8,533 Hits in 10.1 sec

Hybrid CNN and RNN-based shilling attack framework in social recommender networks

Praveena Narayanan, Vivekanandan. K
2018 EAI Endorsed Transactions on Scalable Information Systems  
METHODS: Hybrid CNN and RNNs-based shilling attack framework is proposed for shilling attack detection based on the selection of dynamic features for attaining maximized detection accuracy.  ...  INTRODUCTION: Recommender system is considered to be widely utilized in diversified domain for the purpose of effectively handling information overload.  ...  In this research, machine learning and deep learning-based shilling attack detection schemes are proposed for efficient detection by eliminating the limitations of the existing approaches.  ... 
doi:10.4108/eai.2-11-2021.171754 fatcat:h5qishixmfbidp4er3vq3rii3q

Addressing Competing Objectives in Allocating Funds to Scholarships and Need-based Financial Aid

Vinthuy Phan, Laura Wright, Bridgette Decent, Antonija Mitrovic, Nigel Bosch
2022 Zenodo  
We introduce an approach that couples a gradient boosting classifier for predicting outcomes from an allocation strategy with a local search optimization algorithm, which optimizes strategies based on  ...  Unlike most existing approaches that focus strictly on allocating merit-based awards, ours optimizes simultaneously the allocation of both merit-based awards and need-based aid.  ...  Technical approach. An overview of our approach is depicted in Figure 1 .  ... 
doi:10.5281/zenodo.6853028 fatcat:nbg254fcnzeclmpqigowp3u6gy

A kernelized maximal-figure-of-merit learning approach based on subspace distance minimization

Byungki Byun, Chin-Hui Lee
2011 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
We propose a kernelized maximal-figure-of-merit (MFoM) learning approach to efficiently training a nonlinear model using subspace distance minimization.  ...  This construction of the subset enables us to learn a nonlinear model efficiently while keeping the resulting model nearly optimal compared to the model from the whole training data set.  ...  An interesting example of such techniques is a maximal-figureof-merit (MFoM) learning approach, which has been successfully applied to many pattern classification problems, including text categorization  ... 
doi:10.1109/icassp.2011.5946732 dblp:conf/icassp/ByunL11 fatcat:4tt2iyud6fcyvkbwiwt37w33re

GaN Power Amplifier Digital Predistortion by Multi-Objective Optimization for Maximum RF Output Power

Mattia Mengozzi, Gian Piero Gibiino, Alberto M. Angelotti, Corrado Florian, Alberto Santarelli
2021 Electronics  
We present a predistorer learning procedure based on a constrained optimization algorithm that maximizes the RF output power, while guaranteeing a prescribed linearity level, i.e., a maximum normalized  ...  a multi-objective optimization approach.  ...  Acknowledgments: The authors would like to thank Chalmers University for providing the remote RF weblab [29] . Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/electronics10030244 fatcat:4nfxr5oc4jg65cjitvwol4vone

Model-Based Reinforcement Learning via Meta-Policy Optimization [article]

Ignasi Clavera, Jonas Rothfuss, John Schulman, Yasuhiro Fujita, Tamim Asfour, Pieter Abbeel
2018 arXiv   pre-print
Model-based reinforcement learning approaches carry the promise of being data efficient.  ...  Using an ensemble of learned dynamic models, MB-MPO meta-learns a policy that can quickly adapt to any model in the ensemble with one policy gradient step.  ...  Kurutach for the feedback on the earlier draft of the paper. IC was supported by La Caixa Fellowship.  ... 
arXiv:1809.05214v1 fatcat:txnrke4ejbexrkviuzhruuxtxu

SoDeep: a Sorting Deep net to learn ranking loss surrogates [article]

Martin Engilberge, Louis Chevallier, Patrick Pérez, Matthieu Cord
2019 arXiv   pre-print
Several tasks in machine learning are evaluated using non-differentiable metrics such as mean average precision or Spearman correlation.  ...  Our approach yields very competitive results on these three tasks, which validates the merit and the flexibility of SoDeep as a proxy for sorting operation in ranking-based losses.  ...  Several works studied the problem of optimizing average precision with support vector machines [21, 40] and other works extended these approaches to neural networks [1, 31, 8] .  ... 
arXiv:1904.04272v1 fatcat:sr6cgy7ouvgxnid67nku3y5voq

SoDeep: A Sorting Deep Net to Learn Ranking Loss Surrogates

Martin Engilberge, Louis Chevallier, Patrick Perez, Matthieu Cord
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Several tasks in machine learning are evaluated using non-differentiable metrics such as mean average precision or Spearman correlation.  ...  Our approach yields very competitive results on these three tasks, which validates the merit and the flexibility of SoDeep as a proxy for sorting operation in ranking-based losses.  ...  Several works studied the problem of optimizing average precision with support vector machines [21, 40] and other works extended these approaches to neural networks [1, 31, 8] .  ... 
doi:10.1109/cvpr.2019.01105 dblp:conf/cvpr/EngilbergeCPC19 fatcat:5z732nofr5etheedodh7at5o6u

Inverse design and flexible parameterization of meta-optics using algorithmic differentiation [article]

Shane Colburn, Arka Majumdar
2021 arXiv   pre-print
As an open-source platform adaptable to other algorithms and problems, we enable fast and flexible meta-optical design.  ...  For the adjoint method, this requires explicitly deriving gradients, which is sometimes challenging for certain photonics problems.  ...  figure of merit) to optimize.  ... 
arXiv:2011.03626v2 fatcat:y45b7nf5izbmdicfjeg53uhqji

What Does a Policy Network Learn After Mastering a Pong Game? [chapter]

Somnuk Phon-Amnuaisuk
2017 Lecture Notes in Computer Science  
In this report, we discuss an alternative end-to-end approach where the RL attempts to learn general task representations, in this context, learning how to play the Pong game from a sequence of screen  ...  We apply artificial neural networks to approximate a policy of a reinforcement learning model.  ...  Acknowledgments We wish to thank anonymous reviewers for their comments that have helped improve this paper. We would like to thank the GSR office for their financial support given to this research.  ... 
doi:10.1007/978-3-319-69456-6_18 fatcat:a4ag6dmhbneklm6n57mrwuwede

Computationally Efficient Optimization of Plackett-Luce Ranking Models for Relevance and Fairness [article]

Harrie Oosterhuis
2021 arXiv   pre-print
Unlike existing approaches that are based on policy gradients, PL-Rank makes use of the specific structure of PL models and ranking metrics.  ...  Theoretically, they can be used to optimize ranking metrics via stochastic gradient descent.  ...  These methods optimize PL ranking models to maximize the probability of the optimal ranking.  ... 
arXiv:2105.00855v2 fatcat:pyqceojuvfcslgtburjigal56m

Machine-Learning-Assisted Metasurface Design for High-Efficiency Thermal Emitter Optimization [article]

Zhaxylyk A. Kudyshev and Alexander V. Kildishev and Vladimir M. Shalaev and Alexandra Boltasseva
2020 arXiv   pre-print
In our approach, we merge the topology optimization method with machine learning algorithms such as adversarial autoencoders and show substantial improvement of the optimization process by providing unparalleled  ...  With the emergence of new photonic and plasmonic materials with optimized properties as well as advanced nanofabrication techniques, nanophotonic devices are now capable of providing solutions to global  ...  The spatial distribution of the forward and adjoint fields determines the "heat map" of the dielectric function perturbation inside the optimization domain, which maximizes the figure of merit gradient  ... 
arXiv:1910.12741v2 fatcat:ypnwnpj35jcclagyluynwammpe

Deep Inverse Optimization [article]

Yingcong Tan, Andrew Delong, Daria Terekhov
2018 arXiv   pre-print
Given a set of observations generated by an optimization process, the goal of inverse optimization is to determine likely parameters of that process.  ...  Our method, called deep inverse optimization, is to unroll an iterative optimization process and then use backpropagation to learn parameters that generate the observations.  ...  In Figure 1 (ii), starting from c ini , A ini and b ini , our approach finds c lrn , A lrn and b lrn which make x tru an optimal solution of the learned LP through minimizing x tru − x lrn 2 .  ... 
arXiv:1812.00804v1 fatcat:5svkudvgyfeefcuzcmvhvtsqpu

Quantum enhancements for deep reinforcement learning in large spaces [article]

Sofiene Jerbi, Lea M. Trenkwalder, Hendrik Poulsen Nautrup, Hans J. Briegel, Vedran Dunjko
2020 arXiv   pre-print
for efficiency of computation.  ...  In this work, we study the state-of-the-art neural-network approaches for reinforcement learning with quantum enhancements in mind.  ...  The only optimal policy on the other hand, runs through a loop of length N − 1 and achieves an average reward of 1.  ... 
arXiv:1910.12760v2 fatcat:oz526qk5ljfxjbqsve2g4dkice

Refining activation downsampling with SoftPool [article]

Alexandros Stergiou, Ronald Poppe, Grigorios Kalliatakis
2021 arXiv   pre-print
An important feature of the pooling operation is the minimization of information loss, with respect to the initial activation maps, without a significant impact on the computation and memory overhead.  ...  To meet these requirements, we propose SoftPool: a fast and efficient method for exponentially weighted activation downsampling.  ...  We use an initial learning rate of 0.1 with an SGD optimizer and a step-wise learning rate reduction every 40 epochs for a total of 100 epochs.  ... 
arXiv:2101.00440v3 fatcat:dbw4nrcjdjcvhkamgl7kzvlt6i

Energy Optimization of Wind Turbines via a Neural Control Policy Based on Reinforcement Learning Markov Chain Monte Carlo Algorithm [article]

Vahid Tavakol Aghaei, Arda Ağababaoğlu, Peiman Naseradinmousavi, Sinan Yıldırım, Serhat Yeşilyurt, Ahmet Onat
2022 arXiv   pre-print
Our MCMC-based RL algorithm is a model-free and gradient-free algorithm, where the designer does not have to know the precise dynamics of the plant and their uncertainties.  ...  Through this work, we formulate and implement an RL strategy using Markov chain Monte Carlo (MCMC) algorithm to optimize the long-term energy output of the wind turbine.  ...  In e.g. an MPPT controller, a step wind would produce an immediate rise in the current which is a greedy approach that maximizes instantaneous power but delays the acceleration of the rotor to the optimal  ... 
arXiv:2209.03485v1 fatcat:w6ss6upqbfc4lngv2z2k4yqvui
« Previous Showing results 1 — 15 out of 8,533 results