Michele Donini, left, a senior applied scientist with Amazon Web Services, and Luca Oneto, associate professor of computer engineering at University of Genoa
Michele Donini, left, a senior applied scientist with Amazon Web Services, and Luca Oneto, associate professor of computer engineering at University of Genoa, review different approaches that can make data-driven predictions fairer for underrepresented groups.
Credit: Glynis Condon

Working toward fairer machine learning

Exploring and analyzing possible techniques to make ML algorithms capable of learning fairer models by utilizing empirical risk minimization theory.

Editor’s note: Michele Donini is a senior applied scientist with Amazon Web Services (AWS). He and his co-author, Luca Oneto, associate professor of computer engineering at University of Genoa, have written about how different approaches can make data-driven predictions fairer for underrepresented groups. Oneto also won a 2019 Machine Learning Research award for his work on algorithmic fairness. In this article, Donini and Oneto explore the research they and other collaborators have published related to designing machine learning (ML) models from a human-centered perspective, and building responsible AI.

What is fairness?

Fairness can be defined in many different ways, and many different formal notions exist, such as demographic parity, equal opportunity, and equal odds.

Graphic shows a model of an unfair outcome on the left and a fair outcome on the right
Algorithmic fairness is a topic of great importance, with impact on many applications. The issue requires much further research; even the definition of what “being fair” means for an ML model is still an open research question.

Nevertheless, the basic and common idea behind notions of fairness is that the learned ML model should behave equivalently, or at least similarly, no matter whether it is applied to one subgroup of the population (e.g., males) or to another one (e.g., females).

For example, demographic parity, which arguably is the most common notion of fairness, implies that the probability of a certain output of an ML model (e.g., deciding to make a loan) should not depend on the value of specific demographic attributes (e.g., gender, race, or age).

Moving toward fairer models

Broadly speaking, we can group current literature on algorithmic fairness into three main approaches:

  • The first approach consists of pre-processing the data to remove historical biases and then feeding this data to classical ML models.
  • The second approach consists of post-processing an already learned ML model. This approach is useful when very complex ML models need to be made fairer without touching their inner structure or when re-training them is unfeasible (due to computational cost, or time requirements).
  • The third approach, called in-processing, consists of enforcing fairness notions by imposing specific statistical constraints during the learning phase of the model. This is the most natural approach, but so far, it has required ad hoc solutions tailored to specific tasks and data sets.
Fair Models.png
Broadly speaking, current literature on algorithmic fairness falls into three main approaches: pre-processing data; post-processing an already learned ML model; and in-processing, which consists of enforcing fairness notions by imposing specific statistical constraints during the learning phase of the model.

We decided to explore and analyze possible techniques to make ML algorithms capable of learning fairer models.

We started from the base concepts of statistical learning theory — a mathematical framework for describing machine learning — and, in particular, from empirical risk minimization theory. The core concept of empirical risk minimization is that a model’s performance on test data may not accurately predict its performance on real-world data, as the real-world data may have a different probability distribution.

Empirical-risk-minimization theory provides a way to estimate the “true risk” of a model from its “empirical risk”, which can be computed from the available data. We extended this concept to the true and empirical fairness risk of ML models.

Below is a summary of three papers we’ve published related to these topics.

Empirical risk minimization under fairness constraints

This paper presents a new in-processing method, meaning that we incorporate a fairness constraint into the learning problem. We derive theoretical guarantees on both the accuracy and fairness of the resulting models, and we show how to apply our method to a large family of machine learning algorithms, including linear models and support vector machines for classification (a widely used supervised-learning method).

We observe that, in practice, we can meet our fairness constraint simply by requiring that a scalar product between two vectors remains small (an orthogonality constraint between the vector of the weights describing our model and the vector describing the discrimination between the different subgroups). We further observe that, for linear models, this requirement translates into a simple pre-processing method. Experiments indicate that our approach is empirically effective and performs favorably against state-of-the-art approaches.

Fair regression with Wasserstein barycenters

In this paper, we consider the case in which the ML model learns a regression function (as opposed to a classification task). We propose a post-processing method for transforming a real-valued regression function — the ML model — into one that satisfies the demographic-parity constraint (i.e., the probability of getting a positive outcome should be virtually the same for different subgroups). In particular, the new regression function is as good an approximation of the original as is possible while still satisfying the constraint, making it an optimal fair predictor.

Fair Representation.png
In “Fair regression with Wasserstein barycenters”, we consider the case in which the ML model learns a regression function and propose a post-processing method for transforming a real-valued regression function — the ML model — into one that satisfies the demographic-parity constraint.

We assume that the sensitive attribute — the demographic attribute that should not bias outcome — is available to the ML model at inference time and not only during training. We establish a connection between learning a fair model for regression and optimal transport theory, which describes how to measure distances among probability distributions. On that basis, we derive a closed-form expression for the optimal fair predictor.

Specifically, under the unfair regression function, different populations have different probability distributions; the function skews the probabilities for the population with the sensitive attribute. The difference between subgroups’ distributions can be calculated using the Wasserstein distance. We show that the mean of the distribution of the optimal fair predictor is the mean of the different subgroups’ distributions, as calculated using Wasserstein distance. This mean is known as the Wasserstein barycenter.

This result offers an intuitive interpretation of optimal fair prediction and suggests a simple post-processing algorithm to achieve fairness. We establish fairness-risk guarantees for this procedure. Numerical experiments indicate that our method is very effective in learning fair models, with a relative increase in error rate that is smaller than the relative gain in fairness.

"Exploiting MMD and Sinkhorn divergences for fair and transferable representation learning

Where the first paper described a general learning method, and the second a regression method, this paper concerns deep learning. We show how to improve demographic parity in the multitask-learning setting, in which a deep-learning model learns a single representation of the input data that is useful for multiple tasks. We derive theoretical guarantees on the learned model, establishing that the representation will still reduce bias even when transferred to novel tasks.

We propose a learning algorithm that imposes constraints based on two different ways of measuring distances between probability distributions, maximum mean discrepancy and Sinkhorn divergence. Keeping this distance small ensures that we represent similar inputs in a similar way when they differ only on the sensitive attribute. We present experiments on three real-world datasets, showing that the proposed method outperforms state-of-the-art approaches by a significant margin.

Algorithmic fairness is a topic of great importance, with impact on many applications. In our work, we have attempted to take a small step forward, but the issue requires much further research; even the definition of what “being fair” means for an ML model is still an open research question.

It’s also becoming clearer that we need to keep humans in the loop during the lifecycle of ML models, to evaluate whether the models are acting as we would like them to. In this sense, it is important to note that many other research subjects – such as the explainability, interpretability, and privacy of ML models – are deeply connected to algorithmic fairness. They can work in synergy, with the common goal of increasing the trustworthiness of ML models.

View from space of a connected network around planet Earth representing the Internet of Things.
Sign up for our newsletter

Research areas

Related content

US, CA, Palo Alto
The Amazon Search team creates powerful, customer-focused search and advertising solutions and technologies. Whenever a customer visits an Amazon site worldwide and types in a query or browses through product categories, the Amazon Search services go to work. We design, develop, and deploy high performance, fault-tolerant distributed search systems used by millions of Amazon customers every day. Our team works to maximize the quality and effectiveness of the search experience for visitors to Amazon websites worldwide.
JP, Tokyo
The Amazon Logistics (AMZL) Team is responsible for the acquisition, design, construction, and management of all facilities in the Amazon Delivery Station Network. AMZL is looking for a talented and passionate Data Scientist to help shape its Last Mile business with technical strategies and solutions, by processing, analyzing and interpreting huge data sets. You should be comfortable with ambiguity, problem solving and enjoy working in a fast-paced, diverse and dynamic environment. Using analytical rigor and statistical methods, you mine through data to identify opportunities for Amazon and our delivery channels. And you collaborate with other scientists, engineers, Product and Program Managers to deploy new products and solutions. [More Information] Last Mile Department Data Analyst/BI Engineer Tokyo Office *Amazon is committed to a diverse and inclusive workplace. Amazon is an equal opportunity employer and does not discriminate on the basis of race, national origin, gender, gender identity, sexual orientation, protected veteran status, disability, age, or other legally protected status. For individuals with disabilities who would like to request an accommodation, visit https://www.amazon.jobs/disability/jp Key job responsibilities Creating a roadmap of the most challenging business questions and use data to articulate possible root cause analysis and solutions Managing and executing entire projects or components of large projects from start to finish including project management, data gathering and manipulation, synthesis and modeling, problem solving, and communication of insights Partnering with Product, Program and Engineering teams to design and run models, research new algorithms, and prove incrementality and drive growth Understanding drivers, impacts, and key influences on seller growth dynamics Developing and scaling end-to-end ML Models and solutions Automating feedback loops for algorithms in production Utilizing Amazon systems and tools to effectively work with terabytes of data About the team Last Mile Execution Analytics (LMEA) team of JP works as an integral part of Amazon Logistics to ensure that its business intelligence, analytics, tools and planning needs are met. By providing information, insight, and decision support, we strive to enable success of all parts of AMZL. Our customer set includes senior management, station operations, external vendors, long-term planning, Ops technology (Voice of the Delivery Station, Voice of the Customer), network planning, and pretty much every BI and Ops teams. Voice of Employee [Work Life Harmony] We believe, it is important to spend private time such as spending time with your family or doing anything you like to spur innovation. Amazon promotes a fulfilling and flexible work style according to the work volume and lifestyle of each employee.
US, CA, San Francisco
About Twitch Launched in 2011, Twitch is a global community that comes together each day to create multiplayer entertainment: unique, live, unpredictable experiences created by the interactions of millions. We bring the joy of co-op to everything, from casual gaming to world-class esports to anime marathons, music, and art streams. Twitch also hosts TwitchCon, where we bring everyone together to celebrate, learn, and grow their personal interests and passions. We’re always live at Twitch. Stay up to date on all things Twitch on Linkedin, Twitter and on our Blog. About the role: Twitch builds data-driven machine learning solutions across several rich problem spaces: Natural Language Processing (NLP), Recommendations, Semantic Search, Classification/Categorization, Anomaly Detection, Forecasting, Safety, and HCI/Social Computing/Computational Social Science. As an Intern, you will work with a dedicated Mentor and Manager on a project in one of these problem areas. You will also be supported by an Advisor and participate in cohort activities such as research teach backs and leadership talks. This position can also be located in San Francisco, CA or virtual. You Will: Solve large-scale data problems. Design solutions for Twitch's problem spaces Explore ML and data research
US, CA, San Francisco
About Twitch Launched in 2011, Twitch is a global community that comes together each day to create multiplayer entertainment: unique, live, unpredictable experiences created by the interactions of millions. We bring the joy of co-op to everything, from casual gaming to world-class esports to anime marathons, music, and art streams. Twitch also hosts TwitchCon, where we bring everyone together to celebrate, learn, and grow their personal interests and passions. We’re always live at Twitch. Stay up to date on all things Twitch on Linkedin, Twitter and on our Blog. About the role: Twitch builds data-driven machine learning solutions across several rich problem spaces: Natural Language Processing (NLP), Recommendations, Semantic Search, Classification/Categorization, Anomaly Detection, Forecasting, Safety, and HCI/Social Computing/Computational Social Science. As an Intern, you will work with a dedicated Mentor and Manager on a project in one of these problem areas. You will also be supported by an Advisor and participate in cohort activities such as research teach backs and leadership talks. This position can also be located in San Francisco, CA or virtual. You Will: Solve large-scale data problems. Design solutions for Twitch's problem spaces Explore ML and data research
LU, Luxembourg
Are you a talented and inventive scientist with a strong passion about modern data technologies and interested to improve business processes, extracting value from the data? Would you like to be a part of an organization that is aiming to use self-learning technology to process data in order to support the management of the procurement function? The Global Procurement Technology, as a part of Global Procurement Operations, is seeking a skilled Data Scientist to help build its future data intelligence in business ecosystem, working with large distributed systems of data and providing Machine Learning (ML) and Predictive Modeling expertise. You will be a member of the Data Engineering and ML Team, joining a fast-growing global organization, with a great vision to transform the Procurement field, and become the role model in the market. This team plays a strategic role supporting the core Procurement business domains as well as it is the cornerstone of any transformation and innovation initiative. Our mission is to provide a high-quality data environment to facilitate process optimization and business digitalization, on a global scale. We are supporting business initiatives, including but not limited to, strategic supplier sourcing (e.g. contracting, negotiation, spend analysis, market research, etc.), order management, supplier performance, etc. We are seeking an individual who can thrive in a fast-paced work environment, be collaborative and share knowledge and experience with his colleagues. You are expected to deliver results, but at the same time have fun with your teammates and enjoy working in the company. In Amazon, you will find all the resources required to learn new skills, grow your career, and become a better professional. You will connect with world leaders in your field and you will be tackling Data Science challenges to ensure business continuity, by taking the right decisions for your customers. As a Data Scientist in the team, you will: -be the subject matter expert to support team strategies that will take Global Procurement Operations towards world-class predictive maintenance practices and processes, driving more effective procurement functions, e.g. supplier segmentation, negotiations, shipping supplies volume forecast, spend management, etc. -have strong analytical skills and excel in the design, creation, management, and enterprise use of large data sets, combining raw data from different sources -provide technical expertise to support the development of ML models to facilitate intelligent digital services, such as Contract Lifecycle Management (CLM) and Negotiations platform -cooperate closely with different groups of stakeholders, e.g. data/software engineers, product/program managers, analysts, senior leadership, etc. to evaluate business needs and objectives to set up the best data management environment -create and share with audiences of varying levels technical papers and presentations -deal with ambiguity, prioritizing needs, and delivering results in a dynamic environment Basic qualifications -Master’s Degree in Computer Science/Engineering, Informatics, Mathematics, or a related technical discipline -3+ years of industry experience in data engineering/science, business intelligence or related field -3+ years experience in algorithm design, engineering and implementation for very-large scale applications to solve real problems -Very good knowledge of data modeling and evaluation -Very good understanding of regression modeling, forecasting techniques, time series analysis, machine-learning concepts such as supervised and unsupervised learning, classification, random forest, etc. -SQL and query performance tuning skills Preferred qualifications -2+ years of proficiency in using R, Python, Scala, Java or any modern language for data processing and statistical analysis -Experience with various RDBMS, such as PostgreSQL, MS SQL Server, MySQL, etc. -Experience architecting Big Data and ML solutions with AWS products (Redshift, DynamoDB, Lambda, S3, EMR, SageMaker, Lex, Kendra, Forecast etc.) -Experience articulating business questions and using quantitative techniques to arrive at a solution using available data -Experience with agile/scrum methodologies and its benefits of managing projects efficiently and delivering results iteratively -Excellent written and verbal communication skills including data visualization, especially in regards to quantitative topics discussed with non-technical colleagues
US, CA, San Francisco
About Twitch Launched in 2011, Twitch is a global community that comes together each day to create multiplayer entertainment: unique, live, unpredictable experiences created by the interactions of millions. We bring the joy of co-op to everything, from casual gaming to world-class esports to anime marathons, music, and art streams. Twitch also hosts TwitchCon, where we bring everyone together to celebrate, learn, and grow their personal interests and passions. We’re always live at Twitch. Stay up to date on all things Twitch on Linkedin, Twitter and on our Blog. About the role: Twitch builds data-driven machine learning solutions across several rich problem spaces: Natural Language Processing (NLP), Recommendations, Semantic Search, Classification/Categorization, Anomaly Detection, Forecasting, Safety, and HCI/Social Computing/Computational Social Science. As an Intern, you will work with a dedicated Mentor and Manager on a project in one of these problem areas. You will also be supported by an Advisor and participate in cohort activities such as research teach backs and leadership talks. This position can also be located in San Francisco, CA or virtual. You Will: Solve large-scale data problems. Design solutions for Twitch's problem spaces Explore ML and data research
US, CA, San Francisco
About Twitch Launched in 2011, Twitch is a global community that comes together each day to create multiplayer entertainment: unique, live, unpredictable experiences created by the interactions of millions. We bring the joy of co-op to everything, from casual gaming to world-class esports to anime marathons, music, and art streams. Twitch also hosts TwitchCon, where we bring everyone together to celebrate, learn, and grow their personal interests and passions. We’re always live at Twitch. Stay up to date on all things Twitch on Linkedin, Twitter and on our Blog. About the role: Twitch builds data-driven machine learning solutions across several rich problem spaces: Natural Language Processing (NLP), Recommendations, Semantic Search, Classification/Categorization, Anomaly Detection, Forecasting, Safety, and HCI/Social Computing/Computational Social Science. As an Intern, you will work with a dedicated Mentor and Manager on a project in one of these problem areas. You will also be supported by an Advisor and participate in cohort activities such as research teach backs and leadership talks. This position can also be located in San Francisco, CA or virtual. You Will: Solve large-scale data problems. Design solutions for Twitch's problem spaces Explore ML and data research
US, CA, San Francisco
About Twitch Launched in 2011, Twitch is a global community that comes together each day to create multiplayer entertainment: unique, live, unpredictable experiences created by the interactions of millions. We bring the joy of co-op to everything, from casual gaming to world-class esports to anime marathons, music, and art streams. Twitch also hosts TwitchCon, where we bring everyone together to celebrate, learn, and grow their personal interests and passions. We’re always live at Twitch. Stay up to date on all things Twitch on Linkedin, Twitter and on our Blog. About the role: Twitch builds data-driven machine learning solutions across several rich problem spaces: Natural Language Processing (NLP), Recommendations, Semantic Search, Classification/Categorization, Anomaly Detection, Forecasting, Safety, and HCI/Social Computing/Computational Social Science. As an Intern, you will work with a dedicated Mentor and Manager on a project in one of these problem areas. You will also be supported by an Advisor and participate in cohort activities such as research teach backs and leadership talks. This position can also be located in San Francisco, CA or virtual. You Will: Solve large-scale data problems. Design solutions for Twitch's problem spaces Explore ML and data research
US, CA, San Francisco
About Twitch Launched in 2011, Twitch is a global community that comes together each day to create multiplayer entertainment: unique, live, unpredictable experiences created by the interactions of millions. We bring the joy of co-op to everything, from casual gaming to world-class esports to anime marathons, music, and art streams. Twitch also hosts TwitchCon, where we bring everyone together to celebrate, learn, and grow their personal interests and passions. We’re always live at Twitch. Stay up to date on all things Twitch on Linkedin, Twitter and on our Blog. About the role: Twitch builds data-driven machine learning solutions across several rich problem spaces: Natural Language Processing (NLP), Recommendations, Semantic Search, Classification/Categorization, Anomaly Detection, Forecasting, Safety, and HCI/Social Computing/Computational Social Science. As an Intern, you will work with a dedicated Mentor and Manager on a project in one of these problem areas. You will also be supported by an Advisor and participate in cohort activities such as research teach backs and leadership talks. This position can also be located in San Francisco, CA or virtual. You Will: Solve large-scale data problems. Design solutions for Twitch's problem spaces Explore ML and data research
US, WA, Seattle
We are a team of doers working passionately to apply cutting-edge advances in deep learning in the life sciences to solve real-world problems. As a Senior Applied Science Manager you will participate in developing exciting products for customers. Our team rewards curiosity while maintaining a laser-focus in bringing products to market. Competitive candidates are responsive, flexible, and able to succeed within an open, collaborative, entrepreneurial, startup-like environment. At the leading edge of both academic and applied research in this product area, you have the opportunity to work together with a diverse and talented team of scientists, engineers, and product managers and collaborate with others teams. Location is in Seattle, US Embrace Diversity Here at Amazon, we embrace our differences. We are committed to furthering our culture of inclusion. We have ten employee-led affinity groups, reaching 40,000 employees in over 190 chapters globally. We have innovative benefit offerings, and host annual and ongoing learning experiences, including our Conversations on Race and Ethnicity (CORE) and AmazeCon (gender diversity) conferences. Amazon’s culture of inclusion is reinforced within our 14 Leadership Principles, which remind team members to seek diverse perspectives, learn and be curious, and earn trust Balance Work and Life Our team puts a high value on work-life balance. It isn’t about how many hours you spend at home or at work; it’s about the flow you establish that brings energy to both parts of your life. We believe striking the right balance between your personal and professional life is critical to life-long happiness and fulfillment. We offer flexibility in working hours and encourage you to find your own balance between your work and personal lives Mentor & Grow Careers Our team is dedicated to supporting new members. We have a broad mix of experience levels and tenures, and we’re building an environment that celebrates knowledge sharing and mentorship. Our senior members enjoy one-on-one mentoring and thorough, but kind, code reviews. We care about your career growth and strive to assign projects based on what will help each team member develop into a better-rounded engineer and enable them to take on more complex tasks in the future. Key job responsibilities • Manage high performing engineering and science teams • Hire and develop top-performing engineers, scientists, and other managers • Develop and execute on project plans and delivery commitments • Work with business, data science, software engineer, biological, and product leaders to help define product requirements and with managers, scientists, and engineers to execute on them • Build and maintain world-class customer experience and operational excellence for your deliverables