Filters








17,340 Hits in 7.9 sec

Predicting Query Performance by Query-Drift Estimation [chapter]

Anna Shtok, Oren Kurland, David Carmel
2009 Lecture Notes in Computer Science  
We argue that query-drift can potentially be estimated by measuring the diversity (e.g., standard deviation) of the retrieval scores of these documents.  ...  Predicting query performance, that is, the effectiveness of a search performed in response to a query, is a highly important and challenging problem.  ...  This paper is based upon work supported in part by Google's and IBM's faculty awards.  ... 
doi:10.1007/978-3-642-04417-5_30 fatcat:7fkrswosrbhizfifpgwzjkmyl4

Predicting Query Performance by Query-Drift Estimation

Anna Shtok, Oren Kurland, David Carmel, Fiana Raiber, Gad Markovits
2012 ACM Transactions on Information Systems  
We argue that query-drift can potentially be estimated by measuring the diversity (e.g., standard deviation) of the retrieval scores of these documents.  ...  Predicting query performance, that is, the effectiveness of a search performed in response to a query, is a highly important and challenging problem.  ...  This paper is based upon work supported in part by Google's and IBM's faculty awards.  ... 
doi:10.1145/2180868.2180873 fatcat:sickq2o4ufhyzhhafoiv5ccqni

Predicting Query Performance via Classification [chapter]

Kevyn Collins-Thompson, Paul N. Bennett
2010 Lecture Notes in Computer Science  
In an empirical study we compare the performance of class-based statistics to their language-model counterparts for two performance-related tasks: predicting query difficulty and expansion risk.  ...  Our findings suggest that using class predictions can offer comparable performance to full language models while reducing computation overhead.  ...  Our examination of model drift extends recent studies that find variance to be an important facet of predicting query performance.  ... 
doi:10.1007/978-3-642-12275-0_15 fatcat:fmupa3v3bng6hf47xftbmuvraa

A Unified Framework for Post-Retrieval Query-Performance Prediction [chapter]

Oren Kurland, Anna Shtok, David Carmel, Shay Hummel
2011 Lecture Notes in Computer Science  
The query-performance prediction task is estimating the effectiveness of a search performed in response to a query in lack of relevance judgments.  ...  The framework is based on using a pseudo effective and/or ineffective ranking as reference comparisons to the ranking at hand, the quality of which we want to predict.  ...  This paper is based upon work supported in part by the Israel Science Foundation under grant no. 557/09, and by IBM's SUR award.  ... 
doi:10.1007/978-3-642-23318-0_4 fatcat:a5sglzqpordxpifpmwsjqvxizy

Query-performance prediction and cluster ranking

Oren Kurland, Fiana Raiber, Anna Shtok
2012 Proceedings of the 21st ACM international conference on Information and knowledge management - CIKM '12  
The first is query-performance prediction; i.e., estimating the effectiveness of a search performed in response to a query in the absence of relevance judgments.  ...  Finally, we empirically demonstrate that using insights gained in work on query-performance prediction can help, in many cases, to improve the performance of a previously proposed cluster ranking method  ...  The query performance prediction (QPP) task [3] is estimating the effectiveness of D res using some clustering algorithm.  ... 
doi:10.1145/2396761.2398666 dblp:conf/cikm/KurlandRS12 fatcat:mcoc43qfjnaq5p2sqzrbajnmfe

Aggregate Query Prediction under Dynamic Workloads

Fotis Savva, Christos Anagnostopoulos, Peter Triantafillou
2019 2019 IEEE International Conference on Big Data (Big Data)  
The estimations are performed in milliseconds and are inexepensive as the mechanism learns from past analytical-query patterns.  ...  Therefore, we introduce an adaptive Machine Learning mechanism which is light-weight, stored client-side, can estimate the answers of a variety of aggregate queries and can avoid the big data backend.  ...  Our work contributes with monitoring and detecting real-time query patterns change based on approximating the prediction error, which differentiates with the previous concept drift methods by measuring  ... 
doi:10.1109/bigdata47090.2019.9006267 dblp:conf/bigdataconf/SavvaAT19 fatcat:aor2vcoddzeodhdz4ogomumwyq

Using Query Performance Predictors to Reduce Spoken Queries [chapter]

Jaime Arguello, Sandeep Avula, Fernando Diaz
2017 Lecture Notes in Computer Science  
The goal of query performance prediction is to estimate a query's retrieval effectiveness without user feedback.  ...  Our results show that we are able to outperform the original spoken query by a small, but significant margin.  ...  This work was supported in part by NSF grant IIS-1451668.  ... 
doi:10.1007/978-3-319-56608-5_3 fatcat:tiwpjztv3zbubms4dhbzf3gqxy

Estimating query performance using class predictions

Kevyn Collins-Thompson, Paul N. Bennett
2009 Proceedings of the 32nd international ACM SIGIR conference on Research and development in information retrieval - SIGIR '09  
In an empirical study we compare the performance of class-based statistics to their languagemodel counterparts for predicting two measures: query difficulty and expansion risk.  ...  Our findings suggest that using class predictions can offer comparable performance to full language models while reducing computation overhead.  ...  DISCUSSION & CONCLUSIONS We explored the use of new sources of evidence in estimating two important measures of query performance, query difficulty and expansion risk, by comparing two document representations  ... 
doi:10.1145/1571941.1572071 dblp:conf/sigir/Collins-ThompsonB09 fatcat:65a27ndzafgklmitrcdisqzrzq

Stream-based Active Learning with Verification Latency in Non-stationary Environments [article]

Andrea Castellani, Sebastian Schmitt, Barbara Hammer
2022 arXiv   pre-print
We propose PRopagate (PR), a latency independent utility estimator which also predicts the requested, but not yet known, labels.  ...  Stream-based Active Learning (AL) approaches address this problem by interactively querying a human expert to provide new data labels for the most recent samples, within a limited budget.  ...  For readability, we only report the best performing query strategy (pal ) of Fig. 6 , and we compare the proposed PR with FS.B utility estimator.  ... 
arXiv:2204.06822v1 fatcat:uirw2p7e5zcxvm42vxpsn47b7i

Adaptive learning of aggregate analytics under dynamic workloads

Fotis Savva, Christos Anagnostopoulos, Peter Triantafillou
2020 Future generations computer systems  
The estimations are performed in milliseconds are inexpensive and accurate as the mechanism learns from past analytical-query patterns.  ...  Therefore, we introduce an adaptive, reciprocity-based Machine Learning mechanism which is light-weight, stored client-side, can estimate the answers of a variety of aggregate queries and can avoid the  ...  Then the lower-bound and upper-bound for the given predicates is set and we execute the query over the restricted space given by the bounds over the selected columns by b using dataset d.  ... 
doi:10.1016/j.future.2020.03.063 fatcat:yiz6yhi4s5d63iv3746lya6cya

Selectivity correction with online machine learning [article]

Max Halford and Philippe Saint-Pierre and Franck Morvan
2020 arXiv   pre-print
As an experiment, we teach models to improve the selectivity estimates made by PostgreSQL's cost model.  ...  In the database community, many recent proposals have been made to improve selectivity estimation with batch machine learning methods.  ...  We then multiplied each predicted correction factor with the selectivity estimate made by PostgreSQL's cost model.  ... 
arXiv:2009.09884v1 fatcat:fo6ieco7azaclpg6thgh2cglx4

A Bi-Criteria Active Learning Algorithm for Dynamic Data Streams

Saad Mohamad, Abdelhamid Bouchachia, Moamar Sayed-Mouchaweh
2018 IEEE Transactions on Neural Networks and Learning Systems  
A learner deliberately queries specific instances to tune the classifier's model using as few labels as possible.  ...  Index Terms-Active learning (AL), Bayesian online learning, concept drift, data streams. 2162-237X  ...  To sum up, BAL shows the ability to adapt to different types of drift but its performance may be affected by the data distribution.  ... 
doi:10.1109/tnnls.2016.2614393 pmid:27775910 fatcat:4bj6vywl3fhz7h7lpzg7gw2l6a

Adaptive Learning of Aggregate Analytics under Dynamic Workloads [article]

Fotis Savva, Christos Anagnostopoulos, Peter Triantafillou
2019 arXiv   pre-print
The estimations are performed in milliseconds are inexpensive and accurate as the mechanism learns from past analytical-query patterns.  ...  Therefore, we introduce an adaptive Machine Learning mechanism which is light-weight, stored client-side, can estimate the answers of a variety of aggregate queries and can avoid the big data backend.  ...  Our work contributes with monitoring and detecting real-time query patterns change based on approximating the prediction error, which differentiates with the previous concept drift methods by measuring  ... 
arXiv:1908.04772v1 fatcat:o2jgnqnv7baezodnaxldyyz37a

Warper: Efficiently Adapting Learned Cardinality Estimators to Data and Workload Drifts

Beibin Li, Yao Lu, Srikanth Kandula
2022 Proceedings of the 2022 International Conference on Management of Data  
We also show that faster model adaptation improves query performance by shortening the period for which imperfect query plans are picked by a query optimizer due to incorrect cardinality estimates.  ...  Recent learned cardinality estimation (CE) models are vulnerable when query predicates or the underlying datasets drift from what the models were trained upon.  ...  By improving the model accuracy, Warper improves the query performance in a sizable way; as the figure on the right shows, adapting to workload drift reduces cardinality estimation errors by up to 3× and  ... 
doi:10.1145/3514221.3526179 fatcat:dhht7qkthjertjhpeflzuqz3cu

A fuzzy kernel c-means clustering model for handling concept drift in regression

Yiliao Song, Guangquan Zhang, Jie Lu, Haiyan Lu
2017 2017 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE)  
Experiments show that the FKLL algorithm is better able to respond to drift as soon as the learning sets are updated, and is also suitable for dealing with reoccurring drift, when compared to the original  ...  There is, however, a shortage of research on drift adaptation for regression cases in the literature.  ...  ACKNOWLEDGMENT The work presented in this paper was supported by the Australian Research Council (ARC) under discovery grant DP150101645.  ... 
doi:10.1109/fuzz-ieee.2017.8015515 dblp:conf/fuzzIEEE/SongZLL17 fatcat:nfl4thwxdbhqlg6xrojbqjfppm
« Previous Showing results 1 — 15 out of 17,340 results