Katherine Heller
Katherine is a research scientist in Responsible AI at Google Research, and a member of Context in AI Research (CAIR) team. She works on Machine Learning (ML) research in Healthcare, Vision, Language, and Creativity, focusing on incorporating values for Transparency, Inclusivity, Fairness, and Robustness in our research. Prior to joining Google, she was Statistical Science faculty at Duke University, where she developed a sepsis detection system now in use at Duke University Hospital, and a nationally released iOS app which tries to complete the picture of peoples' Multiple Sclerosis course between clinic visits.
Katherine received a BS in CS and Applied Math from SUNY Stony Brook, an MS in CS from Columbia University, and a PhD in Machine Learning from the Gatsby Computational Neuroscience Unity at UCL. She was then a postdoc on an EPSRC fellowship in Engineering at the University of Cambridge, and an NSF postoc fellow in Brain and Cognitive Sciences at MIT.
Research Areas
Authored Publications
Sort By
A Toolbox for Surfacing Health Equity Harms and Biases in Large Language Models
Heather Cole-Lewis
Nenad Tomašev
Liam McCoy
Leo Anthony Celi
Alanna Walton
Akeiylah DeWitt
Philip Mansfield
Sushant Prakash
Joelle Barral
Ivor Horn
Karan Singhal
Arxiv (2024) (to appear)
Preview abstract
Large language models (LLMs) hold immense promise to serve complex health information needs but also have the potential to introduce harm and exacerbate health disparities. Reliably evaluating equity-related model failures is a critical step toward developing systems that promote health equity. In this work, we present resources and methodologies for surfacing biases with potential to precipitate equity-related harms in long-form, LLM-generated answers to medical questions and then conduct an empirical case study with Med-PaLM 2, resulting in the largest human evaluation study in this area to date. Our contributions include a multifactorial framework for human assessment of LLM-generated answers for biases, and EquityMedQA, a collection of seven newly-released datasets comprising both manually-curated and LLM-generated questions enriched for adversarial queries. Both our human assessment framework and dataset design process are grounded in an iterative participatory approach and review of possible biases in Med-PaLM 2 answers to adversarial queries. Through our empirical study, we find that the use of a collection of datasets curated through a variety of methodologies, coupled with a thorough evaluation protocol that leverages multiple assessment rubric designs and diverse rater groups, surfaces biases that may be missed via narrower evaluation approaches. Our experience underscores the importance of using diverse assessment methodologies and involving raters of varying backgrounds and expertise. We emphasize that while our framework can identify specific forms of bias, it is not sufficient to holistically assess whether the deployment of an AI system promotes equitable health outcomes. We hope the broader community leverages and builds on these tools and methods towards realizing a shared goal of LLMs that promote accessible and equitable healthcare for all.
View details
Batch Calibration: Rethinking Calibration For In-Context Learning And Prompt Engineering
Lev Proleev
International Conference on Learning Representations (ICLR) (2024)
Preview abstract
Prompting and in-context learning (ICL) have become efficient learning paradigms for large language models (LLMs). However, LLMs suffer from prompt brittleness and various bias factors in the prompt, including but not limited to the formatting, the choice verbalizers, and the ICL examples. To address this problem that results in unexpected performance degradation, calibration methods have been developed to mitigate the effects of these biases while recovering LLM performance. In this work, we first conduct a systematic analysis of the existing calibration methods, where we both provide a unified view and reveal the failure cases. Inspired by these analyses, we propose Batch Calibration (BC), a simple yet intuitive method that controls the contextual bias from the batched input, unifies various prior approaches, and effectively addresses the aforementioned issues. BC is zero-shot, inference-only, and incurs negligible additional costs. In the few-shot setup, we further extend BC to allow it to learn the contextual bias from labeled data. We validate the effectiveness of BC with PaLM 2-(S, M, L) and CLIP models and demonstrate state-of-the-art performance over previous calibration baselines across more than 10 natural language understanding and image classification tasks.
View details
Diagnosing failures of fairness transfer across distribution shift in real-world medical settings
Sanmi Koyejo
Eva Schnider
Krista Opsahl-Ong
Alex Brown
Christina Chen
Silvia Chiappa
Proceedings of Neural Information Processing Systems 2022 (2022)
Preview abstract
Diagnosing and mitigating changes in model fairness under distribution shift is an important component of the safe deployment of machine learning in healthcare settings. Importantly, the success of any mitigation strategy strongly depends on the structure of the shift. Despite this, there has been little discussion of how to empirically assess the structure of a distribution shift that one is encountering in practice. In this work, we adopt a causal framing to motivate conditional independence tests as a key tool for characterizing distribution shifts. Using our approach in two medical applications, we show that this knowledge can help diagnose failures of fairness transfer, including cases where real-world shifts are more complex than is often assumed in the literature. Based on these results, we discuss potential remedies at each step of the machine learning pipeline.
View details
Healthsheet: development of a transparency artifact for health datasets
Lauren Wilcox
Razvan Adrian Amironesei
Nyalleng Moorosi
ACM FAccT Conference 2022, ACM (2022)
Preview abstract
Machine learning (ML) approaches have demonstrated promising results in a wide range of healthcare applications. Data plays a crucial role in developing ML-based healthcare systems that directly affect people’s lives. Many of the ethical issues surrounding the use of ML in healthcare stem from structural inequalities underlying the way we collect, use, and handle data. Developing guidelines to improve documentation practices regarding the creation, use, and maintenance of ML healthcare datasets is therefore of critical importance. In this work, we introduce Healthsheet, a contextualized adaptation of the original datasheet questionnaire for health-specific applications. Through a series of semi-structured interviews, we adapt the datasheets for healthcare data documentation. As part of the Healthsheet development process and to understand the obstacles researchers face in creating datasheets, we worked with three publicly-available healthcare datasets as our case studies, each with different types of structured data: Electronic health Records (EHR), clinical trial study data, and smartphone-based performance outcome measures. Our findings from the interviewee study and case studies show 1) that datasheets should be contextualized for healthcare, 2) that despite incentives to adopt accountability practices such as datasheets, there is a lack of consistency in the broader use of these practices 3) how the ML for health community views datasheets and particularly Healthsheets as diagnostic tool to surface the limitations and strength of datasets and 4) the relative importance of different fields in the datasheet to healthcare concerns.
View details
Preview abstract
In order to build trust that a machine learned model is appropriate and responsible within a systems context involving technical and
human components, a broad range of factors typically need to be considered. However in practice model evaluations frequently focus
on only a narrow range of expected predictive behaviours. This paper examines the critical evaluation gap between the idealized
breadth of concerns and the observed narrow focus of actual evaluations. In doing so, we demonstrate which values are centered—and
which are marginalized—within the machine learning community. Through an empirical study of machine learning papers from
recent high profile conferences, we demonstrate the discipline’s general focus on a small set of evaluation methods. By considering the
mathematical formulations of evaluation metrics and the test datasets over which they are calculated, we draw attention to which
properties of models are centered in the field. This analysis also reveals an important gap: the properties of models which are frequently
neglected or sidelined during evaluation. By studying the structure of this gap, we demonstrate the machine learning discipline’s
implicit assumption of a range of commitments which have normative impacts; these include commitments to consequentialism,
abstractability from context, the quantifiability of impacts, the irrelevance of non-predictive features, and the equivalence of different
failure modes. Shedding light on these assumptions and commitments enables us to question their appropriateness for different ML
system contexts, and points the way towards more diverse and contextualized evaluation methodologies which can be used to more
robustly examine the trustworthiness of ML models.
View details
Deep Cox Mixtures for Survival Regression
Proceedings of the 6th Machine Learning for Healthcare Conference, PMLR (2021), pp. 674-708
Preview abstract
Survival analysis is a challenging variation of regression modeling because of the presence of censoring, where the outcome measurement is only partially known, due to, for example, loss to follow up. Such problems come up frequently in medical applications, making survival analysis a key endeavor in biostatistics and machine learning for healthcare, with Cox regression models being amongst the most commonly employed models. We describe a new approach for survival analysis regression models, based on learning mixtures of Cox regressions to model individual survival distributions. We propose an approximation to the Expectation Maximization algorithm for this model that does hard assignments to mixture groups to make optimization efficient. In each group assignment, we fit the hazard ratios within each group using deep neural networks, and the baseline hazard for each mixture component non-parametrically.
We perform experiments on multiple real world datasets, and look at the mortality rates of patients across ethnicity and gender. We emphasize the importance of calibration in healthcare settings and demonstrate that our approach outperforms classical and modern survival analysis baselines, both in terms of discriminative performance and calibration, with large gains in performance on the minority demographics.
View details
Three Directions for the Design of Human-Centered Machine Translation
Samantha Robertson
Wesley Deng
Timnit Gebru
Margaret Mitchell
Samy Bengio
Niloufar Salehi
(2021)
Preview abstract
As people all over the world adopt machine translation (MT) to communicate across languages, there is increased need for affordances that aid users in understanding when to rely on automated translations. Identifying the information and interactions that will most help users meet their translation needs is an open area of research at the intersection of Human-Computer Interaction (HCI) and Natural Language Processing (NLP). This paper advances work in this area by drawing on a survey of users' strategies in assessing translations. We identify three directions for the design of translation systems that support more reliable and effective use of machine translation: helping users craft good inputs, helping users understand translations, and expanding interactivity and adaptivity. We describe how these can be introduced in current MT systems and highlight open questions for HCI and NLP research.
View details
Analyzing the Role of Model Uncertainty for Electronic Health Records
Edward Choi
Jeremy Nixon
Ghassen Jerfel
ACM Conference on Health, Inference, and Learning (ACM CHIL) (2020)
Preview abstract
In medicine, both ethical and monetary costs of incorrect predictions can be significant, and the complexity of the problems often necessitates increasingly complex models. Recent work has shown that changing just the random seed is enough for otherwise well-tuned deep neural networks to vary in their individual predicted probabilities. In light of this, we investigate the role of model uncertainty methods in the medical domain. Using RNN ensembles and various Bayesian RNNs, we show that population-level metrics, such as AUC-PR, AUC-ROC, log-likelihood, and calibration error, do not capture model uncertainty. Meanwhile, the presence of significant variability in patient-specific predictions and optimal decisions motivates the need for capturing model uncertainty. Understanding the uncertainty for individual patients is an area with clear clinical impact, such as determining when a model decision is likely to be brittle. We further show that RNNs with only Bayesian embeddings can be a more efficient way to capture model uncertainty compared to ensembles, and we analyze how model uncertainty is impacted across individual input features and patient subgroups.
View details
Efficient and Scalable Bayesian Neural Nets with Rank-1 Factors
Ghassen Jerfel
Yeming Wen
Yian Ma
International Conference on Machine Learning (ICML) (2020)
Preview abstract
Bayesian neural networks (BNNs) demonstrate promising success in improving the robustness and uncertainty quantification of modern neural networks. However, they generally struggle with underfitting at scale and parameter efficiency. On the other hand, deep ensembles have emerged as an alternative for uncertainty quantification that, while outperforming BNNs on certain problems, also suffers from efficiency issues. It remains unclear how to combine the strengths of these two approaches and remediate their common issues. To tackle this challenge, we propose a rank-1 parameterization of BNNs, where each weight matrix involves only a distribution on a rank-1 subspace. We also revisit the use of mixture approximate posteriors to capture multiple modes where unlike typical mixtures, this approach admits a significantly smaller memory increase (e.g., only a 0.4% increase for a ResNet-50 mixture of size 10). We perform a systematic empirical study on the choices of prior, variational posterior, and methods to improve training. For ResNet-50 on ImageNet and Wide ResNet 28-10 on CIFAR-10/100, rank-1 BNNs outperform baselines across log-likelihood, accuracy, and calibration on the test set and out-of-distribution variants.
View details
Underspecification Presents Challenges for Credibility in Modern Machine Learning
Dan Moldovan
Ben Adlam
Babak Alipanahi
Alex Beutel
Christina Chen
Jon Deaton
Shaobo Hou
Neil Houlsby
Ghassen Jerfel
Yian Ma
Akinori Mitani
Andrea Montanari
Christopher Nielsen
Thomas Osborne
Rajiv Raman
Kim Ramasamy
Martin Gamunu Seneviratne
Shannon Sequeira
Harini Suresh
Victor Veitch
Journal of Machine Learning Research (2020)
Preview abstract
ML models often exhibit unexpectedly poor behavior when they are deployed in real-world domains. We identify underspecification as a key reason for these failures. An ML pipeline is underspecified when it can return many predictors with equivalently strong held-out performance in the training domain. Underspecification is common in modern ML pipelines, such as those based on deep learning. Predictors returned by underspecified pipelines are often treated as equivalent based on their training domain performance, but we show here that such predictors can behave very differently in deployment domains. This ambiguity can lead to instability and poor model behavior in practice, and is a distinct failure mode from previously identified issues arising from structural mismatch between training and deployment domains. We show that this problem appears in a wide variety of practical ML pipelines, using examples from computer vision, medical imaging, natural language processing, clinical risk prediction based on electronic health records, and medical genomics. Our results show the need to explicitly account for underspecification in modeling pipelines that are intended for real-world deployment in any domain.
View details