检索结果-内蒙古大学图书馆

49th IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Xu, Xianbo van Erp, Bart Ignatenko, Tanya Eindhoven Univ Technol Dept Elect Engn Eindhoven Netherlands GN Hearing Res & Technol Eindhoven Netherlands

ISBN: (纸本)9798350344868;9798350344851

We study a context-aware preference learning system that automatically learns user preferences in different environments. The system is based on a Dirichlet process Gaussian mixture model and comprises an environmental model and a preference learning system. The environmental model is used to determine the user's current environment, based on environmental signals. The preference learning system actively learns user preferences through pairwise comparison trials. During these trials, the so-called reference and alternative proposals are generated by the system and presented to a user, who, in his/her turn, indicates the preferred one. We propose a novel inference approach that allows for environmental model personalization and leads to improved system performance. This is achieved by introducing a two-way dependency between environment and preference domains. Our approach is validated through simulations that demonstrate a significant performance improvement with respect to the baseline model.

关键词： Active learning clustering Dirichlet process Bayesian nonparametric modelling preference learning

来源：评论

学校读者我要写书评

暂无评论

Few-Shot preference learning for Human-in-the-Loop RL 6

Few-Shot Preference Learning for Human-in-the-Loop RL

引用

6th Conference on Robot learning (CoRL)

作者： Hejna, Joey Sadigh, Dorsa Stanford Univ Stanford CA 94305 USA

While reinforcement learning (RL) has become a more popular approach for robotics, designing sufficiently informative reward functions for complex tasks has proven to be extremely difficult due their inability to capture human intent and policy exploitation. preference based RL algorithms seek to overcome these challenges by directly learning reward functions from human feedback. Unfortunately, prior work either requires an unreasonable number of queries implausible for any human to answer or overly restricts the class of reward functions to guarantee the elicitation of the most informative queries, resulting in models that are insufficiently expressive for realistic robotics tasks. Contrary to most works that focus on query selection to minimize the amount of data required for learning reward functions, we take an opposite approach: expanding the pool of available data by viewing human-in-the-loop RL through the more flexible lens of multi-task learning. Motivated by the success of meta-learning, we pre-train preference models on prior task data and quickly adapt them for new tasks using only a handful of queries. Empirically, we reduce the amount of online feedback needed to train manipulation policies in Meta-World by 20x, and demonstrate the effectiveness of our method on a real Franka Panda Robot. Moreover, this reduction in query-complexity allows us to train robot policies from actual human users. Videos of our results can be found at https://***/view/few-shot-preference-rl/home.

关键词： preference learning Interactive learning Multi-task learning

来源：评论

学校读者我要写书评

暂无评论

Bayesian preference learning for Interactive Multi-objective Optimisation

Bayesian Preference Learning for Interactive Multi-objective...

引用

2nd Genetic and Evolutionary Computation Conference (GECCO)

作者： Taylor, Kendall Ha, Huong Li, Minyi Chan, Jeffrey Li, Xiaodong RMIT Univ Sch Comp Technol Melbourne Vic Australia

ISBN: (纸本)9781450383509

This work proposes a Bayesian optimisation with Gaussian Process approach to learn decision maker (DM) preferences in the attribute search space of a multi-objective optimisation problem (MOP). The DM is consulted periodically during optimisation of the problem and asked to provide their preference over a series of pairwise comparisons of candidate solutions. After each consultation, the most preferred solution is used as the reference point in an appropriate multiobjective optimisation evolutionary algorithm (MOEA). The rationale for using Bayesian optimisation is to identify the most preferred location in the decision search space with the least number of DM queries, thereby minimising DM cognitive burden and fatigue. This enables non-expert DMs to be involved in the optimisation process and make more informed decisions. We further reduce the number of preference queries required, by progressively redefining the Bayesian search space to reflect the MOEA's decision bounds as it converges toward the Pareto Front. We demonstrate how this approach can locate a reference point close to an unknown preferred location on the Pareto Front, of both benchmark and real-world problems with relatively few pairwise comparisons.

关键词： Multi-objective optimisation preferences evolutionary algorithms interactive algorithms Bayesian optimisation preference learning pairwise comparisons active learning

来源：评论

学校读者我要写书评

暂无评论

Dealing with Intransitivity, Non-Convexity, and Algorithmic Bias in preference learning

Dealing with Intransitivity, Non-Convexity, and Algorithmic ...

引用

作者： Bower, Amanda University of Michigan

学位级别：Ph.D., Doctor of Philosophy

Rankings are ubiquitous since they are a natural way to present information to people who are making decisions. There are seemingly countless scenarios where rankings arise, such as deciding whom to hire at a company, determining what movies to watch, purchasing products, understanding human perception, judging science fair projects, voting for political candidates, and so on. In many of these scenarios, the number of items in consideration is prohibitively large, such that asking someone to rank all of the choices is essentially impossible. On the other hand, collecting preference data on a small subset of the items is feasible, e.g., collecting answers to ``Do you prefer item A or item B?" or ``Is item A closer to item B or item C?". Therefore, an important machine learning task is to learn a ranking of the items based on this preference data. This thesis theoretically and empirically addresses three key challenges of preference learning: intransitivity in preference data, non-convex optimization, and algorithmic bias. Chapter 2 addresses the challenge of learning a ranking given pairwise comparison data that violates rational choice such as intransitivity. Our key observation is that two items compared in isolation from other items may be compared based on only a salient subset of features. Formalizing this framework, we propose the salient feature preference model and prove a sample complexity result for learning the parameters of our model and the underlying ranking with maximum likelihood estimation. Chapter 3 addresses the non-convexity of an optimization problem inspired by ordinal embedding, which is a preference learning task. We aim to understand the landscape, that is local minimizers and global minimizers, of the non-convex objective, which corresponds to the hinge loss arising from quadratic constraints. Under certain assumptions, we give necessary conditions for non-global, local minimizers of our objective and additionally show that in two dimensions

关键词： Algorithmic fairness Ranking preference learning Algorithmic bias Non-convex optimization Intransitivity

来源：评论

学校读者我要写书评

暂无评论

Multi-criteria Recommendations through preference learning 17

Multi-criteria Recommendations through Preference Learning

引用

4th ACM IKDD Conferences on Data Sciences (CODS)

作者： Sreepada, Rama Syamala Patra, Bidyut Kr. Hernando, Antonio Natl Inst Technol Rourkela Rourkela 769008 Odisha India Univ Politecn Madrid E-28040 Madrid Spain

ISBN: (纸本)9781450348461

In today's internet era, recommender system (RS) addresses information overload problem, which is common in many information driven domains. RS helps users chose a set of appropriate options from a plethora of options. Traditional single rating recommender systems have been playing a vital role over the decades in various domains. However, it is limited in a sense of providing user's accurate preferences about an item or services to the recommendation engine. The single rating recommender systems receive a single rating about an item, due to which these systems are inadequate to understand the reasons behind users' choice of items. On the other hand, multi-criteria rating systems allow the users to share more information about user's interest/disinterest through multiple criteria of an item. Therefore, the multi-criteria recommender engine gets more information from the users and provides relevant recommendations to the users. In this paper, we propose a novel technique to learn and rank users' preferences over different criteria. Dominant criteria of each item are also learnt and ranked in the proposed technique. The obtained ranks are exploited to predict the overall rating by adapting the traditional user-based and item-based collaborative filtering techniques. We conducted experiments on two real world datasets (TripAdvisor and Yahoo! Movies) and our approach outperforms the traditional single rating systems and existing approaches on multi-criteria recommender systems.

关键词： Single rating recommender systems Multi-criteria ratings preference learning Collaborative filtering

来源：评论

学校读者我要写书评

暂无评论

A preference learning System for the Autonomous Selection and Personalization of Entertainment Activities during Human-Robot Interaction

A Preference Learning System for the Autonomous Selection an...

引用

IEEE International Conference on Development and learning (ICDL)

作者： Maroto-Gomez, Marcos Marques Villarroya, Sara Malfaz, Maria Castro-Gonzalez, Alvaro Carlos Castillo, Jose Angel Salichs, Miguel Univ Carlos III Syst Engn & Automat Madrid Spain

ISBN: (数字)9781665413114

ISBN: (纸本)9781665413114

Social robots assisting in cognitive stimulation therapies, physical rehabilitation, or entertainment sessions have gained visibility in the last years. In these activities, users may present different features and needs, so personalization is essential. This manuscript presents a preference learning System for social robots to personalize Human-Robot Interaction during entertainment activities. Our system is integrated into Mini, a social robot dedicated to research with a wide repertoire of entertainment activities like games, displaying multimedia content, or storytelling. The learning model we propose consists of four stages. First, the robot creates a unique profile of its users by obtaining their defining features using interaction. Secondly, a preference learning algorithm predicts the users' favorite entertainment activities using their features and a database with the features and preferences of other users. Third, the prediction is adapted using Reinforcement learning while entertainment sessions occur. Finally, the robot personalizes Human-Robot Interaction by autonomously selecting the users' favorite activities. Thus, the robot aims at promoting longer-lasting interactions and sustaining engagement.

关键词： Autonomous Decision-making preference learning Reinforcement learning User Profiling Social robots Robot personalization

来源：评论

学校读者我要写书评

暂无评论

Implementation of medical image retrieval algorithm based on preference learning model 4

Implementation of medical image retrieval algorithm based on...

引用

4th IEEE International Conference on Power, Electronics and Computer Applications, ICPECA 2024

作者： Chen, Yaqi Luo, Chan He, Zijie Shu, Zhong School of Electronic Information Engineering Jingchu University of Technology Hubei Jingmen China Jingmen Rongmedia Network Technology Co. Ltd Hubei Jingmen China

ISBN: (纸本)9798350359589

This paper utilizes the Stacked Denoising Auto Encoder (SDAE) model to capture the general edge features of images. Additionally, it employs the Convolutional Neural Networks model to extract fine edge features from medical images. By combining preference learning algorithms, the paper implements an image retrieval system and designs a medical image retrieval model termed SDAE+CNN+preference learning. In the design of the reconstruction residual function in the SDAE model, noise factors are introduced to ensure the accuracy of image feature extraction by accounting for the impact of noise. Adjustments have been made to the structure of the CNN model, specifically in the convolutional layers, pooling layers, and fully connected layers. These adjustments aim to further enhance the accuracy of image feature extraction and ensure the precision of image feature classification. By conducting learning training on key parameters, an optimal preference learning model is obtained. Following the rules for matching the category attributes of input images with those in the medical image database and optimizing the loss function of the preference learning model, the similarity of category attributes for images is ranked. Ultimately, images with higher category attribute similarity are listed as retrieval results. Relevant experiments indicate that the proposed SDAE+CNN image feature extraction module in this paper achieves high accuracy in extracting image features, effectively controlling the influence of image noise;The designed SDAE+CNN+preference learning medical image retrieval model achieves a retrieval similarity of 96%, a retrieval accuracy of 94%, and a retrieval recall rate of 95%, as per the conducted experiments. © 2024 IEEE.

关键词： Convolutional Neural Networks Image Retrievalt preference learning Stacked Denoising Auto Encoder

来源：评论

学校读者我要写书评

暂无评论

Dealing with Intransitivity, Non-Convexity, and Algorithmic Bias in preference learning

Dealing with Intransitivity, Non-Convexity, and Algorithmic ...

引用

作者： Bower, Amanda Ruth University of Michigan

关键词： algorithmic fairness ranking preference learning algorithmic bias non-convex optimization intransitivity Thesis

来源：评论

学校读者我要写书评

暂无评论

Online Personalized preference learning Method Based on In-Formative Query for Lane Centering Control Trajectory

引用

SENSORS 2023年第11期23卷 5246页

作者： Ran, Wei Chen, Hui Xia, Taokai Nishimura, Yosuke Guo, Chaopeng Yin, Youyu Tongji Univ Sch Automot Studies Shanghai 201804 Peoples R China JTEKT Corp Nara 6348555 Japan JTEKT Res & Dev Ctr WUXI Co Ltd Wuxi 214161 Peoples R China

The personalization of autonomous vehicles or advanced driver assistance systems has been a widely researched topic, with many proposals aiming to achieve human-like or driver-imitating methods. However, these approaches rely on an implicit assumption that all drivers prefer the vehicle to drive like themselves, which may not hold true for all drivers. To address this issue, this study proposes an online personalized preference learning method (OPPLM) that utilizes a pairwise comparison group preference query and the Bayesian approach. The proposed OPPLM adopts a two-layer hierarchical structure model based on utility theory to represent driver preferences on the trajectory. To improve the accuracy of learning, the uncertainty of driver query answers is modeled. In addition, informative query and greedy query selection methods are used to improve learning speed. To determine when the driver's preferred trajectory has been found, a convergence criterion is proposed. To evaluate the effectiveness of the OPPLM, a user study is conducted to learn the driver's preferred trajectory in the curve of the lane centering control (LCC) system. The results show that the OPPLM can converge quickly, requiring only about 11 queries on average. Moreover, it accurately learned the driver's favorite trajectory, and the estimated utility of the driver preference model is highly consistent with the subject evaluation score.

关键词： online learning preference learning utility theory Bayesian approach LCC trajectory

来源：评论

学校读者我要写书评

暂无评论

An active preference learning approach to aid the selection of validators in blockchain environments

引用

OMEGA-INTERNATIONAL JOURNAL OF MANAGEMENT SCIENCE 2023年 118卷

作者： Gehrlein, Jonas Miebs, Grzegorz Brunelli, Matteo Kadzinski, Milosz Web3 Fdn Zug Switzerland Poznati Univ Technol Inst Comp Sci Piotrowo 2 PL-60965 Poznan Poland Univ Trento Dept Ind Engn Trento Italy

We consider a real-world problem faced in some blockchain ecosystems that select their active validators-the actors that maintain the blockchain-from a larger set of candidates through an electionbased mechanism. Specifically, we focus on Polkadot, a protocol that aggregates preference lists from another set of actors, nominators, that contain a limited number of trusted validators and thereby influence the election's outcome. This process is financially incentivized but often overwhelms human decision makers due to the problem's complexity and the multitude of available alternatives. This paper presents a decision support system (DSS) to help the nominators choose the validators in an environment with frequently changing data. The system structures the relevant multiple attribute problem and incorporates a dedicated active learning algorithm. Its goal is to find a sufficiently small set of pairwise elicitation questions to infer nominators' preferences. We test the proposed solution in an experiment with 115 real nominators from the Polkadot ecosystem. The empirical results confirm that our approach outperforms the unaided process in terms of required interaction time, imposed cognitive effort, and offered efficacy. The developed DSS can be easily extended to other blockchain ecosystems.& COPY;2023 Elsevier Ltd. All rights reserved.

关键词： Blockchain ecosystem Validator selection Multiple criteria decision analysis Active learning preference learning Multi-attribute value function

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：