Filters








676 Hits in 4.6 sec

Improving Cooperative Game Theory-based Data Valuation via Data Utility Learning [article]

Tianhao Wang, Yu Yang, Ruoxi Jia
2022 arXiv   pre-print
In this work, we propose to boost the efficiency in computing Shapley value or Least core by learning to estimate the performance of a learning algorithm on unseen data combinations.  ...  Both methods have recently been proposed as a principled solution for data valuation tasks, i.e., quantifying the contribution of individual datum in machine learning.  ...  To showcase the usefulness of data utility function learning, we study its application to data valuation as a concrete example.  ... 
arXiv:2107.06336v2 fatcat:cnf3wbvlobh7tbu57r7fxfiota

Learnability of Learning Performance and Its Application to Data Valuation [article]

Tianhao Wang, Yu Yang, Ruoxi Jia
2021
On the other hand, the ability to efficiently estimate learning performance may benefit a wide spectrum of applications, such as active learning, data quality management, and data valuation.  ...  We then discuss a natural, important use case of learning performance learning -- data valuation, which is known to suffer computational challenges due to the requirement of estimating learning performance  ...  BACKGROUND AND RELATED WORK The goal of data valuation is to quantify the contribution of each training data point to a learning task.  ... 
doi:10.48550/arxiv.2107.06336 fatcat:fk42rpm6hfekpj4wgysmznognu

Learning Valuation Functions [article]

Maria Florina Balcan, Florin Constantin, Satoru Iwata, Lei Wang
2011 arXiv   pre-print
We provide upper and lower bounds regarding the learnability of important subclasses of valuation functions that express no-complementarities.  ...  We provide nearly tight lower and upper bounds of Θ̃(n^1/2) on the approximation factor for learning XOS and subadditive valuations, both widely studied superclasses of submodular valuations.  ...  B Additional Results for Theorem 4 We prove that Algorithm 2 can be used to PAC-learn (i.e. PMAC-learn with α = 1) any unit-demand valuation.  ... 
arXiv:1108.5669v2 fatcat:rgrcww5egvgw5itkxy653lywrm

Game Theory Meets Computational Learning Theory (Dagstuhl Seminar 17251)

Paul W. Goldberg, Yishay Mansour, Paul Dütting, Marc Herbstritt
2017 Dagstuhl Reports  
While there have been many Dagstuhl seminars on various aspects of Algorithmic Game Theory, this was the first one to focus on the emerging field of its intersection with computational learning theory.  ...  This report documents the program and the outcomes of Dagstuhl Seminar 17251 "Game Theory Meets Computational Learning Theory".  ...  Learning, Optimization, and Noise In combinatorial auctions, for example, an agent may not know her valuation but rather learns it by observing data.  ... 
doi:10.4230/dagrep.7.6.68 dblp:journals/dagstuhl-reports/GoldbergMD17 fatcat:ca4mfrf3qbdbhhbo7rvc53myti

The learnable evolution model in agent-based delivery optimization

Janusz Wojtusiak, Tobias Warden, Otthein Herzog
2012 Memetic Computing  
The Learnable Evolution Model is a stochastic optimization method which employs machine learning to guide the optimization process.  ...  The presented research concerns its application within a multi-agent system for autonomous control of container on-carriage operations.  ...  Therefore, it is worthwhile to identify such knowledge in the given application domain and provide it to the LEM system.  ... 
doi:10.1007/s12293-012-0088-9 fatcat:kxv2ebdfrrd5rjm4tmxphj7tky

Usability Considerations of Mobile Learning Applications

Ali Mostakhdemin-Hosseini
2009 International Journal of Interactive Mobile Technologies  
, Learnability, Lack of Errors, and Memorability The assessment and importance of the usability attributes varies based on the application and the type of the users.  ...  For the expert users' lack of errors, reliability and efficiency are important and for the novice users beside mentioned factors the adjustability, learnability and memorability are essentials factors.  ...  It is essential to clarify the context of use of mobile learning platform, which enables to create a successful content.  ... 
doi:10.3991/ijim.v3s1.854 fatcat:dpmi6myp75dmxktoxhtz4evp7q

Data Valuation using Reinforcement Learning [article]

Jinsung Yoon, Sercan O. Arik, Tomas Pfister
2019 arXiv   pre-print
We demonstrate that DVRL yields superior data value estimates compared to alternative methods across different types of datasets and in a diverse set of application scenarios.  ...  To adaptively learn data values jointly with the target task predictor model, we propose a meta learning framework which we name Data Valuation using Reinforcement Learning (DVRL).  ...  Our method, DVRL, differs from the aforementioned as it directly models the value of the data using learnable neural networks (which we refer to as a data value estimator).  ... 
arXiv:1909.11671v1 fatcat:tltgkffp5vf77ig7njn4bnjojq

Prioritizing e-learning websites evaluation and selection criteria using fuzzy set theory

Rakesh Garg, Dimpal Jain
2017 Management Science Letters  
To show the relative importance of each selection criteria, they ranked according to their global weights.  ...  So, for the effective evaluation and selection of E-learning websites, a set of selection criteria should be obtained.  ...  TEACHERS ACCESSING SHARED DATA COMPLETE CONTENT UP-TO-DATE CONTENT USEFUL CONTENT RIGHT & UNDERSTANDABLE CONTENT LEARNING PROGRESS LEARNING PERFORMANCE APPLICABLY IN PHASE REQUIRED TIME REQUIRED  ... 
doi:10.5267/j.msl.2017.1.002 fatcat:tcpgekq5nrcmxixna3srcwvj3q

Finding High-Value Training Data Subset through Differentiable Convex Programming [article]

Soumi Das, Arshdeep Singh, Saptarshi Chatterjee, Suparna Bhattacharya, Sourangshu Bhattacharya
2021 arXiv   pre-print
The key idea is to design a learnable framework for online subset selection, which can be learned using mini-batches of training data, thus making our method scalable.  ...  Extensive evaluation on a synthetic dataset, and three standard datasets, show that our algorithm finds consistently higher value subsets of training data, compared to the recent state-of-the-art methods  ...  Introduction Estimation of "value" of a training datapoint from a Machine Learning model point of view, broadly called data valuation [10, 19, 16] , has become an important problem with many applications  ... 
arXiv:2104.13794v1 fatcat:j5vyms4rtzgwhjoxw53thget6m

AI Reasoning Systems: PAC and Applied Methods [article]

Jeffrey Cheng
2018 arXiv   pre-print
Machine learning has experienced a surge in popularity because it is robust to noise and achieves high performance; however, ML experiences many issues with knowledge transfer and extrapolation.  ...  Learning and logic are distinct and remarkable approaches to prediction.  ...  not suffer at all (curiously, it seems to perform better on the randomized setting).  ... 
arXiv:1807.05054v1 fatcat:o4k5dzjwfzdqxg3g6gsb2dbyau

Task-Based Learning via Task-Oriented Prediction Network with Applications in Finance [article]

Di Chen, Yada Zhu, Xiaodong Cui, Carla P. Gomes
2020 arXiv   pre-print
A major benefit of the proposed TOPNet learning scheme lies in its capability of automatically integrating non-differentiable evaluation criteria, which makes it particularly suitable for diversified and  ...  Real-world applications often involve domain-specific and task-based performance objectives that are not captured by the standard machine learning losses, but are critical for decision making.  ...  to 06-30-2019, 421,225 data points) to validate the performance of models.  ... 
arXiv:1910.09357v4 fatcat:eltssi4gfzfh5juzj3nc74ssua

A proposed index of usability: A method for comparing the relative usability of different software systems

Han X. Lin, Yee-Yin Choong, Gavriel Salvendy
1997 Behavior and Information Technology  
An experiment was conducted to test the validity of PUTQ. The experiment result showed high correlation between PUTQ and the Questionnaire for User Interaction Satisfaction (QUIS version 5.5).  ...  In addition, PUTQ detected the diOE erences in user performance between two experimental interface systems, but QUIS failed to do so.  ...  Acknowledgements We gratefully acknowledge and appreciate the generous support of the NEC corporation in supporting the NEC professorship which made this study possible.  ... 
doi:10.1080/014492997119833 fatcat:rnds6zpabrfoxdtptnmcejq22e

Data Valuation for Offline Reinforcement Learning [article]

Amir Abolfazli, Gregory Palmer, Daniel Kudenko
2022 arXiv   pre-print
To address this, we propose data valuation for offline reinforcement learning (DVORL), which allows us to identify relevant and high-quality transitions, improving the performance and transferability of  ...  The field of offline reinforcement learning addresses these issues through outsourcing the collection of data to a domain expert or a carefully monitored program and subsequently searching for a batch-constrained  ...  ACKNOWLEDGMENT The authors gratefully acknowledge, that the proposed research is a result of the research project "IIP-Ecosphere", granted by the German Federal Ministry for Economics and Climate Action  ... 
arXiv:2205.09550v1 fatcat:lap2jiz2grhb3kb2rvd4pcsggq

Managing Editor's Letter

Francesco A. Fabozzi
2021 The Journal of Financial Data Science  
of report cards, and the relationship between a dataset's structure of information content and its potential to enhance investment returns.  ...  In the opening article, "Alternative Data in Investment Management: Usage, Challenges, and Valuation," Gene Ekster and Petter N.  ...  of clusters and by preprocessing the data using distance metric learning.  ... 
doi:10.3905/jfds.2021.3.4.001 fatcat:k7a4y4jkybelxpwbmehbmpvwdq

Agent-based Pickup and Delivery Planning: The Learnable Evolution Model Approach

Janusz Wojtusiak, Tobias Warden, Otthein Herzog
2011 2011 International Conference on Complex, Intelligent, and Software Intensive Systems  
Implementation and experimental evaluation of the method is performed within the PlaSMA multiagent simulation platform.  ...  In order to compile transport plans and render optimized decisions agents managing transport vehicles employ a guided evolutionary computation method, called the learnable evolution model (LEM).  ...  Current research on the learnable evolution model is supported by the National Institute of Standards and Technology grant.  ... 
doi:10.1109/cisis.2011.11 dblp:conf/cisis/WojtusiakWH11 fatcat:yn5ccbvixfg2tppsolther5kte
« Previous Showing results 1 — 15 out of 676 results