检索结果-内蒙古大学图书馆

arXiv 2022年

作者： Tang, Kun Cao, Xu Cao, Zhipeng Zhou, Tong Li, Erlong Liu, Ao Zou, Shengtao Liu, Chang Mei, Shuqi Sizikova, Elena Zheng, Chao T Lab Tencent Beijing China Center of Data Science New York University New York United States

Nowadays, autonomous vehicle technology is becoming more and more mature. Critical to progress and safety, high-definition (HD) maps, a type of centimeter-level map collected using a laser sensor, provide accurate descriptions of the surrounding environment. The key challenge of HD map production is efficient, high-quality collection and annotation of large-volume datasets. Due to the demand for high quality, HD map production requires significant manual human effort to create annotations, a very time-consuming and costly process for the map industry. In order to reduce manual annotation burdens, many artificial intelligence (AI) algorithms have been developed to pre-label the HD maps. However, there still exists a large gap between AI algorithms and the traditional manual HD map production pipelines in accuracy and robustness. Furthermore, it is also very resource-costly to build large-scale annotated datasets and advanced machine learning algorithms for AI-based HD map automatic labeling systems. In this paper, we introduce the Tencent HD Map AI (THMA) system, an innovative end-to-end, AI-based, active learning HD map labeling system capable of producing and labeling HD maps with a scale of hundreds of thousands of kilometers. In THMA, we train AI models directly from massive HD map datasets via supervised, self-supervised, and weakly supervised learning to achieve high accuracy and efficiency required by downstream users. THMA has been deployed by the Tencent Map team to provide services to downstream companies and users, serving over 1,000 labeling workers and producing more than 30,000 kilometers of HD map data per day at most. More than 90 percent of the HD map data in Tencent Map is labeled automatically by THMA, accelerating the traditional HD map labeling process by more than ten times. Copyright © 2022, The Authors. All rights reserved.

关键词： learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Sketching for First Order Method: Efficient Algorithm for Low-Bandwidth Channel and Vulnerability

arXiv

引用

arXiv 2022年

作者： Song, Zhao Wang, Yitan Yu, Zheng Zhang, Lichen Adobe Research United States Yale University United States Alibaba Inc. China MIT United States

Sketching is one of the most fundamental tools in large-scale machine learning. It enables runtime and memory saving via randomly compressing the original large problem into lower dimensions. In this paper, we propose a novel sketching scheme for the first order method in large-scale distributed learning setting, such that the communication costs between distributed agents are saved while the convergence of the algorithms is still guaranteed. Given gradient information in a high dimension d, the agent passes the compressed information processed by a sketching matrix R ∈ s×d with s d, and the receiver de-compressed via the de-sketching matrix RT to "recover" the information in original dimension. Using such a framework, we develop algorithms for federated learning with lower communication costs. However, such random sketching does not protect the privacy of local data directly. We show that the gradient leakage problem still exists after applying the sketching technique by presenting a specific gradient attack method. As a remedy, we prove rigorously that the algorithm will be differentially private by adding additional random noises in gradient information, which results in a both communication-efficient and differentially private first order approach for federated learning tasks. Our sketching scheme can be further generalized to other learning settings and might be of independent interest itself. © 2022, CC BY-NC-SA.

关键词： learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Quantifying French Document Complexity

arXiv

引用

arXiv 2022年

作者： Primpied, Vincent Beauchemin, David Khoury, Richard Department of Computer Science and Software Engineering Université Laval Canada

Measuring a document’s complexity level is an open challenge, particularly when one is working on a diverse corpus of documents rather than comparing several documents on a similar topic or working on a language other than English. In this paper, we define a methodology to measure the complexity of French documents, using a new general and diversified corpus of texts, the "French Canadian complexity level corpus", and a wide range of metrics. We compare different learning algorithms to this task and contrast their performances and their observations on which characteristics of the texts are more significant to their complexity. Our results show that our methodology gives a general-purpose measurement of text complexity in French. © 2022, CC BY.

关键词： learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Q-Ensemble for Offline RL: Don’t Scale the Ensemble, Scale the Batch Size

arXiv

引用

arXiv 2022年

作者： Nikulin, Alexander Kurenkov, Vladislav Tarasov, Denis Akimov, Dmitry Kolesnikov, Sergey Tinkoff Russia

Training large neural networks is known to be time-consuming, with the learning duration taking days or even weeks. To address this problem, large-batch optimization was introduced. This approach demonstrated that scaling mini-batch sizes with appropriate learning rate adjustments can speed up the training process by orders of magnitude. While long training time was not typically a major issue for model-free deep offline RL algorithms, recently introduced Q-ensemble methods achieving state-of-the-art performance made this issue more relevant, notably extending the training duration. In this work, we demonstrate how this class of methods can benefit from large-batch optimization, which is commonly overlooked by the deep offline RL community. We show that scaling the mini-batch size and naively adjusting the learning rate allows for (1) a reduced size of the Q-ensemble, (2) stronger penalization of out-of-distribution actions, and (3) improved convergence time, effectively shortening training duration by 3-4x times on average.1 © 2022, CC BY.

关键词： learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Consistent Multiclass algorithms for Complex Metrics and Constraints

arXiv

引用

arXiv 2022年

作者： Narasimhan, Harikrishna Ramaswamy, Harish G. Tavker, Shiv Kumar Khurana, Drona Netrapalli, Praneeth Agarwal, Shivani Google Research Mountain View United States Indian Institute of Technology Madras Chennai India Amazon Inc. Bengaluru India RBCDSAI Indian Institute of Technology Madras Chennai India Google Research India Bengaluru India University of Pennsylvania Philadelphia United States

We present consistent algorithms for multiclass learning with complex performance metrics and constraints, where the objective and constraints are defined by arbitrary functions of the confusion matrix. This setting includes many common performance metrics such as the multiclass G-mean and micro F1-measure, and constraints such as those on the classifier’s precision and recall and more recent measures of fairness discrepancy. We give a general framework for designing consistent algorithms for such complex design goals by viewing the learning problem as an optimization problem over the set of feasible confusion matrices. We provide multiple instantiations of our framework under different assumptions on the performance metrics and constraints, and in each case show rates of convergence to the optimal (feasible) classifier (and thus asymptotic consistency). Experiments on a variety of multiclass classification tasks and fairness constrained problems show that our algorithms compare favorably to the state-of-the-art baselines. © 2022, CC BY.

关键词： learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Minimax rate of consistency for linear models with missing values

arXiv

引用

arXiv 2022年

作者： Ayme, Alexis Boyer, Claire Dieuleveut, Aymeric Scornet, Erwan ParisF-75005 France MOKAPLAN INRIA Paris France CMAP UMR7641 Ecole Polytechnique IP Paris Palaiseau91128 France

Missing values arise in most real-world data sets due to the aggregation of multiple sources and intrinsically missing information (sensor failure, unanswered questions in surveys...). In fact, the very nature of missing values usually prevents us from running standard learning algorithms. In this paper, we focus on the extensively-studied linear models, but in presence of missing values, which turns out to be quite a challenging task. Indeed, the Bayes rule can be decomposed as a sum of predictors corresponding to each missing pattern. This eventually requires to solve a number of learning tasks, exponential in the number of input features, which makes predictions impossible for current real-world datasets. First, we propose a rigorous setting to analyze a least-square type estimator and establish a bound on the excess risk which increases exponentially in the dimension. Consequently, we leverage the missing data distribution to propose a new algorithm, and derive associated adaptive risk bounds that turn out to be minimax optimal. Numerical experiments highlight the benefits of our method compared to state-of-the-art algorithms used for predictions with missing values. Copyright © 2022, The Authors. All rights reserved.

关键词： learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Decentralized Competing Bandits in Non-Stationary Matching Markets

arXiv

引用

arXiv 2022年

作者： Ghosh, Avishek Sankararaman, Abishek Ramchandran, Kannan Javidi, Tara Mazumdar, Arya UC San Diego United States AWS AI Palo Alto United States Electrical Engg. and Computer Sciences UC Berkeley United States Electrical and Computer Engineering UC San Diego United States

Understanding complex dynamics of two-sided online matching markets, where the demand-side agents compete to match with the supply-side (arms), has recently received substantial interest. To that end, in this paper, we introduce the framework of decentralized two-sided matching market under non stationary (dynamic) environments. We adhere to the serial dictatorship setting, where the demand-side agents have unknown and different preferences over the supply-side (arms), but the arms have fixed and known preference over the agents. We propose and analyze a decentralized and asynchronous learning algorithm, namely Decentralized Non-stationary Competing Bandits (DNCB), where the agents play (restrictive) successive elimination type learning algorithms to learn their preference over the arms. The complexity in understanding such a system stems from the fact that the competing bandits choose their actions in an asynchronous fashion, and the lower ranked agents only get to learn from a set of arms, not dominated by the higher ranked agents, which leads to forced exploration. With carefully defined complexity parameters, we characterize this forced exploration and obtain sub-linear (logarithmic) regret of DNCB. Furthermore, we validate our theoretical findings via experiments. Copyright © 2022, The Authors. All rights reserved.

关键词： learning algorithms

来源：评论

学校读者我要写书评

暂无评论

The Sample Complexity of Online Contract Design

arXiv

引用

arXiv 2022年

作者： Zhu, Banghua Bates, Stephen Yang, Zhuoran Wang, Yixin Jiao, Jiantao Jordan, Michael I. Department of Electrical Engineering and Computer Sciences UC Berkeley United States Department of Statistics UC Berkeley United States Department of Statistics and Data Science Yale University United States Department of Statistics University of Michigan United States

We study the hidden-action principal-agent contract design problem in an online setting. In each round, the principal posts a contract that specifies the payment to the agent based on each outcome. The agent then makes a strategic choice of action that maximizes her own utility, but the action is not directly observable by the principal. The principal observes the outcome and receives utility from the agent’s choice of action. Based on past observations, the principal dynamically adjusts the contracts with the goal of maximizing her utility. We treat this problem as a continuum-armed bandit problem, where we think of each potential contract as an arm. We introduce an online learning algorithm and provide an upper bound on its Stackelberg regret. In particular, we show that when the contract space is [0, 1]m, the Stackelberg regret is upper bounded by Oe(√m · T1−1/(2m+1)), and lower bounded by Ω(T1−1/(m+2)), where Oe omits logarithmic factors. This result shows that exponential-in-m samples are both sufficient and necessary to learn a near-optimal contract, resolving an open problem in Ho et al. [2016] on the hardness of online contract design. Moreover, when contracts are restricted to some subset F ⊂ [0, 1]m, we define an intrinsic dimension of F that depends on the covering number of the spherical code in the space and bound the regret in terms of this intrinsic dimension. When F is the family of linear contracts, we show that the Stackelberg regret grows exactly as Θ(T2/3). Technically, the contract design problem is challenging because the utility function is discontinuous. Indeed, bounding the discretization error in this setting has been an open problem [Ho et al., 2016]. In this paper, we identify a limited set of directions in which the utility function is continuous, allowing us to design a new discretization method and bound its error. This approach enables the first upper bound with no restrictions on the contract and action space. The techniques we intro

关键词： learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Training Language Models with Language Feedback

arXiv

引用

arXiv 2022年

作者： Scheurer, Jérémy Campos, Jon Ander Chan, Jun Shern Chen, Angelica Cho, Kyunghyun Perez, Ethan New York University United States University of the Basque Country Spain Genentech United States CIFAR LMB

Pretrained language models often do not perform tasks in ways that are in line with our preferences, e.g., generating offensive text or factually incorrect summaries. Recent work approaches the above issue by learning from a simple form of human evaluation: comparisons between pairs of model-generated task outputs. Comparison feedback conveys limited information about human preferences per human evaluation. Here, we propose to learn from natural language feedback, which conveys more information per human evaluation. We learn from language feedback on model outputs using a three-step learning algorithm. First, we condition the language model on the initial output and feedback to generate many refinements. Second, we choose the refinement with the highest similarity to the feedback. Third, we finetune a language model to maximize the likelihood of the chosen refinement given the input. In synthetic experiments, we first evaluate whether language models accurately incorporate feedback to produce refinements, finding that only large language models (175B parameters) do so. Using only 100 samples of human-written feedback, our learning algorithm finetunes a GPT-3 model to roughly human-level summarization ability. Copyright © 2022, The Authors. All rights reserved.

关键词： learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Patterns of near-crash events in a naturalistic driving dataset: applying rules mining

arXiv

引用

arXiv 2022年

作者： Kong, Xiaoqiang Das, Subasish Zhou, Hongmin Zhang, Yunlong Texas A&M University 3136 TAMU College StationTX77843 United States Texas A&M Transportation Institute 1111 RELLIS Parkway BryanTX77807 United States

This study aims to explore the associations between near-crash events and road geometry and trip features by investigating a naturalistic driving dataset and a corresponding roadway inventory dataset using an association rule mining method - the Apriori algorithm. To provide more insights into near-crash behavior, this study classified near-crash events into two severity levels: trivial near-crash events (-7.5 g ≤ deceleration rate ≤ -4.5 g) and non-trivial near-crash events (≤ -7.5 g). Each category for all variables is considered an item, and a set of items is considered an itemset. From the perspective of descriptive statistics, the frequency of the itemsets generated by the Apriori algorithm suggests that near-crash events are highly associated with several factors, including roadways without access control, driving during non-peak hours, roadways without a shoulder or a median, roadways with the minor arterial functional class, and roadways with a speed limit between 30 and 60 mph. By comparing the frequency of the occurrence of the itemset during trivial and non-trivial near-crash events, the results indicate that the length of the trip is a strong indicator of the near-crash event type. The results show that non-trivial near-crash events are more likely to occur if the trip is longer than 2 hours. After applying the association rule mining algorithm, more interesting patterns for the two near-crash events were generated through the rules. The main findings include: 1) trivial near-crash events are more likely to occur on roadways without a median and shoulder that have a relatively lower functional class;2) relatively higher functional roadways with relatively wide medians and shoulders could be an intriguing combination for non-trivial near-crash events;3) non-trivial near-crash events often occur on long trips (more than 2 hours);4) congestion on roadways that have a lower functional class is a dominant rule of non-trivial near-crash events. This study asso

关键词： learning algorithms

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：