Filters








1,801 Hits in 3.1 sec

Accelerated Convergence for Counterfactual Learning to Rank [article]

Rolf Jagerman, Maarten de Rijke
2020 pre-print
Counterfactual Learning to Rank (LTR) algorithms learn a ranking model from logged user interactions, often collected using a production system.  ...  One of the major difficulties in applying Stochastic Gradient Descent (SGD) approaches to counterfactual learning problems is the large variance introduced by the propensity weights.  ...  ACKNOWLEDGMENTS We thank Chang Li, Harrie Oosterhuis and Ilya Markov for helpful discussions and feedback. We thank the anonymous reviewers for their feedback.  ... 
doi:10.1145/3397271.3401069 arXiv:2005.10615v1 fatcat:rjgwhs27tjcxxjqcyb2ti3cbei

Counterfactual Online Learning to Rank [chapter]

Shengyao Zhuang, Guido Zuccon
2020 Lecture Notes in Computer Science  
Two main methods have arisen for optimizing rankers based on implicit feedback: counterfactual learning to rank (CLTR), which learns a ranker from the historical click-through data collected from a deployed  ...  In this paper, we propose a counterfactual online learning to rank algorithm (COLTR) that combines the key components of both CLTR and OLTR.  ...  Counterfactual Online Learning to Rank Counterfactual Evaluation for Online Learning to Rank The proposed COLTR method uses counterfactual evaluation to estimate the effectiveness of candidate rankers  ... 
doi:10.1007/978-3-030-45439-5_28 fatcat:gpqp6bfqgza6tmd67sfdbqtyky

Algorithms in Multi-Agent Systems: A Holistic Perspective from Reinforcement Learning and Game Theory [article]

Yunlong Lu, Kai Yan
2020 arXiv   pre-print
Counterfactual regret minimization is an important tool to solve games with incomplete information, and has shown great strength when combined with deep learning.  ...  Solution concepts from game theory give inspiration to algorithms which try to evaluate the agents or find better solutions in multi-agent systems.  ...  UN-REAL framework [36] is based on A3C and proposes unsupervised auxiliary tasks like reward prediction to accelerate the learning process.  ... 
arXiv:2001.06487v3 fatcat:o2iovnsbxfgp5omk67jwuoonma

Double Neural Counterfactual Regret Minimization [article]

Hui Li, Kailiang Hu, Zhibang Ge, Tao Jiang, Yuan Qi, Le Song
2018 arXiv   pre-print
To make neural learning efficient, we also developed several novel techniques including a robust sampling method, mini-batch Monte Carlo Counterfactual Regret Minimization (MCCFR) and Monte Carlo Counterfactual  ...  Experimentally, we demonstrate that the proposed double neural algorithm converges significantly better than the reinforcement learning counterpart.  ...  To avoid the algorithm converging to potential local minima or saddle point, we will reset the learning rate to 0.001 and help the optimizer to learn a better performance. θ T best is the best parameters  ... 
arXiv:1812.10607v1 fatcat:jtnlxl7iobh3pkrgxub2nocsva

Solving Games with Functional Regret Estimation

Kevin Waugh, Dustin Morrill, James Bagnell, Michael Bowling
2015 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
The approach learns a function approximator online to estimate the regret for choosing a particular action.  ...  We propose a novel online learning method for minimizing regret in large extensive-form games.  ...  Thanks to Compute Canada for computational resources and the Computer Poker Research Group (CPRG) for software infrastructure support.  ... 
doi:10.1609/aaai.v29i1.9445 fatcat:ajtp6urom5fbxktdi6vvhckb2q

Solving Games with Functional Regret Estimation [article]

Kevin Waugh and Dustin Morrill and J. Andrew Bagnell and Michael Bowling
2014 arXiv   pre-print
The approach learns a function approximator online to estimate the regret for choosing a particular action.  ...  We propose a novel online learning method for minimizing regret in large extensive-form games.  ...  Thanks to Compute Canada for computational resources and the Computer Poker Research Group (CPRG) for software infrastructure support.  ... 
arXiv:1411.7974v2 fatcat:57tt42ntbbcvdlqt2jsuqdjway

Trumping Preemption

Jonathan Schaffer
2000 Journal of Philosophy  
exclusively sensitive to ranking orders (as in trumping preemption).  ...  not have occurred.) 20 Now, C is minimally counterfactually sufficient for E if and only if: (i) C is counterfactually sufficient for E, (ii) no proper subset of C is counterfactually sufficient for E  ... 
doi:10.2307/2678388 fatcat:pnc3zsreljasviyuob7wlbrmfi

Multi-Objective Counterfactual Explanations [chapter]

Susanne Dandl, Christoph Molnar, Martin Binder, Bernd Bischl
2020 Lecture Notes in Computer Science  
Counterfactual explanations are one of the most popular methods to make predictions of black box machine learning models interpretable by providing explanations in the form of 'what-if scenarios'.  ...  This enables a more detailed post-hoc analysis to facilitate better understanding and also more options for actionable user responses to change the predicted outcome.  ...  Introduction Interpretable machine learning methods have become very important in recent years to explain the behavior of black box machine learning (ML) models.  ... 
doi:10.1007/978-3-030-58112-1_31 fatcat:qiww7zypcjaffmqonr2iww3bri

SCouT: Synthetic Counterfactuals via Spatiotemporal Transformers for Actionable Healthcare [article]

Bhishma Dedhia, Roshini Balasubramanian, Niraj K. Jha
2022 arXiv   pre-print
to support randomized controlled trials, and a medical intervention for patients with Friedreich's ataxia to improve clinical decision-making and promote personalized therapy.  ...  We instead propose an approach to use local spatiotemporal information before the onset of the intervention as a promising way to estimate the counterfactual sequence.  ...  algorithm to learn a deep representation c i for the units using pre-intervention covariates.  ... 
arXiv:2207.04208v1 fatcat:b2ya3l3vzbcoho7mlgsjkh3uzu

Generating Plausible Counterfactual Explanations for Deep Transformers in Financial Text Classification [article]

Linyi Yang, Eoin M. Kenny, Tin Lok James Ng, Yi Yang, Barry Smyth, Ruihai Dong
2020 arXiv   pre-print
To address these issues, this paper proposes a novel methodology for producing plausible counterfactual explanations, whilst exploring the regularization benefits of adversarial training on language models  ...  Exhaustive quantitative experiments demonstrate that not only does this approach improve the model accuracy when compared to the current state-of-the-art and human performance, but it also generates counterfactual  ...  Also, we would like to thank the anonymous reviewers for their insightful comments and suggestions to help improve the paper.  ... 
arXiv:2010.12512v1 fatcat:ynkbj6j6srhj5fy6yymtqs7tfa

Teasing out the overall survival benefit with adjustment for treatment switching to other therapies [article]

Yuqing Xu, Meijing Wu, Weili He, Qiming Liao, Yabing Mai
2019 arXiv   pre-print
Several commonly used statistical methods are available to estimate overall survival benefit while adjusting for treatment switching, ranging from naive exclusion or censoring approaches to more advanced  ...  methods including inverse probability of censoring weighting (IPCW), iterative parameter estimation (IPE) algorithm or rank-preserving structural failure time models (RPSFTM).  ...  This iterative process continues until the new estimate for exp ( ) converges.  ... 
arXiv:1908.00654v1 fatcat:6kwvpbcmd5dztmi3phs27hmblq

Leveraging Sparse Linear Layers for Debuggable Deep Networks [article]

Eric Wong, Shibani Santurkar, Aleksander Mądry
2021 arXiv   pre-print
We show how fitting sparse linear models over learned deep feature representations can lead to more debuggable neural networks.  ...  The code for our toolkit can be found at https://github.com/madrylab/debuggabledeepnetworks.  ...  Acknowledgements We thank Dimitris Tsipras for helpful discussions. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein.  ... 
arXiv:2105.04857v1 fatcat:fqwhkbr7krhnzal5ys74f5euci

Accelerating Economic Growth: the Science beneath the Art

Michele Peruzzi, Alessio Terzi
2021 Economic Modelling  
However, a set of counterfactual analyses suggest that such major improvements in growth determinants are a necessary but hardly sufficient condition for success, failing to accelerate growth in 9 out  ...  However, a set of counterfactual analyses suggest that such major improvements in growth determinants are a necessary but hardly sufficient condition for success, failing to accelerate growth in 9 out  ...  This does not spell well for convergence, as it implies that richer countries make the most of growth accelerations.  ... 
doi:10.1016/j.econmod.2021.105593 fatcat:h2eibcbvhzbf5nwerev6zezuxa

Benchmark Evaluation of Counterfactual Algorithms for XAI: From a White Box to a Black Box [article]

Yu-Liang Chou and Chihcheng Hsieh and Catarina Moreira and Chun Ouyang and Joaquim Jorge and João Madeiras Pereira
2022 arXiv   pre-print
metrics; (2) the counterfactual generated are not impacted by the different types of machine learning models; (3) DiCE was the only tested algorithm that was able to generate actionable and plausible counterfactuals  ...  We evaluated the different counterfactual algorithms using several metrics including proximity, interpretability and functionality for five datasets.  ...  This work was also partially supported by Queensland University of Technology (QUT) Centre for Data Science First Byte Funding Program  ... 
arXiv:2203.02399v1 fatcat:fbbri4u4xvccxjmcg6tbkm2jxa

A Data-Based Perspective on Transfer Learning [article]

Saachi Jain, Hadi Salman, Alaa Khaddaj, Eric Wong, Sung Min Park, Aleksander Madry
2022 arXiv   pre-print
In this work, we take a closer look at the role of the source dataset's composition in transfer learning and present a framework for probing its impact on downstream performance.  ...  Our framework gives rise to new capabilities such as pinpointing transfer learning brittleness as well as detecting pathologies such as data-leakage and the presence of misleading examples in the source  ...  To do so, we use the transfer influences computed for CIFAR-10 in order to perform the counterfactual experiments for other datasets.  ... 
arXiv:2207.05739v1 fatcat:dn5rtwaa65efpjkgpe5lw7y7ya
« Previous Showing results 1 — 15 out of 1,801 results