Runs - Session 2011¶

CWIpostRW.RL1¶

Run ID: CWIpostRW.RL1
Participant: CWI
Track: Session
Year: 2011
Submission: 8/2/2011
Type: automatic
Task: RL1
MD5: 3ca380009a3552679c96fa7da083ecca
Run description: The previous queries and clicks are used to retrieve related queries using a random walk on a graph consisting of queries and clicks submitted to a major search engine. The related queries are used to rerank the documents retrieved by the original query.

CWIpostRW.RL2¶

Run ID: CWIpostRW.RL2
Participant: CWI
Track: Session
Year: 2011
Submission: 8/2/2011
Type: automatic
Task: RL2
MD5: 08beac7e1263a00650af33fb6649e21b
Run description: The previous queries and clicks are used to retrieve related queries using a random walk on a graph consisting of queries and clicks submitted to a major search engine. The related queries are used to rerank the documents retrieved by the original query.

CWIpostRW.RL3¶

Run ID: CWIpostRW.RL3
Participant: CWI
Track: Session
Year: 2011
Submission: 8/2/2011
Type: automatic
Task: RL3
MD5: 08beac7e1263a00650af33fb6649e21b
Run description: The previous queries and clicks are used to retrieve related queries using a random walk on a graph consisting of queries and clicks submitted to a major search engine. The related queries are used to rerank the documents retrieved by the original query.

CWIpostRW.RL4¶

Run ID: CWIpostRW.RL4
Participant: CWI
Track: Session
Year: 2011
Submission: 8/2/2011
Type: automatic
Task: RL4
MD5: 11c1639e915f75e2641b29e2bc18f675
Run description: The previous queries and clicks are used to retrieve related queries using a random walk on a graph consisting of queries and clicks submitted to a major search engine. The related queries are used to rerank the documents retrieved by the original query.

CWIrun1.RL1¶

Run ID: CWIrun1.RL1
Participant: CWI
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 7261eee2362bd24847e6a53bd9884d1f
Run description: static user modeling

CWIrun1.RL2¶

Run ID: CWIrun1.RL2
Participant: CWI
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: c005ca325b7f8a2ba31e22e4e8382471
Run description: static user modeling

CWIrun1.RL3¶

Run ID: CWIrun1.RL3
Participant: CWI
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 6c225edc7ef833472b6994c75c45ae5c
Run description: static user modeling

CWIrun1.RL4¶

Run ID: CWIrun1.RL4
Participant: CWI
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: 6c225edc7ef833472b6994c75c45ae5c
Run description: static user modeling

CWIrun2.RL1¶

Run ID: CWIrun2.RL1
Participant: CWI
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: c68c605c436e3a17036dc1ca18d9c454
Run description: dynamic user modeling

CWIrun2.RL2¶

Run ID: CWIrun2.RL2
Participant: CWI
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: e9ba4f58a28d0f5394a689a41466d58b
Run description: dynamic user modeling

CWIrun2.RL3¶

Run ID: CWIrun2.RL3
Participant: CWI
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 7fb04532cb7bf944b7074f6b2c8fdfd5
Run description: dynamic user modelling

CWIrun2.RL4¶

Run ID: CWIrun2.RL4
Participant: CWI
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: 87944235d0ecac8227a407ca9f97baab
Run description: dynamic user modelling

DUTIR2011A.RL1¶

Run ID: DUTIR2011A.RL1
Participant: DUTIR
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 60fdd38fa2626dfc75fe7c3e64c90004
Run description: rl1

DUTIR2011A.RL2¶

Run ID: DUTIR2011A.RL2
Participant: DUTIR
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: 54778b46377342fa872a025870b7d028
Run description: rl2

DUTIR2011A.RL3¶

Run ID: DUTIR2011A.RL3
Participant: DUTIR
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: b1ac7df2ca11b01eb3a2e482a7a0103a
Run description: rl3

DUTIR2011A.RL4¶

Run ID: DUTIR2011A.RL4
Participant: DUTIR
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: dc0b20dac61bae133dd989ae13ff2565
Run description: rl4

essexAnchor.RL1¶

Run ID: essexAnchor.RL1
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 6e09b1b867802567929a632a850c3f2a
Run description: RL1 is simply generated by submitting the current query to the Indri search engine.

essexAnchor.RL2¶

Run ID: essexAnchor.RL2
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: 7901b2949bdac7fb49420a8f9cba8c31
Run description: The approach we propose here can be seen as extension to our anchor log technique proposed in the previous year. To generate RL2 we use a similar approach to the one used in our anchor log system last year `essex3'. We consider the first query and the last query in the session q1, qm and use the same method describe in our last year report to generate query expansions from the anchor log to the reformulated query qm. For RL3, we expand the reformulated query qm with the anchor text of the displayed documents. We take up to ten query phrases from the anchor text weighted by their normalised frequency. For RL4, like in RL3, we follow the same procedure but consider only the anchor text of the clicked documents.

essexAnchor.RL3¶

Run ID: essexAnchor.RL3
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 61048252c60d29e7b41f9fa91758ffa7
Run description: The approach we propose here can be seen as extension to our anchor log technique proposed in the previous year. To generate RL2 we use a similar approach to the one used in our anchor log system last year `essex3'. We consider the first query and the last query in the session q1, qm and use the same method describe in our last year report to generate query expansions from the anchor log to the reformulated query qm. For RL3, we expand the reformulated query qm with the anchor text of the displayed documents. We take up to ten query phrases from the anchor text weighted by their normalised frequency. For RL4, like in RL3, we follow the same procedure but consider only the anchor text of the clicked documents.

essexAnchor.RL4¶

Run ID: essexAnchor.RL4
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: ae58a2c9949848c9f36d93f4951c6722
Run description: The approach we propose here can be seen as extension to our anchor log technique proposed in the previous year. To generate RL2 we use a similar approach to the one used in our anchor log system last year `essex3'. We consider the first query and the last query in the session q1, qm and use the same method describe in our last year report to generate query expansions from the anchor log to the reformulated query qm. For RL3, we expand the reformulated query qm with the anchor text of the displayed documents. We take up to ten query phrases from the anchor text weighted by their normalised frequency. For RL4, like in RL3, we follow the same procedure but consider only the anchor text of the clicked documents.

essexNooNeg.RL1¶

Run ID: essexNooNeg.RL1
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 6363d2ee8754cb5c091d46cb6e8cca1b
Run description: Nootropia is a biologically inspired Information Filtering system that has been used successfully in news recommendation. Throughout the session we take the documents that has been displayed to the user by the search engine to create a Nootropia model. For the last query, the current query, we submit the query to the search engine and then we use the so far built profile in a form of a Nootropia network to give a score the returned documents reflecting how much they match the profile. The scores can be used to re-rank the documents. To generate RL3, we can use all the displayed documents each time build a Nootropia profile and update it. Here we take an optimistic approach and consider displayed documents as useful documents to the user. The documents that match this profile should be promoted. Therefore we rank the documents returned by the search engine for the current query by their Nootropia score in an descending order and that would be our RL3. For RL4, we take only the clicked documents to build a profile of user interests and use this profile to re-rank the documents. To feed a document into Nootropia, we perform NLP to extract all the noun phrases in the snippet of the document and consider these as representative features of the document that can be fed in or evaluated by the Nootropia network. For RL1 we simply submit the current query to the search engine and for RL2 we submit a concatenation of the first query in the session and the current query to the search engine. The Indri search engine was used in this run.

essexNooNeg.RL2¶

Run ID: essexNooNeg.RL2
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: f16d2ab2e0bf42377be2baa86f65cfc2
Run description: Nootropia is a biologically inspired Information Filtering system that has been used successfully in news recommendation. Throughout the session we take the documents that has been displayed to the user by the search engine to create a Nootropia model. Documents are represented by taking their snippets and peforming NLP to extract all the noun phrases in the snippets and feed that into the Nootropia network. For the last query, the current query, we submit the query to the search engine and then we use the so far built profile in a form of a Nootropia network to give a score the returned documents reflecting how much they matches the profile. The scores can be used to re-rank the documents. To generate RL3, we can use all the displayed documents each time build a Nootropia profile and update it. Here we take a pessimistic approach and consider the displayed documents not useful to the user as they had to reformulate the query and therefore any document that matches this profile should be penalised. Therefore we rank the documents returned by the search engine for the current query by their Nootropia score in an ascending order and that would be our RL3. As for RL4, we take only the abandoned documents to build a profile of documents that the user is not interested in. The documents that match this profile should be penalised. Therefore, we rank the documents returned by the search engine for the current query by their Nootropia score in an ascending order and that would be our RL4. To feed a document into Nootropia, we perform NLP to extract all the noun phrases in the snippet of the document and consider these as representative features of the document that can be fed in or evaluated by the Nootropia network. For RL1 we simply submit the current query to the search engine and for RL2 we submit a concatenation of the first query in the session and the current query to the search engine. The Indri search engine was used in this run.

essexNooNeg.RL3¶

Run ID: essexNooNeg.RL3
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 68b1fd0a6a9e889f989864b4d7e0fd67
Run description: Nootropia is a biologically inspired Information Filtering system that has been used successfully in news recommendation. Throughout the session we take the documents that has been displayed to the user by the search engine to create a Nootropia model. Documents are represented by taking their snippets and peforming NLP to extract all the noun phrases in the snippets and feed that into the Nootropia network. For the last query, the current query, we submit the query to the search engine and then we use the so far built profile in a form of a Nootropia network to give a score the returned documents reflecting how much they matches the profile. The scores can be used to re-rank the documents. To generate RL3, we can use all the displayed documents each time build a Nootropia profile and update it. Here we take a pessimistic approach and consider the displayed documents not useful to the user as they had to reformulate the query and therefore any document that matches this profile should be penalised. Therefore we rank the documents returned by the search engine for the current query by their Nootropia score in an ascending order and that would be our RL3. As for RL4, we take only the abandoned documents to build a profile of documents that the user is not interested in. The documents that match this profile should be penalised. Therefore, we rank the documents returned by the search engine for the current query by their Nootropia score in an ascending order and that would be our RL4. To feed a document into Nootropia, we perform NLP to extract all the noun phrases in the snippet of the document and consider these as representative features of the document that can be fed in or evaluated by the Nootropia network. For RL1 we simply submit the current query to the search engine and for RL2 we submit a concatenation of the first query in the session and the current query to the search engine. The Indri search engine was used in this run.

essexNooNeg.RL4¶

Run ID: essexNooNeg.RL4
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: b34481c423fedeffa88a703f8a92d534
Run description: Nootropia is a biologically inspired Information Filtering system that has been used successfully in news recommendation. Throughout the session we take the documents that has been displayed to the user by the search engine to create a Nootropia model. For the last query, the current query, we submit the query to the search engine and then we use the so far built profile in a form of a Nootropia network to give a score the returned documents reflecting how much they matches the profile. The scores can be used to re-rank the documents. To generate RL3, we can use all the displayed documents each time build a Nootropia profile and update it. Here we take a pessimistic approach and consider the displayed documents not useful to the user as they had to reformulate the query and therefore any document that matches this profile should be penalised. Therefore we rank the documents returned by the search engine for the current query by their Nootropia score in an ascending order and that would be our RL3. As for RL4, we take only the abandoned documents to build a profile of documents that the user is not interested in. The documents that match this profile should be penalised. Therefore, we rank the documents returned by the search engine for the current query by their Nootropia score in an ascending order and that would be our RL4. To feed a document into Nootropia, we perform NLP to extract all the noun phrases in the snippet of the document and consider these as representative features of the document that can be fed in or evaluated by the Nootropia network. For RL1 we simply submit the current query to the search engine and for RL2 we submit a concatenation of the first query in the session and the current query to the search engine. The Indri search engine was used in this run.

essexNooPos.RL1¶

Run ID: essexNooPos.RL1
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: da6d396dfda8aa4620c8755c2be77748
Run description: Nootropia is a biologically inspired Information Filtering system that has been used successfully in news recommendation. Throughout the session we take the documents that has been displayed to the user by the search engine to create a Nootropia model. For the last query, the current query, we submit the query to the search engine and then we use the so far built profile in a form of a Nootropia network to give a score the returned documents reflecting how much they match the profile. The scores can be used to re-rank the documents. To generate RL3, we can use all the displayed documents each time build a Nootropia profile and update it. Here we take an optimistic approach and consider displayed documents as useful documents to the user. The documents that match this profile should be promoted. Therefore we rank the documents returned by the search engine for the current query by their Nootropia score in an descending order and that would be our RL3. For RL4, we take only the clicked documents to build a profile of user interests and use this profile to re-rank the documents. To feed a document into Nootropia, we peform NLP to extract all the noun phrases in the snippet of the document and consider these as reprsentative features of the document that can be fed in or evaluated by the Nootropia network. For RL1 we simply submit the current query to the search engine and for RL2 we submit a concatenation of the first query in the session and the current query to the search engine. The Indri search engine was used in this run.

essexNooPos.RL2¶

Run ID: essexNooPos.RL2
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: 47a3fa9c721af389d1b7654206d97396
Run description: Nootropia is a biologically inspired Information Filtering system that has been used successfully in news recommendation. Throughout the session we take the documents that has been displayed to the user by the search engine to create a Nootropia model. For the last query, the current query, we submit the query to the search engine and then we use the so far built profile in a form of a Nootropia network to give a score the returned documents reflecting how much they match the profile. The scores can be used to re-rank the documents. To generate RL3, we can use all the displayed documents each time build a Nootropia profile and update it. Here we take an optimistic approach and consider displayed documents as useful documents to the user. The documents that match this profile should be promoted. Therefore we rank the documents returned by the search engine for the current query by their Nootropia score in an descending order and that would be our RL3. For RL4, we take only the clicked documents to build a profile of user interests and use this profile to re-rank the documents. To feed a document into Nootropia, we peform NLP to extract all the noun phrases in the snippet of the document and consider these as reprsentative features of the document that can be fed in or evaluated by the Nootropia network. For RL1 we simply submit the current query to the search engine and for RL2 we submit a concatenation of the first query in the session and the current query to the search engine. The Indri search engine was used in this run.

essexNooPos.RL3¶

Run ID: essexNooPos.RL3
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: b6bfb886cc98fa14dda2aff094c354fc
Run description: Nootropia is a biologically inspired Information Filtering system that has been used successfully in news recommendation. Throughout the session we take the documents that has been displayed to the user by the search engine to create a Nootropia model. For the last query, the current query, we submit the query to the search engine and then we use the so far built profile in a form of a Nootropia network to give a score the returned documents reflecting how much they match the profile. The scores can be used to re-rank the documents. To generate RL3, we can use all the displayed documents each time build a Nootropia profile and update it. Here we take an optimistic approach and consider displayed documents as useful documents to the user. The documents that match this profile should be promoted. Therefore we rank the documents returned by the search engine for the current query by their Nootropia score in an descending order and that would be our RL3. For RL4, we take only the clicked documents to build a profile of user interests and use this profile to re-rank the documents. To feed a document into Nootropia, we peform NLP to extract all the noun phrases in the snippet of the document and consider these as reprsentative features of the document that can be fed in or evaluated by the Nootropia network. For RL1 we simply submit the current query to the search engine and for RL2 we submit a concatenation of the first query in the session and the current query to the search engine. The Indri search engine was used in this run.

essexNooPos.RL4¶

Run ID: essexNooPos.RL4
Participant: essexuni
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: 5a274ffbc35a0f3fd370e5cc0ed9bdc2
Run description: Nootropia is a biologically inspired Information Filtering system that has been used successfully in news recommendation. Throughout the session we take the documents that has been displayed to the user by the search engine to create a Nootropia model. For the last query, the current query, we submit the query to the search engine and then we use the so far built profile in a form of a Nootropia network to give a score the returned documents reflecting how much they match the profile. The scores can be used to re-rank the documents. To generate RL3, we can use all the displayed documents each time build a Nootropia profile and update it. Here we take an optimistic approach and consider displayed documents as useful documents to the user. The documents that match this profile should be promoted. Therefore we rank the documents returned by the search engine for the current query by their Nootropia score in an descending order and that would be our RL3. For RL4, we take only the clicked documents to build a profile of user interests and use this profile to re-rank the documents. To feed a document into Nootropia, we peform NLP to extract all the noun phrases in the snippet of the document and consider these as reprsentative features of the document that can be fed in or evaluated by the Nootropia network. For RL1 we simply submit the current query to the search engine and for RL2 we submit a concatenation of the first query in the session and the current query to the search engine. The Indri search engine was used in this run.

ICTNET11SER1.RL1¶

Run ID: ICTNET11SER1.RL1
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL1
MD5: b3dded97c4ecfe67682fd4b399ef151a
Run description: apply bm25 on content field.

ICTNET11SER1.RL2¶

Run ID: ICTNET11SER1.RL2
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL2
MD5: 9fd24e573b0754dba1aa6eac310cb5bf
Run description: Make a intersection from the previous query lists to optimize the result

ICTNET11SER1.RL3¶

Run ID: ICTNET11SER1.RL3
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL3
MD5: 77667a6865f4150465fa95b2d9b159c4
Run description: Use the snippet information from xml file to sort the data,then repeat using the intersection.

ICTNET11SER1.RL4¶

Run ID: ICTNET11SER1.RL4
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL4
MD5: 77667a6865f4150465fa95b2d9b159c4
Run description: Use the clicked list data to sort the data from the third data.

ICTNET11SER2.RL1¶

Run ID: ICTNET11SER2.RL1
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL1
MD5: 3110d28437a5a0d67bc5431d9a082f4b
Run description: The data is the same with the RL1 data with ICTNET11SER2,we apply bm25 on content field.

ICTNET11SER2.RL2¶

Run ID: ICTNET11SER2.RL2
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL2
MD5: 1a5e0dcb854bd9e9551c651cb1b5a8b8
Run description: Use the session type data and the intersection data to optimization.

ICTNET11SER2.RL3¶

Run ID: ICTNET11SER2.RL3
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL3
MD5: 3110d28437a5a0d67bc5431d9a082f4b
Run description: The method is similar with this using in ICTNET11SER1,but different parameters.

ICTNET11SER2.RL4¶

Run ID: ICTNET11SER2.RL4
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL4
MD5: 3110d28437a5a0d67bc5431d9a082f4b
Run description: The method is similar with this using in ICTNET11SER1,but different parameters.

ICTNET11SER3.RL1¶

Run ID: ICTNET11SER3.RL1
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/30/2011
Type: automatic
Task: RL1
MD5: 34050033a8b4a90ae787e30d61ec881b
Run description: Apply bm25 on content field.

ICTNET11SER3.RL2¶

Run ID: ICTNET11SER3.RL2
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/30/2011
Type: automatic
Task: RL2
MD5: f9d5693e459a57ea3ec3d6c0d0e2f923
Run description: Using Session Type method and intersection ranking method to optimize the result.

ICTNET11SER3.RL3¶

Run ID: ICTNET11SER3.RL3
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/30/2011
Type: automatic
Task: RL3
MD5: 34050033a8b4a90ae787e30d61ec881b
Run description: Using the snippet information from the xml file to classify the primary results from our own search enginethen repeat the session type and intersection rank method to optmize the result further.

ICTNET11SER3.RL4¶

Run ID: ICTNET11SER3.RL4
Participant: ICTNET
Track: Session
Year: 2011
Submission: 7/30/2011
Type: automatic
Task: RL4
MD5: 34050033a8b4a90ae787e30d61ec881b
Run description: Firstly we use the description and narration information of each query to compute match score with content . Then repeat the procedures based on the third Run.

PITTSIS.RL1¶

Run ID: PITTSIS.RL1
Participant: PITTSIS
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: 1f3425d1492cced03dbb092914bea3cb
Run description: RL1 result only uses current query. We apply query likelihood language model and tuned sequential dependence model to current query. Documents with lower than 70 waterloo spam filter score are filtered.

PITTSIS.RL2¶

Run ID: PITTSIS.RL2
Participant: PITTSIS
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: a44fc6c65f162d98ea265011dfb94b63
Run description: RL2 result uses current query and previous queries. We estimate query term necessity estimation based on previous queries in order to weight query terms and phrases (ordered and unordered). Documents with lower than 70 waterloo spam filter score are filtered.

PITTSIS.RL3¶

Run ID: PITTSIS.RL3
Participant: PITTSIS
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: c1c30b314e04385c0bb6c6761a98e3fb
Run description: RL3 result further weight query terms and phrases using pseudo relevance feedback documents of current query, and mixed with RL2's term necessity weight.

PITTSIS.RL4¶

Run ID: PITTSIS.RL4
Participant: PITTSIS
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 17e24809f185c9b90991f234b2699074
Run description: Instead of using pseudo relevance feedback documents for weighting terms and phrases, RL3 uses clicked results of all previous queries for weighting, and mixed with RL2's term necessity weight.

Rgposneg.RL1¶

Run ID: Rgposneg.RL1
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: d201ab55d230644db0aeb261d85bdaf5
Run description: baselinerun with pseudoRF

Rgposneg.RL2¶

Run ID: Rgposneg.RL2
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: ba702f72df2971b7631d690f278cdd95
Run description: the run which takes account of query characteristics only

Rgposneg.RL3¶

Run ID: Rgposneg.RL3
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 9f8e306a203d5e298c433dbf65e28ce4
Run description: the run which takes account of URL-by-query evidence only

Rgposneg.RL4¶

Run ID: Rgposneg.RL4
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: bde30b67ca69ddea7b38847eb57d9f7a
Run description: the general model, with positive and negative feedback, no task type considered.

rguBase.RL1¶

Run ID: rguBase.RL1
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: 3dfdd47b1fefd52bf2e56c2287fc4866
Run description: This is our baseline system. It is based on the observation that using the term 'wikipedia' to expand the query resulted in a better retrieval performance in last year's task (session track 2010). For RL1 we just submit the current query to the Indri search engine. For RL2,RL3,RL4 we expand the current query with the term wikipedia.

rguBase.RL2¶

Run ID: rguBase.RL2
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: 69d99070eff63195e987ac8794e09818
Run description: This is our baseline system. It is based on the observation that using the term 'wikipedia' to expand the query resulted in a better retrieval performance in last year's task (session track 2010). For RL1 we just submit the current query to the Indri search engine. For RL2,RL3,RL4 we expand the current query with the term wikipedia.

rguBase.RL3¶

Run ID: rguBase.RL3
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: 4b3657bb8fe5f7c85df14fde79970a46
Run description: This is our baseline system. It is based on the observation that using the term 'wikipedia' to expand the query resulted in a better retrieval performance in last year's task (session track 2010). For RL1 we just submit the current query to the Indri search engine. For RL2,RL3,RL4 we expand the current query with the term wikipedia.

rguBase.RL4¶

Run ID: rguBase.RL4
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 4b3657bb8fe5f7c85df14fde79970a46
Run description: This is our baseline system. It is based on the observation that using the term 'wikipedia' to expand the query resulted in a better retrieval performance in last year's task (session track 2010). For RL1 we just submit the current query to the Indri search engine. For RL2,RL3,RL4 we expand the current query with the term wikipedia.

rguPisaSS.RL1¶

Run ID: rguPisaSS.RL1
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 950a2e2f756b339327fb72a0c69ef8d7
Run description: These runs are obtained by using the SeachShortcuts (Generating Suggestions for Queries in the Long Tail with an Inverted Index, Broccolo et al., IP&M, to appear) recommender system. The SearchShortcuts technique uses an inverted index and the concept of satisfactory sessions present in Web search engines query log in order to produce effective recommendations for both frequent and rare/unseen queries. We adapt the above technique to work as a query expansion tool. We use such expansion tool to expand the TREC queries (expansion is generated by using a method aiming at considering also the past queries in the session). The expansion terms obtained are then used to build a global, uniformly weighted, representation of the user session (RL2). Furthermore, the expansion terms are then combined with ranked lists of results in order to boost terms appearing more frequently in the final results lists (RL3). Finally, we also integrate dwell times and the weighting method obtained takes both result lists and clicks into account for assigning weights to the terms expanding the final query of the session.

rguPisaSS.RL2¶

Run ID: rguPisaSS.RL2
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: 0d65f9e024947176297956396400cce5
Run description: These runs are obtained by using the SeachShortcuts (Generating Suggestions for Queries in the Long Tail with an Inverted Index, Broccolo et al., IP&M, to appear) recommender system. The SearchShortcuts technique uses an inverted index and the concept of satisfactory sessions present in Web search engines query log in order to produce effective recommendations for both frequent and rare/unseen queries. We adapt the above technique to work as a query expansion tool. We use such expansion tool to expand the TREC queries (expansion is generated by using a method aiming at considering also the past queries in the session). The expansion terms obtained are then used to build a global, uniformly weighted, representation of the user session (RL2). Furthermore, the expansion terms are then combined with ranked lists of results in order to boost terms appearing more frequently in the final results lists (RL3). Finally, we also integrate dwell times and the weighting method obtained takes both result lists and clicks into account for assigning weights to the terms expanding the final query of the session.

rguPisaSS.RL3¶

Run ID: rguPisaSS.RL3
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 2ca42373916dc0702b9fe91764c07028
Run description: These runs are obtained by using the SeachShortcuts (Generating Suggestions for Queries in the Long Tail with an Inverted Index, Broccolo et al., IP&M, to appear) recommender system. The SearchShortcuts technique uses an inverted index and the concept of satisfactory sessions present in Web search engines query log in order to produce effective recommendations for both frequent and rare/unseen queries. We adapt the above technique to work as a query expansion tool. We use such expansion tool to expand the TREC queries (expansion is generated by using a method aiming at considering also the past queries in the session). The expansion terms obtained are then used to build a global, uniformly weighted, representation of the user session (RL2). Furthermore, the expansion terms are then combined with ranked lists of results in order to boost terms appearing more frequently in the final results lists (RL3). Finally, we also integrate dwell times and the weighting method obtained takes both result lists and clicks into account for assigning weights to the terms expanding the final query of the session.

rguPisaSS.RL4¶

Run ID: rguPisaSS.RL4
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: fa0103756beb216533eeaed3fdccaf23
Run description: These runs are obtained by using the SeachShortcuts (Generating Suggestions for Queries in the Long Tail with an Inverted Index, Broccolo et al., IP&M, to appear) recommender system. The SearchShortcuts technique uses an inverted index and the concept of satisfactory sessions present in Web search engines query log in order to produce effective recommendations for both frequent and rare/unseen queries. We adapt the above technique to work as a query expansion tool. We use such expansion tool to expand the TREC queries (expansion is generated by using a method aiming at considering also the past queries in the session). The expansion terms obtained are then used to build a global, uniformly weighted, representation of the user session (RL2). Furthermore, the expansion terms are then combined with ranked lists of results in order to boost terms appearing more frequently in the final results lists (RL3). Finally, we also integrate dwell times and the weighting method obtained takes both result lists and clicks into account for assigning weights to the terms expanding the final query of the session.

rguPisaSST.RL1¶

Run ID: rguPisaSST.RL1
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 9a25d56ced7e6bdb071c6db47fc6df11
Run description: These runs are obtained by using the SeachShortcuts (Generating Suggestions for Queries in the Long Tail with an Inverted Index, Broccolo et al., IP&M, to appear) recommender system. The SearchShortcuts technique uses an inverted index and the concept of satisfactory sessions present in Web search engines query log in order to produce effective recommendations for both frequent and rare/unseen queries. We adapt the above technique to work as a query expansion tool. We use such expansion tool to expand the TREC queries (expansion is generated by using a method aiming at considering also the past queries in the session). The expansion terms obtained are then used to build a global, uniformly weighted, representation of the user session (RL2). Furthermore, the expansion terms are then combined with ranked lists of results in order to boost terms appearing more frequently in the final results lists (RL3). Finally, we also integrate dwell times and the weighting method obtained takes both result lists and clicks into account for assigning weights to the terms expanding the final query of the session.

rguPisaSST.RL2¶

Run ID: rguPisaSST.RL2
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: ab604e9dfe0f3321b8263839c871daf2
Run description: These runs are obtained by using the SeachShortcuts (Generating Suggestions for Queries in the Long Tail with an Inverted Index, Broccolo et al., IP&M, to appear) recommender system. The SearchShortcuts technique uses an inverted index and the concept of satisfactory sessions present in Web search engines query log in order to produce effective recommendations for both frequent and rare/unseen queries. We adapt the above technique to work as a query expansion tool. We use such expansion tool to expand the TREC queries (expansion is generated by using a method aiming at considering also the past queries in the session). The expansion terms obtained are then used to build a global, uniformly weighted, representation of the user session (RL2). Furthermore, the expansion terms are then combined with ranked lists of results in order to boost terms appearing more frequently in the final results lists (RL3). Finally, we also integrate dwell times and the weighting method obtained takes both result lists and clicks into account for assigning weights to the terms expanding the final query of the session.

rguPisaSST.RL3¶

Run ID: rguPisaSST.RL3
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 4b68d896147e0819ced49149c07f0576
Run description: These runs are obtained by using the SeachShortcuts (Generating Suggestions for Queries in the Long Tail with an Inverted Index, Broccolo et al., IP&M, to appear) recommender system. The SearchShortcuts technique uses an inverted index and the concept of satisfactory sessions present in Web search engines query log in order to produce effective recommendations for both frequent and rare/unseen queries. We adapt the above technique to work as a query expansion tool. We use such expansion tool to expand the TREC queries (expansion is generated by using a method aiming at considering also the past queries in the session). The expansion terms obtained are then used to build a global, uniformly weighted, representation of the user session (RL2). Furthermore, the expansion terms are then combined with ranked lists of results in order to boost terms appearing more frequently in the final results lists (RL3). Finally, we also integrate dwell times and the weighting method obtained takes both result lists and clicks into account for assigning weights to the terms expanding the final query of the session.

rguPisaSST.RL4¶

Run ID: rguPisaSST.RL4
Participant: RGU_AutoAdapt
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: 7624a595657788ef7b35e432e8a3edd0
Run description: These runs are obtained by using the SeachShortcuts (Generating Suggestions for Queries in the Long Tail with an Inverted Index, Broccolo et al., IP&M, to appear) recommender system. The SearchShortcuts technique uses an inverted index and the concept of satisfactory sessions present in Web search engines query log in order to produce effective recommendations for both frequent and rare/unseen queries. We adapt the above technique to work as a query expansion tool. We use such expansion tool to expand the TREC queries (expansion is generated by using a method aiming at considering also the past queries in the session). The expansion terms obtained are then used to build a global, uniformly weighted, representation of the user session (RL2). Furthermore, the expansion terms are then combined with ranked lists of results in order to boost terms appearing more frequently in the final results lists (RL3). Finally, we also integrate dwell times and the weighting method obtained takes both result lists and clicks into account for assigning weights to the terms expanding the final query of the session.

RMIT1.RL1¶

Run ID: RMIT1.RL1
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: 6c77e07e709551b5fceffae02a7c1272
Run description: Just using the current Query

RMIT1.RL2¶

Run ID: RMIT1.RL2
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: db146342a157d449e12b98d786ac2834
Run description: Just using the current Query and all queries in interactions

RMIT1.RL3¶

Run ID: RMIT1.RL3
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: 2cc4be3583f7e106bcb1c9a0b3dd1140
Run description: we are using query expansion from top five results in all interactions.

RMIT1.RL4¶

Run ID: RMIT1.RL4
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 0451368de950bd65739a514c183f38c2
Run description: We are using expanded queries from clicked document in all interactions.

RMIT2.RL1¶

Run ID: RMIT2.RL1
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: 06dc3edae23193f1eadc2a1442589549
Run description: Just using the current Query

RMIT2.RL2¶

Run ID: RMIT2.RL2
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: 01282066b304b72b5854cd492e276923
Run description: Just using all queries in interactions

RMIT2.RL3¶

Run ID: RMIT2.RL3
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: ce22b5f3822c41b70219956bc4c2db48
Run description: we are using query expansion from top five results in all interactions.

RMIT2.RL4¶

Run ID: RMIT2.RL4
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 85fd83ea7ea9fa22c12b1087370b4cf1
Run description: We are using expanded queries from clicked snippets in all interactions.

RMIT3.RL1¶

Run ID: RMIT3.RL1
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: 2f59f5e0176bc27873ca9798067e0489
Run description: Just using the current Query

RMIT3.RL2¶

Run ID: RMIT3.RL2
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: 518a07454e0946641be69cb44abaade4
Run description: Just using current query and all queries in interactions

RMIT3.RL3¶

Run ID: RMIT3.RL3
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: 8ad5e64d8cc62af4f5353ace783ff190
Run description: we are using query expansion from top five results in all interactions.

RMIT3.RL4¶

Run ID: RMIT3.RL4
Participant: RMIT
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 082569109c46b17ef42f08772dc03632
Run description: We are using expanded queries which is used clicked and non clicked results snippets.

Rspos.RL1¶

Run ID: Rspos.RL1
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: c7ededa303d456e96aceaddde121b1bf
Run description: baseline run, with pseudoRF

Rspos.RL2¶

Run ID: Rspos.RL2
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: 6702d29157baf376f1e48ea276d8f4d7
Run description: the run which takes account of query characteristics only

Rspos.RL3¶

Run ID: Rspos.RL3
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 632d2e79435f1c7dbe0d7555c74b5795
Run description: the run which takes account of URL-by-query evidence only

Rspos.RL4¶

Run ID: Rspos.RL4
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: manual
Task: RL4
MD5: 3260eedb9f38754a6a557048790b26eb
Run description: the specific model, with positive feedback only, manually classified tasks

Rsposneg.RL1¶

Run ID: Rsposneg.RL1
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 78c66593b8c955e5ef240d9a18263aed
Run description: baseline run, with pseudoRF

Rsposneg.RL2¶

Run ID: Rsposneg.RL2
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: 7242ee6d40b89ba7bede71309b345b34
Run description: the run which takes account of query characteristics only

Rsposneg.RL3¶

Run ID: Rsposneg.RL3
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 60ffeb658586e2ffd95608904f5ecb1d
Run description: the run which takes account of URL-by-query evidence only

Rsposneg.RL4¶

Run ID: Rsposneg.RL4
Participant: SCI_TREC_2011
Track: Session
Year: 2011
Submission: 8/3/2011
Type: manual
Task: RL4
MD5: 6a5bcaa87064dd1b397b5c99d8bec005
Run description: the specific model, with positive and negative feedback, manual classified tasks

udelASFe1new.RL1¶

Run ID: udelASFe1new.RL1
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: 57a6c19f3bea55a373d9902bbb429fb5
Run description: waterloo spam scores to filter. RL1: basic indri system RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using clicked docs

udelASFe1new.RL2¶

Run ID: udelASFe1new.RL2
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: c8b878947b1c9588ad66e3af60a220af
Run description: waterloo spam scores to filter. RL1: basic indri system RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using clicked docs

udelASFe1new.RL3¶

Run ID: udelASFe1new.RL3
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: 9fb83454fca56f80402f70a4604b6a23
Run description: waterloo spam scores to filter. RL1: basic indri system RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using clicked docs

udelASFe1new.RL4¶

Run ID: udelASFe1new.RL4
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 0fa59da3940d7d2832fe5805a434f0ea
Run description: waterloo spam scores to filter. RL1: basic indri system RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using clicked docs

udelBe2.RL1¶

Run ID: udelBe2.RL1
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: bba164cb219a544bc67b4febc21683fe
Run description: waterloo spam scores to filter. RL1: basic indri system RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using unclicked docs---note that this one is cat A and should be compared to ASF.RL4

udelBe2.RL2¶

Run ID: udelBe2.RL2
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: 004c0703be6de6f5e6498f159f2aadf2
Run description: waterloo spam scores to filter. RL1: basic indri system RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using unclicked docs---note that this one is cat A and should be compared to ASF.RL4

udelBe2.RL3¶

Run ID: udelBe2.RL3
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: 6e76794dd6d8d45bb90b0bb580222949
Run description: waterloo spam scores to filter. RL1: basic indri system RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using unclicked docs---note that this one is cat A and should be compared to ASF.RL4

udelBe2.RL4¶

Run ID: udelBe2.RL4
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 44cb6be30c7bf7bc31db2c99b9daad7d
Run description: waterloo spam scores to filter. RL1: basic indri system on WP RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using unclicked docs---note that this one is cat A and should be compared to ASF.RL4

udelWPmnz.RL1¶

Run ID: udelWPmnz.RL1
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: 1eb905b076817e8330a6957a0ccfb8b6
Run description: waterloo spam scores to filter. RL1: basic indri system on WP RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using all docs, or maybe using combMNZ to combine the other two RL4s---note that this one is cat A and should be compared to ASF.RL4

udelWPmnz.RL2¶

Run ID: udelWPmnz.RL2
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: 3a4b62c7ee2d5404c419b3bffad8e6ea
Run description: waterloo spam scores to filter. RL1: basic indri system on WP RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using all docs, or maybe using combMNZ to combine the other two RL4s---note that this one is cat A and should be compared to ASF.RL4

udelWPmnz.RL3¶

Run ID: udelWPmnz.RL3
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: a747bad6fb50b652e5eaf7de2969af40
Run description: waterloo spam scores to filter. RL1: basic indri system on WP RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using all docs, or maybe using combMNZ to combine the other two RL4s---note that this one is cat A and should be compared to ASF.RL4

udelWPmnz.RL4¶

Run ID: udelWPmnz.RL4
Participant: udel
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: ec33ad4cf73550d6e34a85cc4093db43
Run description: waterloo spam scores to filter. RL1: basic indri system on WP RL2: use same system to get results from prev queries, filter dups from RL1 RL3: use provided results from prev queries to filter dups from RL1 RL4: query expansion using all docs, or maybe using combMNZ to combine the other two RL4s---note that this one is cat A and should be compared to ASF.RL4

umasscontext.RL1¶

Run ID: umasscontext.RL1
Participant: CIIR
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL1
MD5: e6af2fdbec0a9553925f2eb21cdb6ded
Run description: Sequential dependency model over the query.

umasscontext.RL2¶

Run ID: umasscontext.RL2
Participant: CIIR
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL2
MD5: aedb749673a729e692a80f8299cd3934
Run description: Expands current query with most likely terms from previously submitted queries.

umasscontext.RL3¶

Run ID: umasscontext.RL3
Participant: CIIR
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL3
MD5: aedb749673a729e692a80f8299cd3934
Run description: Expands current query with most likely terms from previously submitted queries (same as RL2).

umasscontext.RL4¶

Run ID: umasscontext.RL4
Participant: CIIR
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL4
MD5: 6054f4b7410fff5fc55f55dd08840234
Run description: Expands current query with most likely terms from previously submitted queries and summaries of clicked documents.

umassqdist.RL1¶

Run ID: umassqdist.RL1
Participant: CIIR
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL1
MD5: a1027da56aed938afde1adb6b8b6e0ed
Run description: Score distribution over queries seeded by current query.

umassqdist.RL2¶

Run ID: umassqdist.RL2
Participant: CIIR
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL2
MD5: b1a9366cfa59c7e05ea87c1f2d80ed29
Run description: Score distribution over queries seeded by current query, expanded with most likely terms from previously submitted queries.

umassqdist.RL3¶

Run ID: umassqdist.RL3
Participant: CIIR
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL3
MD5: b1a9366cfa59c7e05ea87c1f2d80ed29
Run description: Score distribution over queries seeded by current query, expanded with most likely terms from previously submitted queries (same as RL2).

umassqdist.RL4¶

Run ID: umassqdist.RL4
Participant: CIIR
Track: Session
Year: 2011
Submission: 7/29/2011
Type: automatic
Task: RL4
MD5: ebf5ecea88e934cb5d00ae44a35d05ba
Run description: Score distribution over queries seeded by current query, expanded with most likely terms from previously submitted queries and clicked document summaries.

UvAlearning.RL1¶

Run ID: UvAlearning.RL1
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: d25bcf0214d5ddd23621340b6fda29f6
Run description: combination of results from UvAmodeling.RL1, UvAsemantic.RL1, and a search of the anchortext using the current query. Combinations are learned using a set of 20 self-made training sessions

UvAlearning.RL2¶

Run ID: UvAlearning.RL2
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: d25bcf0214d5ddd23621340b6fda29f6
Run description: combination of results from UvAmodeling.RL2, UvAsemantic.Rl2, and a search of the anchortext using the current query. Combinations are learned using a set of 20 self-made training sessions

UvAlearning.RL3¶

Run ID: UvAlearning.RL3
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: faca3abcdc1c587a4e025ddabf6bc9d0
Run description: combination of results from UvAmodeling.RL3, UvAsemantic.RL3, and a search of the anchortext using the current query. Combinations are learned using a set of 20 self-made training sessions

UvAlearning.RL4¶

Run ID: UvAlearning.RL4
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 209e5da8ea7f2b71c8a4cedf4b640032
Run description: combination of results from UvAmodeling.RL3, UvAsemantic.RL3, and a search of the anchortext using the current query. Combinations are learned using the clicks from other sessions in the run

UvAmodeling.RL1¶

Run ID: UvAmodeling.RL1
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: 5a3644225469bbb83fd1788dbe280658
Run description: baseline - original query only

UvAmodeling.RL2¶

Run ID: UvAmodeling.RL2
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: 59484dc86810633951760d08d53c65b0
Run description: results from original query interpolated with results from previous queries

UvAmodeling.RL3¶

Run ID: UvAmodeling.RL3
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: b3eafdbef7c65e384b7d07ff064bd99a
Run description: results from original query interpolated with results from previous queries, and result from queries formed using previously displayed snippets

UvAmodeling.RL4¶

Run ID: UvAmodeling.RL4
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: c80be9ec5eb1d6416cb6a0dd6de2e4ec
Run description: results from original query interpolated with results from previous queries, and result from queries formed using previously clicked snippets

UvAsemantic.RL1¶

Run ID: UvAsemantic.RL1
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: 7d020206b11be1171635be0ef68fb7d4
Run description: original query, expanded using wikipedia sections

UvAsemantic.RL2¶

Run ID: UvAsemantic.RL2
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: 4e5522daccb6e3395fe284506699255d
Run description: results from original query, interpolated with results from previous queries, all expanded using wikipedia sections

UvAsemantic.RL3¶

Run ID: UvAsemantic.RL3
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: cac04e88691621a3fe337b1c0a649327
Run description: results from original query, interpolated with results from previous queries, all expanded using wikipedia sections. This is further interpolated with results from displayed snippets

UvAsemantic.RL4¶

Run ID: UvAsemantic.RL4
Participant: UvA
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 048eb5a6e97f282c8f0f0e1d9fdede87
Run description: results from original query, interpolated with results from previous queries, all expanded using wikipedia sections. This is further interpolated with results from clicked snippets

webis11cnb.RL1¶

Run ID: webis11cnb.RL1
Participant: Webis
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: bb95238a5e5db3ecd07a3d918d6c4591
Run description: Used ChatNoir Search Engine

webis11cnb.RL2¶

Run ID: webis11cnb.RL2
Participant: Webis
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: 7d8a0d3ef3b0a1b1a8e89bd21af11e50
Run description: Used ChatNoir Search Engine

webis11cnb.RL3¶

Run ID: webis11cnb.RL3
Participant: Webis
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 8f502f83f3f61d40a7ccb46249859a87
Run description: Used ChatNoir Search Engine

webis11cnb.RL4¶

Run ID: webis11cnb.RL4
Participant: Webis
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 33f1e0b5c77c278f93c635e5e5c4fb6e
Run description: Used ChatNoir Search Engine

webis11cnw.RL1¶

Run ID: webis11cnw.RL1
Participant: Webis
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL1
MD5: aae16c7511a450ee4b5c955e894b3784
Run description: Used ChatNoir Search Engine

webis11cnw.RL2¶

Run ID: webis11cnw.RL2
Participant: Webis
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: c0dfb322ffc926e45339253dd420f185
Run description: Used ChatNoir Search Engine

webis11cnw.RL3¶

Run ID: webis11cnw.RL3
Participant: Webis
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: 01eae4d299243e8656ddbea4516b31e0
Run description: Used ChatNoir Search Engine

webis11cnw.RL4¶

Run ID: webis11cnw.RL4
Participant: Webis
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: 4fc2605626e9a88152e0ea8c7cbaa300
Run description: Used ChatNoir Search Engine

webis11ind.RL1¶

Run ID: webis11ind.RL1
Participant: Webis
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 6d0db8ebbd378e155405d8d75e24e3e1
Run description: Used Indri search engine

webis11ind.RL2¶

Run ID: webis11ind.RL2
Participant: Webis
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL2
MD5: c6372798eacc7c81e09255bf4080131d
Run description: Used Indri search engine

webis11ind.RL3¶

Run ID: webis11ind.RL3
Participant: Webis
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL3
MD5: e7e50c0e6a17a8fbc58d13df76d4ff61
Run description: Used Indri search engine

webis11ind.RL4¶

Run ID: webis11ind.RL4
Participant: Webis
Track: Session
Year: 2011
Submission: 8/4/2011
Type: automatic
Task: RL4
MD5: a483ae50b7f8eb22646269c27d491150
Run description: Used Indri search engine

wildcat1.RL1¶

Run ID: wildcat1.RL1
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 6bc4d94affeae134dc9494d7efe13c18
Run description: use spam filter, search the current query in indri, then calc similarity between current query and docs

wildcat1.RL2¶

Run ID: wildcat1.RL2
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: be9af976b38ab38c02d16ab1c0034344
Run description: use past query to reform current query, search the reform query in indri, then calc similarity between reform query and docs

wildcat1.RL3¶

Run ID: wildcat1.RL3
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 21f24a99b92696f15086c88deab1c664
Run description: use clicked docs' anchor log to reform current query, search it in indri, then calc the similarity between current query and docs

wildcat1.RL4¶

Run ID: wildcat1.RL4
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: faa4db37772e5cf81e0ec636d371a9cb
Run description: use past query and clicked titles to reform current query, search it in indri,then calc the similarity between current query and docs

wildcat2.RL1¶

Run ID: wildcat2.RL1
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 199452bbdd5030f34742a7954dc17b30
Run description: use spam filter

wildcat2.RL2¶

Run ID: wildcat2.RL2
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: 7f6e0ce304feeb4f067d4f3935ec883c
Run description: use past query to reform the current query, search it in the indri system, then calc the similarity between the reform query and docs

wildcat2.RL3¶

Run ID: wildcat2.RL3
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: 08c2c2f38ad6a610134be3ed5ddf3bcb
Run description: use anchor log

wildcat2.RL4¶

Run ID: wildcat2.RL4
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: f6941917a2c02e7c176b83c6b8248fee
Run description: use past clicked url and its time to calc simulation with current search docs

wildcat3.RL1¶

Run ID: wildcat3.RL1
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL1
MD5: 61f3c57d5d19e151ce57b0d87153c5fc
Run description: use spam filter and pagerank

wildcat3.RL2¶

Run ID: wildcat3.RL2
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL2
MD5: 9ad4fe810972903631f932e20e4ba130
Run description: use spam filter

wildcat3.RL3¶

Run ID: wildcat3.RL3
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL3
MD5: a15ab4b550c6383df2b7cc90f2ff0aea
Run description: use anchor log

wildcat3.RL4¶

Run ID: wildcat3.RL4
Participant: BUPT_WILDCAT
Track: Session
Year: 2011
Submission: 8/3/2011
Type: automatic
Task: RL4
MD5: de502e438d261d1f8aa569cfaf9a2a77
Run description: use past clicked url and time to calc the similarity between clicked docs and search docs, then predict the time for search docs