Learning Fair Representations Without Demographics PDF Download
Are you looking for read ebook online? Search for your book and save it on your Kindle device, PC, phones or tablets. Download Learning Fair Representations Without Demographics PDF full book. Access full book title Learning Fair Representations Without Demographics by Xiaoxue Wang. Download full books in PDF and EPUB format.
Author: Xiaoxue Wang Publisher: ISBN: Category : Electronic dissertations Languages : en Pages : 0
Book Description
Due to hard accessibility, real-world adoption of fair representation learning algorithms lacks the prior knowledge of the sensitive attributes that we wish to be fair with. To address the challenge in fairness without explicit demographics, our solution is based on the idea of maximally randomizing the representation while being as informative as possible about the target task. We operationalize this goal through the concept of maximizing the entropy of the learned representation. For this purpose, we propose two new avenues for entropy maximization in the absence of demographic information: intra-class and inter-class entropy maximization. For 1) intra-class entropy maximization, it maximizes the entropy of the non-target class predictions (excluding the probability of the ground truth class label for classification problems), thus encouraging the model to discard spurious correlations between the different target classes, and for 2) inter-class entropy maximization, it maximizes the entropy of the representation conditioned on the target label, thus encouraging randomization of the samples within each target class label and minimizing the leakage of potential demographic information in the representation. Quantitative and qualitative results of our Maximum Entropy method (MaxEnt) on COMPAS and UCI Adult datasets show that 1) our method can outperform the State-of-the-art (SOTA) Adversarially Reweighted Learning (ARL) method and will enhance the difficulty of extracting sensitive demographic information in representation without prior demographic knowledge 2) our method reaches a good trade-off between utility and fairness.
Author: Xiaoxue Wang Publisher: ISBN: Category : Electronic dissertations Languages : en Pages : 0
Book Description
Due to hard accessibility, real-world adoption of fair representation learning algorithms lacks the prior knowledge of the sensitive attributes that we wish to be fair with. To address the challenge in fairness without explicit demographics, our solution is based on the idea of maximally randomizing the representation while being as informative as possible about the target task. We operationalize this goal through the concept of maximizing the entropy of the learned representation. For this purpose, we propose two new avenues for entropy maximization in the absence of demographic information: intra-class and inter-class entropy maximization. For 1) intra-class entropy maximization, it maximizes the entropy of the non-target class predictions (excluding the probability of the ground truth class label for classification problems), thus encouraging the model to discard spurious correlations between the different target classes, and for 2) inter-class entropy maximization, it maximizes the entropy of the representation conditioned on the target label, thus encouraging randomization of the samples within each target class label and minimizing the leakage of potential demographic information in the representation. Quantitative and qualitative results of our Maximum Entropy method (MaxEnt) on COMPAS and UCI Adult datasets show that 1) our method can outperform the State-of-the-art (SOTA) Adversarially Reweighted Learning (ARL) method and will enhance the difficulty of extracting sensitive demographic information in representation without prior demographic knowledge 2) our method reaches a good trade-off between utility and fairness.
Author: Ian Foster Publisher: CRC Press ISBN: 1498751431 Category : Mathematics Languages : en Pages : 493
Book Description
Both Traditional Students and Working Professionals Acquire the Skills to Analyze Social Problems. Big Data and Social Science: A Practical Guide to Methods and Tools shows how to apply data science to real-world problems in both research and the practice. The book provides practical guidance on combining methods and tools from computer science, statistics, and social science. This concrete approach is illustrated throughout using an important national problem, the quantitative study of innovation. The text draws on the expertise of prominent leaders in statistics, the social sciences, data science, and computer science to teach students how to use modern social science research principles as well as the best analytical and computational tools. It uses a real-world challenge to introduce how these tools are used to identify and capture appropriate data, apply data science models and tools to that data, and recognize and respond to data errors and limitations. For more information, including sample chapters and news, please visit the author's website.
Author: Michael Kearns Publisher: Oxford University Press ISBN: 0190948221 Category : Computers Languages : en Pages : 288
Book Description
Over the course of a generation, algorithms have gone from mathematical abstractions to powerful mediators of daily life. Algorithms have made our lives more efficient, more entertaining, and, sometimes, better informed. At the same time, complex algorithms are increasingly violating the basic rights of individual citizens. Allegedly anonymized datasets routinely leak our most sensitive personal information; statistical models for everything from mortgages to college admissions reflect racial and gender bias. Meanwhile, users manipulate algorithms to "game" search engines, spam filters, online reviewing services, and navigation apps. Understanding and improving the science behind the algorithms that run our lives is rapidly becoming one of the most pressing issues of this century. Traditional fixes, such as laws, regulations and watchdog groups, have proven woefully inadequate. Reporting from the cutting edge of scientific research, The Ethical Algorithm offers a new approach: a set of principled solutions based on the emerging and exciting science of socially aware algorithm design. Michael Kearns and Aaron Roth explain how we can better embed human principles into machine code - without halting the advance of data-driven scientific exploration. Weaving together innovative research with stories of citizens, scientists, and activists on the front lines, The Ethical Algorithm offers a compelling vision for a future, one in which we can better protect humans from the unintended impacts of algorithms while continuing to inspire wondrous advances in technology.
Author: Aileen Nielsen Publisher: O'Reilly Media ISBN: 1492075701 Category : Computers Languages : en Pages : 346
Book Description
Fairness is an increasingly important topic as machine learning and AI more generally take over the world. While this is an active area of research, many realistic best practices are emerging at all steps along the data pipeline, from data selection and preprocessing to blackbox model audits. This book will guide you through the technical, legal, and ethical aspects of making your code fair and secure while highlighting cutting edge academic research and ongoing legal developments related to fairness and algorithms. There is mounting evidence that the widespread deployment of machine learning and artificial intelligence in business and government is reproducing the same biases we are trying to fight in the real world. For this reason, fairness is an increasingly important consideration for the data scientist. Yet discussions of what fairness means in terms of actual code are few and far between. This code will show you how to code fairly as well as cover basic concerns related to data security and privacy from a fairness perspective.
Author: Andrea Vedaldi Publisher: Springer Nature ISBN: 3030585743 Category : Computers Languages : en Pages : 830
Book Description
The 30-volume set, comprising the LNCS books 12346 until 12375, constitutes the refereed proceedings of the 16th European Conference on Computer Vision, ECCV 2020, which was planned to be held in Glasgow, UK, during August 23-28, 2020. The conference was held virtually due to the COVID-19 pandemic. The 1360 revised papers presented in these proceedings were carefully reviewed and selected from a total of 5025 submissions. The papers deal with topics such as computer vision; machine learning; deep neural networks; reinforcement learning; object recognition; image classification; image processing; object detection; semantic segmentation; human pose estimation; 3d reconstruction; stereo vision; computational photography; neural networks; image coding; image reconstruction; object recognition; motion estimation.
Author: K. Gal Publisher: IOS Press ISBN: 164368437X Category : Computers Languages : en Pages : 3328
Book Description
Artificial intelligence, or AI, now affects the day-to-day life of almost everyone on the planet, and continues to be a perennial hot topic in the news. This book presents the proceedings of ECAI 2023, the 26th European Conference on Artificial Intelligence, and of PAIS 2023, the 12th Conference on Prestigious Applications of Intelligent Systems, held from 30 September to 4 October 2023 and on 3 October 2023 respectively in Kraków, Poland. Since 1974, ECAI has been the premier venue for presenting AI research in Europe, and this annual conference has become the place for researchers and practitioners of AI to discuss the latest trends and challenges in all subfields of AI, and to demonstrate innovative applications and uses of advanced AI technology. ECAI 2023 received 1896 submissions – a record number – of which 1691 were retained for review, ultimately resulting in an acceptance rate of 23%. The 390 papers included here, cover topics including machine learning, natural language processing, multi agent systems, and vision and knowledge representation and reasoning. PAIS 2023 received 17 submissions, of which 10 were accepted after a rigorous review process. Those 10 papers cover topics ranging from fostering better working environments, behavior modeling and citizen science to large language models and neuro-symbolic applications, and are also included here. Presenting a comprehensive overview of current research and developments in AI, the book will be of interest to all those working in the field.
Author: Mary Lee Publisher: Springer Nature ISBN: 3030587487 Category : Mathematics Languages : en Pages : 134
Book Description
This volume features a variety of research projects at the intersection of mathematics and public policy. The topics included here fall in the areas of cybersecurity and climate change, two broad and impactful issues that benefit greatly from mathematical techniques. Each chapter in the book is a mathematical look into a specific research question related to one of these issues, an approach that offers the reader insight into the application of mathematics to important public policy questions. The articles in this volume are papers inspired by a Workshop for Women in Mathematics and Public Policy, held January 22-25, 2019 at the Institute for Pure and Applied Mathematics and the Luskin Center at the University of California, Los Angeles. The workshop was created to promote and develop women at all levels of their careers as researchers in mathematics and public policy. The idea was modeled after other successful Research Collaboration Conferences for Women, where junior and senior women come together at week-long conferences held at mathematics institutes to work on pre-defined research projects. The workshop focused on how mathematics can be used in public policy research and was designed to foster collaborative networks for women to help address the gender gap in mathematics and science.
Author: Heiko Ludwig Publisher: Springer Nature ISBN: 3030968960 Category : Computers Languages : en Pages : 531
Book Description
Federated Learning: A Comprehensive Overview of Methods and Applications presents an in-depth discussion of the most important issues and approaches to federated learning for researchers and practitioners. Federated Learning (FL) is an approach to machine learning in which the training data are not managed centrally. Data are retained by data parties that participate in the FL process and are not shared with any other entity. This makes FL an increasingly popular solution for machine learning tasks for which bringing data together in a centralized repository is problematic, either for privacy, regulatory or practical reasons. This book explains recent progress in research and the state-of-the-art development of Federated Learning (FL), from the initial conception of the field to first applications and commercial use. To obtain this broad and deep overview, leading researchers address the different perspectives of federated learning: the core machine learning perspective, privacy and security, distributed systems, and specific application domains. Readers learn about the challenges faced in each of these areas, how they are interconnected, and how they are solved by state-of-the-art methods. Following an overview on federated learning basics in the introduction, over the following 24 chapters, the reader will dive deeply into various topics. A first part addresses algorithmic questions of solving different machine learning tasks in a federated way, how to train efficiently, at scale, and fairly. Another part focuses on providing clarity on how to select privacy and security solutions in a way that can be tailored to specific use cases, while yet another considers the pragmatics of the systems where the federated learning process will run. The book also covers other important use cases for federated learning such as split learning and vertical federated learning. Finally, the book includes some chapters focusing on applying FL in real-world enterprise settings.
Author: Angela Daly Publisher: Lulu.com ISBN: 9492302284 Category : Data protection Languages : en Pages : 372
Book Description
Moving away from the strong body of critique of pervasive ?bad data? practices by both governments and private actors in the globalized digital economy, this book aims to paint an alternative, more optimistic but still pragmatic picture of the datafied future. The authors examine and propose ?good data? practices, values and principles from an interdisciplinary, international perspective. From ideas of data sovereignty and justice, to manifestos for change and calls for activism, this collection opens a multifaceted conversation on the kinds of futures we want to see, and presents concrete steps on how we can start realizing good data in practice.
Author: Xiaoxue Wang
Author: Xiaoxue Wang
Author: Ian Foster
Author: Michael Kearns
Author: Aileen Nielsen
Author: Andrea Vedaldi
Author: Helmut Degen
Author: K. Gal
Author: Mary Lee
Author: Heiko Ludwig
Author: Angela Daly