Advances in trustworthy machine learning at Alexa AI

The team’s latest research on privacy-preserving machine learning, federated learning, and bias mitigation.

At Amazon, we take the protection of customer data very seriously. We are also committed to eliminating the biases that can exist in off-the-shelf language models — such as GPT-3 and RoBERTa — that are the basis of most modern natural-language processing. Trained on public texts, these language models are known to reflect the biases implicit in those texts.

Related content
Calibrating noise addition to word density in the embedding space improves utility of privacy-protected text.

These two topics — privacy protection and fairness — are at the core of trustworthy machine learning, an important area of research at Alexa AI. In 2021, we made contributions in the following areas:

  • Privacy-preserving machine learningDifferential privacy provides a rigorous way to quantify the privacy of machine learning models. We investigated vulnerabilities presented in the differential-privacy literature and propose computationally efficient mechanisms for protecting against them.
  • Federated learning: Federated learning (FL) is a distributed-training technique that keeps customer data on-device. Devices send only model parameter updates to the cloud, not raw data. We studied several FL challenges arising in an industrial setting.
  • Fairness in machine learning: Machine learning (ML) models should perform equally well regardless of who’s using them. But even knowing how to quantify fairness is a challenge. We introduced measures of fairness and methods to mitigate bias in ML models.
Counterfactuals.png
To reduce binary-gender disparity in a distilled GPT-2 language model, we introduce counterfactual examples, in which binary genders in real-world training examples are swapped.

Below, we summarize our research in these areas, which will be presented at ACL and ICASSP later this year. We also invite readers to participate in workshops and sessions we are organizing at NAACL 2022 and Interspeech 2022.

1. Privacy-preserving ML

The intuition behind differential privacy (DP) is that access to the outputs of a model should not provide any hint about what inputs were used to train the model. DP quantifies that intuition as a difference (in probabilities) between the outputs of a model trained on a given dataset and the outputs of the same model trained on the same dataset after a single input is removed.

One way to meet a DP privacy guarantee is to add some noise to the model parameters during training in order to obfuscate their relationship to training data. But this can compromise accuracy. The so-called privacy/utility tradeoff appears in every DP application.

Another side effect of adding a DP mechanism is increased training time. Given that training natural-language-understanding (NLU) models with large volumes of data can be prohibitively slow and that industry standards require fast training and deployment — e.g., when new features are being released — we developed a training method that meets DP requirements but remains efficient. We describe the method in a paper we’re presenting at this year’s ICASSP, “An efficient DP-SGD mechanism for large scale NLP models”.

In this work, we study the most popular DP mechanism for deep neural networks, DP-SGD, and build a computationally efficient alternative, eDP-SGD, in which we use a batch-processing scheme that leverages the GPU architecture and automates part of the hyperparameter-tuning process. While both DP-SGD and eDP-SGD provide the same privacy guarantees, we show that the training time for our mechanism is very similar to its non-DP counterpart’s. The original DP-SGD extends training time as much as 130-fold.

Related content
ADePT model transforms the texts used to train natural-language-understanding models while preserving semantic coherence.

Since we did our study, researchers have developed methods with stronger theoretical DP guarantees than the ones we impose in our paper, but our approach is consistent with those methods. Overall, this work makes DP more generally accessible and helps us integrate NLU models with DP guarantees into our production systems, where new models are frequently released, and a significant increase in training time is prohibitive.

While DP provides theoretical privacy guarantees, we are also interested in practical guarantees, i.e., measuring the amount of information that could potentially leak from a given model. In addition to the performance and training time of eDP-SGD, we also studied the correlation between theoretical and practical privacy guarantees. We measured practical privacy leakage using the most common method in the field, the success rate of membership inference attacks on a given model. Our experiments provide a general picture of how to optimize the privacy/utility trade-off using DP techniques for NLU models.

We also expanded the set of mechanisms for protecting NLU models against other types of attacks. In “Canary extraction in natural language understanding models”, which we will present at ACL 2022, we study the vulnerability of text classification models to a certain kind of white-box attack called a model inversion attack (ModIvA), where a fictional attack has access to the entire set of model parameters and intends to retrieve examples used during training. Existing model inversion techniques are applied to models with either continuous inputs or continuous outputs. In our work, we adopt a similar approach to text classification tasks where both inputs and outputs are discrete.

As new model architectures are developed that might display new types of vulnerabilities, we will continue innovating efficient ways of protecting our customers’ privacy.

Upcoming activities

2. Federated Learning

The idea behind federated learning (FL) is that, during the training of an ML model, part of the computation is delegated to customers’ devices, leveraging the processing power of those devices while avoiding the centralization of privacy-sensitive datasets. Each device modifies a common, shared model according to locally stored data, then sends an updated model to a central server that aggregates model updates and sends a new shared model to all the devices. At each round, the central server randomly selects a subset of active devices and requests that they perform updates.

Federated Learning Animation.gif
With federated learning, devices send model updates, not data, to a central server.

In the past year, we have made progress toward more-efficient FL and adapted common FL techniques to the industrial setting. For instance, in “Learnings from federated learning in the real world”, which we will present at ICASSP this year, we explore device selection strategies that differ from the standard uniform selection. In particular, we present the first study of device selection based on device “activity” — i.e., the number of available training samples.

These simple selection strategies are lightweight compared to existing methods, which require heavy computation from all the devices. They are thus more suitable to industrial applications, where millions of devices are involved. We study two different settings: the standard “static” setting, where all the data are available at once, and the more realistic “continual” setting, where customers generate new data over time, and past examples might have to be deleted to save storage space. Our experiments on training a language model with FL show that non-uniform sampling outperforms uniform sampling when applied to real-world data, for both the static and continual settings.

Related content
Amazon researchers optimize the distributed-training tool to run efficiently on the Elastic Fabric Adapter network interface.

We also expanded our understanding of FL for natural-language processing (NLP) and, in the process, made FL more accessible to the NLP community. In “FedNLP: A research platform for federated learning in natural language processing”, which will be presented later this year at NAACL, we and our colleagues at the University of Southern California and FedML systematically compare the most popular FL algorithms for four mainstream NLP tasks. We also present different methods to generate dataset partitions that are not independent and identically distributed (IID), as real-world FL methods must be robust against shifts in the distributions of the data used to train ML models.

Our analysis reveals that there is still a large gap between centralized and decentralized training under various settings, and we highlight several directions in which FL for NLP can advance. The paper represents Amazon’s contribution to the open-source framework FedNLP, which is capable of evaluating, analyzing, and developing FL methods for NLP. The codebase contains non-IID partitioning methods, enabling easy experimentation to advance the state of FL research for NLP.

We also designed methods to account for the naturally heterogeneous character of customer-generated data and applied FL to a wide variety of NLP tasks. We are aware that FL still presents many challenges, such as how to do evaluation when access to data is removed, on-device label generation for supervised tasks, and privacy-preserving communication between the server and the different devices. We are actively addressing each of these and plan to leverage our findings to improve FL-based model training and enhance associated capabilities such as analytics and model evaluation.

Upcoming activities

3. Fairness in ML

Natural-language-processing applications’ increased reliance on large language models trained on intrinsically biased web-scale corpora has amplified the importance of accurate fairness metrics and procedures for building more robust models.

In “On the intrinsic and extrinsic fairness evaluation metrics for contextualized language representations”, which we are presenting at ACL 2022, we compare two families of fairness metrics — namely extrinsic and intrinsic — that are widely used for language models. Intrinsic metrics directly probe into the fairness of language models, while extrinsic metrics evaluate the fairness of a whole system through predictions on downstream tasks.

Related content
Method significantly reduces bias while maintaining comparable performance on machine learning tasks.

For example, the contextualized embedding association test (CEAT), an intrinsic metric, measures bias through word embedding distances in semantic vector spaces, and the extrinsic metric HateXPlain measures the bias in a downstream hate speech detection system.

Our experiments show that inconsistencies between intrinsic and extrinsic metrics often reflect inconsistencies between the datasets used to evaluate them, and a clear understanding of bias in ML models requires more careful alignment of evaluation data. The results we report in the paper can help guide the NLP community as to how to best conduct fairness evaluations.

We have also designed new measures of fairness that are adapted to language-processing applications. In “Measuring fairness of text classifiers via prediction sensitivity”, which we will present at ACL 2022, we looked at sensitivity to perturbations of input as a way to measure fairness in ML models. The metric attempts to quantify the extent to which a single prediction depends on an input feature that encodes membership in an underrepresented group.

Accumulated prediction sensitivity.png
Our new bias measure, accumulated prediction sensitivity, combines the outputs of tow models, a task classifier (TC) and a protected status model (PSM).

We provide a theoretical analysis of our formulation and show a statistically significant difference between our metric’s correlation with the human notion of fairness and the existing counterfactual fairness metric’s.

Finally, we proposed a method to mitigate the biases of large language models during knowledge distillation, in which a smaller, more efficient model is trained to match the language model’s output on a particular task. Because large language models are trained on public texts, they can be biased in multiple ways, including the unfounded association of male or female genders with gender-neutral professions.

Distillation examples.png
Examples of texts generated by language models in response to gendered prompts before and after the application of our distillation method.

In another ACL paper, “Mitigating gender bias in distilled language models via counterfactual role reversal”, we introduce two modifications to the standard distillation mechanisms: data augmentation and teacher prediction perturbation.

We use our method to distill a GPT-2 language model for a text-generation task and demonstrate a substantial reduction in gender disparity, with only a minor reduction in utility. Interestingly, we find that reduced disparity in open-ended text generation may not necessarily lead to fairness on other downstream tasks. This finding underscores the importance of evaluating language model fairness along multiple metrics and tasks.

Our work on fairness in ML for NLP applications should help enable models that are more robust against the inherent biases of text datasets. There remain plenty of challenges in this field, but we strive to build models that offer the same experience to any customer, wherever and however they choose to interact with Alexa.

Upcoming activities

Related content

US, WA, Bellevue
Conversational AI ModEling and Learning (CAMEL) team is part of Amazon Devices organization where our mission is to build a best-in-class Conversational AI that is intuitive, intelligent, and responsive, by developing superior Large Language Models (LLM) solutions and services which increase the capabilities built into the model and which enable utilizing thousands of APIs and external knowledge sources to provide the best experience for each request across millions of customers and endpoints. We are looking for a passionate, talented, and resourceful Applied Scientist in the field of LLM, Artificial Intelligence (AI), Natural Language Processing (NLP), Recommender Systems and/or Information Retrieval, to invent and build scalable solutions for a state-of-the-art context-aware conversational AI. A successful candidate will have strong machine learning background and a desire to push the envelope in one or more of the above areas. The ideal candidate would also have hands-on experiences in building Generative AI solutions with LLMs, enjoy operating in dynamic environments, be self-motivated to take on challenging problems to deliver big customer impact, moving fast to ship solutions and then iterating on user feedback and interactions. Key job responsibilities As an Applied Scientist, you will leverage your technical expertise and experience to collaborate with other talented applied scientists and engineers to research and develop novel algorithms and modeling techniques to reduce friction and enable natural and contextual conversations. You will analyze, understand and improve user experiences by leveraging Amazon’s heterogeneous data sources and large-scale computing resources to accelerate advances in artificial intelligence. You will work on core LLM technologies, including Prompt Engineering and Optimization, Supervised Fine-Tuning, Learning from Human Feedback, Evaluation, Self-Learning, etc. Your work will directly impact our customers in the form of novel products and services.
IL, Tel Aviv
Are you an inventive, curious, and driven Applied Scientist with a strong background in AI and Deep Learning? Join Amazon’s AWS Multimodal generative AI science team and be a catalyst for groundbreaking advancements in Computer Vision, Generative AI, and foundational models. As part of the AWS Multimodal generative AI science team, you’ll lead innovative research projects, develop state-of-the-art algorithms, and pioneer solutions that will directly impact millions of Amazon customers. Leveraging Amazon’s vast computing power, you’ll work alongside a supportive and diverse group of top-tier scientists and engineers, contributing to products that redefine the industry. Key job responsibilities * Lead research initiatives in Multimodal generative AI, pushing the boundaries of model efficiency, accuracy, and scalability. * Design, implement, and evaluate deep learning models in a production environment. * Collaborate with cross-functional teams to transfer research outcomes into scalable AWS services. * Publish in top-tier conferences and journals, keeping Amazon at the forefront of innovation. * Mentor and guide other scientists and engineers, fostering a culture of scientific curiosity and excellence.
GB, Cambridge
The Artificial General Intelligence team (AGI) has an exciting position for an Applied Scientist with a strong background NLP and Large Language Models to help us develop state-of-the-art conversational systems. As part of this team, you will collaborate with talented scientists and software engineers to enable conversational assistants capabilities to support the use of external tools and sources of information, and develop novel reasoning capabilities to revolutionise the user experience for millions of Alexa customers. Key job responsibilities As an Applied Scientist, you will develop innovative solutions to complex problems to extend the functionalities of conversational assistants . You will use your technical expertise to research and implement novel algorithms and modelling solutions in collaboration with other scientists and engineers. You will analyse customer behaviours and define metrics to enable the identification of actionable insights and measure improvements in customer experience. You will communicate results and insights to both technical and non-technical audiences through written reports, presentations and external publications.
US, WA, Bellevue
Conversational AI ModEling and Learning (CAMEL) team is part of Amazon Artificial General Intelligence (AGI) organization where our mission is to create a best-in-class Conversational AI that is intuitive, intelligent, and responsive, by developing superior Large Language Models (LLM) solutions and services which increase the capabilities built into the model and which enable utilizing thousands of APIs and external knowledge sources to provide the best experience for each request across millions of customers and endpoints. We are looking for a passionate, talented, and resourceful Applied Scientist in the field of LLM, Artificial Intelligence (AI), Natural Language Processing (NLP), Recommender Systems and/or Information Retrieval, to invent and build scalable solutions for a state-of-the-art context-aware conversational AI. A successful candidate will have strong machine learning background and a desire to push the envelope in one or more of the above areas. The ideal candidate would also have hands-on experiences in building Generative AI solutions with LLMs, enjoy operating in dynamic environments, be self-motivated to take on challenging problems to deliver big customer impact, moving fast to ship solutions and then iterating on user feedback and interactions. Key job responsibilities As an Applied Scientist, you will leverage your technical expertise and experience to collaborate with other talented applied scientists and engineers to research and develop novel algorithms and modeling techniques to reduce friction and enable natural and contextual conversations. You will analyze, understand and improve user experiences by leveraging Amazon’s heterogeneous data sources and large-scale computing resources to accelerate advances in artificial intelligence. You will work on core LLM technologies, including Supervised Fine-Tuning (SFT), In-Context Learning (ICL), Learning from Human Feedback (LHF), etc. Your work will directly impact our customers in the form of novel products and services.
US, WA, Seattle
We are seeking a highly skilled economist to measure and understand how each Customer Service activity impacts customers. This candidate's analysis will assist teams across Amazon to prioritize defect elimination efforts and optimize how we respond to customer contacts. This candidate will partner closely with our product, program, and tech teams to deliver their findings to users via systems and dashboards that guide Customer Service planning and policy rules. Key job responsibilities - Develop Causal, Economic, and Machine Learning models at scale. - Engage in economic analysis and raise the bar for research. - Inform strategic discussions with senior leaders across the company to guide policies. A day in the life If you are not sure that every qualification on the list above describes you exactly, we'd still love to hear from you! At Amazon, we value people with unique backgrounds, experiences, and skillsets. If you’re passionate about this role and want to make an impact on a global scale, please apply! Amazon offers a full range of benefits that support you and eligible family members, including domestic partners and their children. Benefits can vary by location, the number of regularly scheduled hours you work, length of employment, and job status such as seasonal or temporary employment. The benefits that generally apply to regular, full-time employees include: 1. Medical, Dental, and Vision Coverage 2. Maternity and Parental Leave Options 3. Paid Time Off (PTO) 4. 401(k) Plan About the team The Worldwide defect elimination team's mission is to understand and resolve all issues impacting customers at scale. The Customer Service Economics and Optimization team is a force multiplier within this group, helping to understand the impact of these issues and our actions to optimize the customer experience.
US, WA, Seattle
We are building GenAI based shopping assistant for Amazon. We reimage Amazon Search with an interactive conversational experience that helps you find answers to product questions, perform product comparisons, receive personalized product suggestions, and so much more, to easily find the perfect product for your needs. We’re looking for the best and brightest across Amazon to help us realize and deliver this vision to our customers right away. This will be a once in a generation transformation for Search, just like the Mosaic browser made the Internet easier to engage with three decades ago. If you missed the 90s—WWW, Mosaic, and the founding of Amazon and Google—you don’t want to miss this opportunity.
US, WA, Seattle
At Amazon, we believe that scientific innovation is essential to being the most customer-centric company in the world. Our scientists' ability to have an impact at scale allows us to attract some of the brightest minds in machine learning, artificial intelligence and related fields. Amazon scientists employ the company's working backwards method to identify problems to solve on behalf of customers in research areas ranging from machine learning to operations, GenAI, robotics, quantum computing, computer vision, economics, search, sustainability and more. Learn more about Amazon Science here: https://www.amazon.science/ We are hiring across multiple businesses and in many locations across the US. Apply here to learn more about open roles that could be a compelling fit for your background. Key job responsibilities You will be responsible for defining key research directions, adopting or inventing new machine learning techniques, conducting rigorous experiments, publishing results, and ensuring that research is translated into practice. You will develop long-term strategies, persuade teams to adopt those strategies, propose goals and deliver on them. You will also participate in organizational planning, hiring, mentorship and leadership development. You will be technically fearless and with a passion for building scalable science and engineering solutions. You will serve as a key scientific resource in full-cycle development (conception, design, implementation, testing to documentation, delivery, and maintenance).
NL, Amsterdam
Are you a MS or PhD student interested in a 2025 Internship in the field of machine learning, deep learning, speech, robotics, computer vision, optimization, quantum computing, automated reasoning, or formal methods? If so, we want to hear from you! We are looking for students interested in using a variety of domain expertise to invent, design and implement state-of-the-art solutions for never-before-solved problems. You can find more information about the Amazon Science community as well as our interview process via the links below; https://www.amazon.science/ https://amazon.jobs/content/en/career-programs/university/science https://amazon.jobs/content/en/how-we-hire/university-roles/applied-science Key job responsibilities As an Applied Science Intern, you will own the design and development of end-to-end systems. You’ll have the opportunity to write technical white papers, create roadmaps and drive production level projects that will support Amazon Science. You will work closely with Amazon scientists, and other science interns to develop solutions and deploy them into production. You will have the opportunity to design new algorithms, models, or other technical solutions whilst experiencing Amazon’s customer focused culture. The ideal intern must have the ability to work with diverse groups of people and cross-functional teams to solve complex business problems. A day in the life At Amazon, you will grow into the high impact, visionary person you know you’re ready to be. Every day will be filled with developing new skills and achieving personal growth. How often can you say that your work changes the world? At Amazon, you’ll say it often. Join us and define tomorrow. Some more benefits of an Amazon Science internship include; • All of our internships offer a competitive stipend/salary • Interns are paired with an experienced manager and mentor(s) • Interns receive invitations to different events such as intern program initiatives or site events • Interns can build their professional and personal network with other Amazon Scientists • Interns can potentially publish work at top tier conferences each year About the team Applicants will be reviewed on a rolling basis and are assigned to teams aligned with their research interests and experience prior to interviews. Start dates are available throughout the year and durations can vary in length from 3-6 months for full time internships. This role may available across multiple locations in the EMEA region (Austria, Estonia, France, Germany, Ireland, Israel, Italy, Luxembourg, Netherlands, Poland, Romania, Spain, UAE, and UK). Please note these are not remote internships.
US, WA, Seattle
Come be a part of a rapidly expanding $35 billion-dollar global business. At Amazon Business, a fast-growing startup passionate about building solutions, we set out every day to innovate and disrupt the status quo. We stand at the intersection of tech & retail in the B2B space developing innovative purchasing and procurement solutions to help businesses and organizations thrive. At Amazon Business, we strive to be the most recognized and preferred strategic partner for smart business buying. Bring your insight, imagination and a healthy disregard for the impossible. Join us in building and celebrating the value of Amazon Business to buyers and sellers of all sizes and industries. Unlock your career potential. The AB Sales Analytics, Data, Product and Tech (ADAPTech) team uses CRM, data, product, and science to improve Sales productivity and performance. It has four pillars: 1) SalesTech maintains Salesforce to enable Sales workflows, and supports >2K users in nine countries; 2) Product and Science builds tools embedded with bespoke Machine Learning (ML) and GenAI large language models to enable sales reps to prioritize top accounts, position the right Amazon Business (AB) product features, and take actions based on critical customer events; 3) Sales Data Management (SDM) and Sales Account Management (SAM) enrich customer profiles and business hierarchies while improving productivity through automation and integration of internal/external tools; and 4) Business Intelligence (BI) enables self-service reporting simplifying access to key insights through WBRs and dashboards. Sales teams leverage these products to identify which customers to target, what features to target them with, and when to target them, in order to capture their share of wallet. A successful Applied Scientist at Amazon demonstrates bias for action and operates in a startup environment, with outstanding leadership skills, and proven ability to build and manage medium-scale modeling projects, identify data requirements, build methodology and tools that are statistically grounded. We need great leaders to think big and design new solutions to solve complex problems using machine learning (ML) and Generative AI techniques to improve our customers’ experience when using AB. You have hands-on experience making the right decisions about technology, models and methodology choices. Key job responsibilities As an Applied Scientist, you will primarily leverage machine learning techniques and generative AI to outreach customers based on their life cycle stage, behavioral patterns, and purchase history. You may also perform text mining and insight analysis of real-time customer conversations and make the model learn and recommend the solutions. Your work will directly impact the trust customers place in Amazon Business. You will partner with product management and technical leadership to identify opportunities to innovate customer journey experiences. You will identify new areas of investment and work to align product roadmaps to deliver on these opportunities. As a science leader, you will not only develop unique scientific solutions, but also play a crucial role in shaping strategies. Additional responsibilities include: -Design, implement, test, deploy and maintain innovative data and machine learning solutions to further the customer experience. -Create experiments and prototype implementations of new learning algorithms and prediction techniques -Develop algorithms for new capabilities and trace decisions in the data and assess how proposed changes could potentially impact business metrics to cater needs of Amazon Business Sales -Build models that measure incremental value, predict growth, define and conduct experiments to optimize engagement of AB customers, and communicate insights and recommendations to product, sales, and finance partners. A day in the life In this role, you will be a technical expert with significant scope and impact. You will work with Technical Product Managers, Data Engineers, other Scientists, and Salesforce developers, to build new and enhance existing ML models to optimize customer experience. You will prototype and test new ideas, iterate quickly, and deploy models to production. Also, you will conduct in-depth data analysis and feature engineering to build robust ML models.
US, WA, Seattle
Amazon continues to invest heavily in building our world class advertising business. Our products are strategically important to our Retail and Marketplace businesses, driving long term growth. We deliver billions of ad impressions and millions of clicks daily, breaking fresh ground to create world-class products. We are highly motivated, collaborative and fun-loving with an entrepreneurial spirit and strong bias for action. With a broad mandate to experiment and innovate, we are growing at an unprecedented rate with a seemingly endless range of new opportunities. The Sponsored Products Monetization team is broadly responsible for pricing of ads on Amazon search pages, balancing short-term and long-term ad revenue growth to drive sustainable marketplace health. As a Senior Applied Scientist on our team, you will be responsible for defining the science and technical strategy for one of our most impactful marketplace controls, creating lasting value for Amazon and our advertising customers. You will help to identify unique opportunities to create customized and delightful shopping experience for our growing marketplaces worldwide. Your job will be identify big opportunities for the team that can help to grow Sponsored Products business working with retail partner teams, Product managers, Software engineers and PMs. You will have opportunity to design, run and analyze A/B experiments to improve the experience of millions of Amazon shoppers while driving quantifiable revenue impact. More importantly, you will have the opportunity to broaden your technical skills in an environment that thrives on creativity, experimentation, and product innovation. Key job responsibilities - Lead science, tech and business strategy and roadmap for Sponsored Products Monetization - Drive alignment across multiple organizations for science, engineering and product strategy to achieve business goals - Lead and mentor scientists and engineers across teams to develop, test, launch and improve of science models designed to optimize the shopper experience and deliver long term value for Amazon and advertisers - Develop state of the art experimental approaches and ML models - Drive end-to-end Machine Learning projects that have a high degree of ambiguity, scale, complexity - Establish scalable, efficient, automated processes for large-scale data analysis, machine-learning model development, model validation and serving - Research new and innovative machine learning approaches - Recruit Scientists to the team and provide mentorship