Filters








48,206 Hits in 8.4 sec

Suppressing outliers in pairwise preference ranking

Vitor R. Carvalho, Jonathan L. Elsas, William W. Cohen, Jaime G. Carbonell
2008 Proceeding of the 17th ACM conference on Information and knowledge mining - CIKM '08  
Many of the recently proposed algorithms for learning feature-based ranking functions are based on the pairwise preference framework, in which instead of taking documents in isolation, document pairs are  ...  In this paper we study the effects of outlying pairs in rank learning with pairwise preferences and introduce a new meta-learning algorithm capable of suppressing these undesirable effects.  ...  Given the model learned, we calculated the pairwise decision scores P l [2] for all training data instances and constructed a histogram, as shown in the top of that the learned ranking model correctly  ... 
doi:10.1145/1458082.1458348 dblp:conf/cikm/CarvalhoECC08 fatcat:4f56irgtevgzpocqh5ceg7fycm

Predicting Strategies for Lead Optimization via Learning to Rank

Nobuaki Yasuo, Keisuke Watanabe, Hideto Hara, Kentaro Rikimaru, Masakazu Sekijima
2018 IPSJ Transactions on Bioinformatics  
to rank methods.  ...  We present a concept for a computational compound optimization system that outputs optimized compounds from hit compounds by using previous lead optimization data from a pharmaceutical company.  ...  The authors would like to thank N. Arai for useful discussions. This work was partially supported by the Research Complex  ... 
doi:10.2197/ipsjtbio.11.41 fatcat:v3zysdnnzfhsbox6flczofdjaa

Constrained Clustering and Multiple Kernel Learning without Pairwise Constraint Relaxation [article]

Benedikt Boecking, Vincent Jeanselme, Artur Dubrawski
2022 arXiv   pre-print
We introduce a new constrained clustering algorithm that jointly clusters data and learns a kernel in accordance with the available pairwise constraints.  ...  These pairwise constraints, which come in the form of must-link and cannot-link pairs, arise naturally in many applications and are intuitive for users to provide.  ...  In addition, for large datasets one may choose to downsample data for which no constraints are known to further reduce training complexity while learning an appropriate kernel.  ... 
arXiv:2203.12546v1 fatcat:h43kwo6a6reyjj6kly4w7z5zu4

Active Learning to Rank using Pairwise Supervision [chapter]

Buyue Qian, Hongfei Li, Jun Wang, Xiang Wang, Ian Davidson
2013 Proceedings of the 2013 SIAM International Conference on Data Mining  
As the performance of a learnt ranking model is predominantly determined by the quality and quantity of training data, in this work we explore an active learning to rank approach.  ...  This paper investigates learning a ranking function using pairwise constraints in the context of human-machine interaction.  ...  In each trial, we randomly select 100 pairs as the initial training data to learn the ranking function.  ... 
doi:10.1137/1.9781611972832.33 dblp:conf/sdm/DavidsonLQWW13 fatcat:a57qzorzzrajfb7s25utzomcpy

Learning to rank for why-question answering

Suzan Verberne, Hans van Halteren, Daphne Theijssen, Stephan Raaijmakers, Lou Boves
2010 Information retrieval (Boston)  
We provide an in-depth analysis of the effect of data imbalance and hyperparameter tuning, and we relate our findings to previous research on learning to rank for Information Retrieval.  ...  In this paper, we evaluate a number of machine learning techniques for the task of ranking answers to why-questions.  ...  Learning to rank is the task of learning the optimal ranking for the answers within each cluster.  ... 
doi:10.1007/s10791-010-9136-6 fatcat:pzoi5ntpqbd7hfjnnqfzh5gdve

Rank-smoothed Pairwise Learning In Perceptual Quality Assessment [article]

Hossein Talebi, Ehsan Amid, Peyman Milanfar, Manfred K. Warmuth
2020 arXiv   pre-print
Training a model on these pairwise preferences is a common deep learning approach.  ...  However, optimizing by gradient descent through mini-batch learning means that the "global" ranking of the images is not explicitly taken into account.  ...  RankNet [14] is perhaps the most commonly used approach for learning to rank from pairwise comparisons.  ... 
arXiv:2011.10893v1 fatcat:ivwa54vj5nc5tbjpsymukjsw5y

HybridDTA: Hybrid Data Fusion through Pairwise Training for Drug-Target Affinity Prediction [article]

Hongyu Luo, Yingfei Xiang, Xiaomin Fang, Wei Lin, Fan Wang, Hua Wu, Haifeng Wang
2021 bioRxiv   pre-print
Since the ranking orders of the affinity scores with respect to measurements and experimental batches are more consistent, we adopt a pairwise paradigm to enable the DNNs to learn from ranking orders instead  ...  We propose a novel paradigm for effective training on hybrid DTA data to alleviate the data inconsistency issue.  ...  Secondly, a pairwise training method is applied to optimize the model parameters of the DTA backbond by learning the orders of pairs.  ... 
doi:10.1101/2021.11.23.469641 fatcat:6k6v5ispcrewta54rc2ookeqsy

ListBERT: Learning to Rank E-commerce products with Listwise BERT [article]

Lakshya Kumar, Sagnik Sarkar
2022 arXiv   pre-print
model that can be easily deployed and leads to ~10 times lower ranking latency.  ...  The approxNDCG based RoBERTa model also achieves an NDCG improvement of 20.6% compared to the pairwise RankNet based RoBERTa model.  ...  [20] for all 5 For every query, we have a ranked list of products which is indicated by N. the queries in the training set to optimize for the pairwise ranking in RankNet.  ... 
arXiv:2206.15198v1 fatcat:uwdfmnvo3fgenl2n3vi7u62gwm

Case-Based Label Ranking [chapter]

Klaus Brinker, Eyke Hüllermeier
2006 Lecture Notes in Computer Science  
Label ranking studies the problem of learning a mapping from instances to rankings over a predefined set of labels.  ...  It exhibits the appealing property of transparency and is based on an aggregation model which allows to incorporate a broad class of pairwise loss functions on label ranking.  ...  If the training data would offer the utility scores directly, preference learning in the label ranking scenario would reduce to a standard regression problem.  ... 
doi:10.1007/11871842_53 fatcat:zcms462hgrezva6wsuao672bgu

SOLAR: Scalable Online Learning Algorithms for Ranking

Jialei Wang, Ji Wan, Yongdong Zhang, Steven Hoi
2015 Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)  
Traditional learning to rank methods learn ranking models from training data in a batch and offline learning mode, which suffers from some critical limitations, e.g., poor scalability as the model has  ...  To overcome the limitations, this paper presents SO-LAR -a new framework of Scalable Online Learning Algorithms for Ranking, to tackle the challenge of scalable learning to rank.  ...  The goal of learning to rank is to build a ranking model from training data of a set of queries by optimizing some IR performance measures using machine learning techniques.  ... 
doi:10.3115/v1/p15-1163 dblp:conf/acl/WangWZH15 fatcat:vx62h4zrijcbvb6bucqn2m73fq

A Pre-processing Method for Fairness in Ranking [article]

Ryosuke Sonoda
2022 arXiv   pre-print
closed solution of the minimization problem augmented by weights to training data.  ...  In this paper, we propose a fair ranking framework that evaluates the order of training data in a pairwise manner as well as various fairness measurements in ranking.  ...  ACKNOWLEDGMENTS We would like to thank Hiroya Inakoshi for his in-depth feedback on this paper. We would also like to thank Enago (www.enago.jp) for the English language review.  ... 
arXiv:2110.15503v2 fatcat:3lf2xg34nbcbpgfd2ynczjpfxe

End-to-End Cross-Modality Retrieval with CCA Projections and Pairwise Ranking Loss [article]

Matthias Dorfer and Jan Schlüter and Andreu Vall and Filip Korzeniowski and Gerhard Widmer
2017 arXiv   pre-print
network using freely learned projections optimized by a pairwise ranking loss, especially when little training data is available (the code for all three methods is released at: https://github.com/CPJKU  ...  In contrast to previous approaches, the CCA Layer (CCAL) allows us to combine existing objectives for embedding space learning, such as pairwise ranking losses, with the optimal projections of CCA.  ...  In summary, pairwise ranking losses optimize objectives to learn embeddings that are useful for retrieval, and allow incorporating domain knowledge into the loss function.  ... 
arXiv:1705.06979v1 fatcat:mnm4dmpl2nd2jbetsthvivbpyy

Deep Multi-view Learning to Rank [article]

Guanqun Cao and Alexandros Iosifidis and Moncef Gabbouj and Vijay Raghavan and Raju Gottumukkala
2018 arXiv   pre-print
Though multi-view learning and learning to rank have been studied extensively leading to a wide range of applications, multi-view learning to rank as a synergy of both topics has received little attention  ...  We study the problem of learning to rank from multiple sources.  ...  RELATED WORK Learning to rank Learning to rank aims to optimize the combination of data representation for ranking problems [18] .  ... 
arXiv:1801.10402v1 fatcat:esgar3k3fzgthlgd235sfe3wpy

A Learning-to-Rank-Based Investment Portfolio Optimization Framework for Smart Grid Planning

Wenxin Zhao, Xubin Liu, Yujie Wu, Tao Zhang, Luao Zhang
2022 Frontiers in Energy Research  
Visualization and contribution to the discussion of the topic: TZ and LZ. FUNDING This work was supported by the Hunan Natural Science Foundation of China under Grant 2021JJ10019.  ...  The final ranking formula in the L2R algorithms is obtained by automatic learning, while people only need to provide relevant training data for the L2R algorithm.  ...  A typical L2R model consists of two parts: learning system and ranking system. The L2R algorithms obtain the optimal ranking model from the training data through the learning system.  ... 
doi:10.3389/fenrg.2022.852520 fatcat:zsvyahjyg5eoje2nazben2kt5m

Pairwise Fairness for Ranking and Regression [article]

Harikrishna Narasimhan, Andrew Cotter, Maya Gupta, Serena Wang
2020 arXiv   pre-print
We show that the resulting training problems can be efficiently and effectively solved using existing constrained optimization and robust optimization techniques developed for fair classification.  ...  We present pairwise fairness metrics for ranking models and regression models that form analogues of statistical fairness notions such as equal opportunity, equal accuracy, and statistical parity.  ...  Figure 2 : Plot of learned hyperplanes on simulated ranking data with 2 groups.  ... 
arXiv:1906.05330v3 fatcat:z5lkk57i45egbpyqnkjrdy4jim
« Previous Showing results 1 — 15 out of 48,206 results