result patent to score the result again.

DUT11TSRun1¶

Run ID: DUT11TSRun1
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 7/27/2011
Type: automatic
Task: techsurv
MD5: 7edcfa0311a24d74091fde74a74a5f17
Run description: DUT11TSRun1 uses the query-likelihood retrieval model and only searches title field in query. Additional, we use the concept of 'phrase expansion' proposed in [Indri at TREC 2004:Terabyte Track] to construct our query.

DUT11TSRun2¶

Run ID: DUT11TSRun2
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 7/27/2011
Type: automatic
Task: techsurv
MD5: 2cf4b6406d4c8767778fd72f40961d2c
Run description: DUT11TSRun2 also employs the query-likelihood retrieval model,and uses the phrase expansion method to construct our query. In addition, we also use query expansion after we get the result.

DUT11TSRun3¶

Run ID: DUT11TSRun3
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 7/27/2011
Type: automatic
Task: techsurv
MD5: 44ae7a44e083e90801b6fdece4069931
Run description: In DUT11TSRun3, we only use title field in the query,and use DFR model as our retrieval model.we only search the description field in the document collection.

DUT11TSRun4¶

Run ID: DUT11TSRun4
Participant: DUTIR
Track: Chemical
Year: 2011
Submission: 7/27/2011
Type: automatic
Task: techsurv
MD5: 977a14dcf2da35742a889887561dac92
Run description: In this run, we also employ the DFR retrieval model, and only use the title field in the query. Besides, query expansion is used after we get the retrieval result and we use description field as the expansion field.

GGA_Imago2011a¶

Run ID: GGA_Imago2011a
Participant: GGA
Track: Chemical
Year: 2011
Submission: 8/25/2011
Task: image2struct
Run description: Molfiles were generated from the set of test images using command line interface for Imago recognition tool (http://ggasoftware.com/opensource/imago). After that each file was automatically processed with Indigo toolkit for better results. Most abbreviations are expanded.

GGA_Imago_2011¶

Run ID: GGA_Imago_2011
Participant: GGA
Track: Chemical
Year: 2011
Submission: 8/25/2011
Task: image2struct
Run description: Molfiles were generated from the set of test images using command line interface for Imago recognition tool (http://ggasoftware.com/opensource/imago). After that each file was automatically processed with Indigo toolkit for better results. Abbreviations are not expanded in this set.

osra_1.3.8_def¶

Run ID: osra_1.3.8_def
Participant: SAIC_Frederick
Track: Chemical
Year: 2011
Submission: 7/26/2011
Task: image2struct
Run description: OSRA 1.3.8 run with default options (and -f sdf for SDF format).

osra_1.3.8_r300¶

Run ID: osra_1.3.8_r300
Participant: SAIC_Frederick
Track: Chemical
Year: 2011
Submission: 7/26/2011
Task: image2struct
Run description: OSRA-1.3.8 with command line options "-r 300 -f sdf" (try only resolution of 300 dpi, produce SDF formatted output)

Run1¶

Run ID: Run1
Participant: UoB
Track: Chemical
Year: 2011
Submission: 8/7/2011
Task: image2struct
Run description: This is the first run for the I2S task. Team is University of Birmingham, UK. Team ID is UoB. MOL-files are named according to their corresponding images.

Run2¶

Run ID: Run2
Participant: UoB
Track: Chemical
Year: 2011
Submission: 8/7/2011
Task: image2struct
Run description: This is the second run for the I2S task. Team is University of Birmingham, UK. Team ID is UoB. MOL-files are named according to their corresponding images.

SCAIRUN1¶

Run ID: SCAIRUN1
Participant: fraunhofer.scai
Track: Chemical
Year: 2011
Submission: 8/1/2011
Type: automatic
Task: techsurv
MD5: f76ddc9796bcd68b9d30bf686eceb330
Run description: Keyword search in Patent and Full-text documents

SCAIRUN2¶

Run ID: SCAIRUN2
Participant: fraunhofer.scai
Track: Chemical
Year: 2011
Submission: 8/1/2011
Type: automatic
Task: techsurv
MD5: 988e24a5b9dd990feeb2e827777a0059
Run description: Semantic search in Patent and Full-text documents

SCAIRUN3¶

Run ID: SCAIRUN3
Participant: fraunhofer.scai
Track: Chemical
Year: 2011
Submission: 8/1/2011
Type: automatic
Task: techsurv
MD5: 026703a74666a06a911c556803189b09
Run description: Semantic search in Patent and Full-text documents with co-citation based re-ranking

SCAIRUN4¶

Run ID: SCAIRUN4
Participant: fraunhofer.scai
Track: Chemical
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: techsurv
MD5: 3a88b483a5457e34f98a28bd0b89798e
Run description: Keyword + semantic search in patent and full-text