Aligning Explanations with Human Communication

Machine learning explainability aims to make the decision-making process of black-box models more transparent by finding the most important input features for a given prediction task. Recent works have proposed composing explanations from semantic concepts (e.g., colors, patterns, shapes) that are inherently interpretable to the user of a model. However, these methods generally ignore the communicative context of explanation-the ability of the user to understand the prediction of the model from the explanation. For example, while a medical doctor might understand an explanation in terms of clinical markers, a patient may need a more accessible explanation to make sense of the same diagnosis. In this paper, we address this gap with listener-adaptive explanations. We propose an iterative procedure grounded in principles of pragmatic reasoning and the rational speech act to generate explanations that maximize communicative utility. Our procedure only needs access to pairwise preferences between candidate explanations, relevant in real-world scenarios where a listener model may not be available. We evaluate our method in image classification tasks, demonstrating improved alignment between explanations and listener preferences across three datasets. Furthermore, we perform a user study that demonstrates our explanations increase communicative utility.

Multiaccuracy and Multicalibration via Proxy Groups

As the use of predictive machine learning algorithms increases in high-stakes decision-making, it is imperative that these algorithms are fair across sensitive groups. Unfortunately, measuring and enforcing fairness in real-world applications can be challenging due to missing or incomplete sensitive group data. Proxy-sensitive attributes have been proposed as a practical and effective solution in these settings, but only for parity-based fairness notions. Knowing how to evaluate and control for fairness with missing sensitive group data for newer and more flexible frameworks, such as multiaccuracy and multicalibration, remains unexplored. In this work, we address this gap by demonstrating that in the absence of sensitive group data, proxy-sensitive attributes can provably be used to derive actionable upper bounds on the true multiaccuracy and multicalibration, providing insights into a model's potential worst-case fairness violations. Additionally, we show that adjusting models to satisfy multiaccuracy and multicalibration across proxy-sensitive attributes can significantly mitigate these violations for the true, but unknown, sensitive groups. Through several experiments on real-world datasets, we illustrate that approximate multiaccuracy and multicalibration can be achieved even when sensitive group information is incomplete or unavailable.

Expanding the Horizon: Enabling Hybrid Quantum Transfer Learning for Long-Tailed Chest X-Ray Classification

Quantum machine learning (QML) has the potential for improving the multi-label classification of rare, albeit critical, diseases in large-scale chest x-ray (CXR) datasets due to theoretical quantum advantages over classical machine learning (CML) in sample efficiency and generalizability. While prior literature has explored QML with CXRs, it has focused on binary classification tasks with small datasets due to limited access to quantum hardware and computationally expensive simulations. To that end, we implemented a Jax-based framework that enables the simulation of medium-sized qubit architectures with significant improvements in wall-clock time over current software offerings. We evaluated the performance of our Jax-based framework in terms of efficiency and performance for hybrid quantum transfer learning for long-tailed classification across 8, 14, and 19 disease labels using large-scale CXR datasets. The Jax-based framework resulted in up to a 58% and 95% speed-up compared to PyTorch and TensorFlow implementations, respectively. However, compared to CML, QML demonstrated slower convergence and an average AUROC of 0.70, 0.73, and 0.74 for the classification of 8, 14, and 19 CXR disease labels. In comparison, the CML models had an average AUROC of 0.77, 0.78, and 0.80 respectively. In conclusion, our work presents an accessible implementation of hybrid quantum transfer learning for long-tailed CXR classification with a computationally efficient Jax-based framework.

Improving Multi-Center Generalizability of GAN-Based Fat Suppression using Federated Learning

Generative Adversarial Network (GAN)-based synthesis of fat suppressed (FS) MRIs from non-FS proton density sequences has the potential to accelerate acquisition of knee MRIs. However, GANs trained on single-site data have poor generalizability to external data. We show that federated learning can improve multi-center generalizability of GANs for synthesizing FS MRIs, while facilitating privacy-preserving multi-institutional collaborations.

Anytime, Anywhere, Anyone: Investigating the Feasibility of Segment Anything Model for Crowd-Sourcing Medical Image Annotations

Curating annotations for medical image segmentation is a labor-intensive and time-consuming task that requires domain expertise, resulting in "narrowly" focused deep learning (DL) models with limited translational utility. Recently, foundation models like the Segment Anything Model (SAM) have revolutionized semantic segmentation with exceptional zero-shot generalizability across various domains, including medical imaging, and hold a lot of promise for streamlining the annotation process. However, SAM has yet to be evaluated in a crowd-sourced setting to curate annotations for training 3D DL segmentation models. In this work, we explore the potential of SAM for crowd-sourcing "sparse" annotations from non-experts to generate "dense" segmentation masks for training 3D nnU-Net models, a state-of-the-art DL segmentation model. Our results indicate that while SAM-generated annotations exhibit high mean Dice scores compared to ground-truth annotations, nnU-Net models trained on SAM-generated annotations perform significantly worse than nnU-Net models trained on ground-truth annotations ($p<0.001$, all).

Out-of-Distribution Detection and Data Drift Monitoring using Statistical Process Control

Background: Machine learning (ML) methods often fail with data that deviates from their training distribution. This is a significant concern for ML-enabled devices in clinical settings, where data drift may cause unexpected performance that jeopardizes patient safety. Method: We propose a ML-enabled Statistical Process Control (SPC) framework for out-of-distribution (OOD) detection and drift monitoring. SPC is advantageous as it visually and statistically highlights deviations from the expected distribution. To demonstrate the utility of the proposed framework for monitoring data drift in radiological images, we investigated different design choices, including methods for extracting feature representations, drift quantification, and SPC parameter selection. Results: We demonstrate the effectiveness of our framework for two tasks: 1) differentiating axial vs. non-axial computed tomography (CT) images and 2) separating chest x-ray (CXR) from other modalities. For both tasks, we achieved high accuracy in detecting OOD inputs, with 0.913 in CT and 0.995 in CXR, and sensitivity of 0.980 in CT and 0.984 in CXR. Our framework was also adept at monitoring data streams and identifying the time a drift occurred. In a simulation with 100 daily CXR cases, we detected a drift in OOD input percentage from 0-1% to 3-5% within two days, maintaining a low false-positive rate. Through additional experimental results, we demonstrate the framework's data-agnostic nature and independence from the underlying model's structure. Conclusion: We propose a framework for OOD detection and drift monitoring that is agnostic to data, modality, and model. The framework is customizable and can be adapted for specific applications.

Hidden in Plain Sight: Undetectable Adversarial Bias Attacks on Vulnerable Patient Populations

The proliferation of artificial intelligence (AI) in radiology has shed light on the risk of deep learning (DL) models exacerbating clinical biases towards vulnerable patient populations. While prior literature has focused on quantifying biases exhibited by trained DL models, demographically targeted adversarial bias attacks on DL models and its implication in the clinical environment remains an underexplored field of research in medical imaging. In this work, we demonstrate that demographically targeted label poisoning attacks can introduce adversarial underdiagnosis bias in DL models and degrade performance on underrepresented groups without impacting overall model performance. Moreover, our results across multiple performance metrics and demographic groups like sex, age, and their intersectional subgroups indicate that a group's vulnerability to undetectable adversarial bias attacks is directly correlated with its representation in the model's training data.

One Copy Is All You Need: Resource-Efficient Streaming of Medical Imaging Data at Scale

Large-scale medical imaging datasets have accelerated development of artificial intelligence tools for clinical decision support. However, the large size of these datasets is a bottleneck for users with limited storage and bandwidth. Many users may not even require such large datasets as AI models are often trained on lower resolution images. If users could directly download at their desired resolution, storage and bandwidth requirements would significantly decrease. However, it is impossible to anticipate every users' requirements and impractical to store the data at multiple resolutions. What if we could store images at a single resolution but send them at different ones? We propose MIST, an open-source framework to operationalize progressive resolution for streaming medical images at multiple resolutions from a single high-resolution copy. We demonstrate that MIST can dramatically reduce imaging infrastructure inefficiencies for hosting and streaming medical images by >90%, while maintaining diagnostic quality for deep learning applications.

High-Throughput AI Inference for Medical Image Classification and Segmentation using Intelligent Streaming

As the adoption of AI systems within the clinical setup grows, limitations in bandwidth could create communication bottlenecks when streaming imaging data, leading to delays in patient diagnosis and treatment. As such, healthcare providers and AI vendors will require greater computational infrastructure, therefore dramatically increasing costs. To that end, we developed intelligent streaming, a state-of-the-art framework to enable accelerated, cost-effective, bandwidth-optimized, and computationally efficient AI inference for clinical decision making at scale. For classification, intelligent streaming reduced the data transmission by 99.01% and decoding time by 98.58%, while increasing throughput by 27.43x. For segmentation, our framework reduced data transmission by 90.32%, decoding time by 90.26%, while increasing throughput by 4.20x. Our work demonstrates that intelligent streaming results in faster turnaround times, and reduced overall cost of data and transmission, without negatively impacting clinical decision making using AI systems.

Text2Cohort: Democratizing the NCI Imaging Data Commons with Natural Language Cohort Discovery

The Imaging Data Commons (IDC) is a cloud-based database that provides researchers with open access to cancer imaging data, with the goal of facilitating collaboration in medical imaging research. However, querying the IDC database for cohort discovery and access to imaging data has a significant learning curve for researchers due to its complex nature. We developed Text2Cohort, a large language model (LLM) based toolkit to facilitate user-friendly and intuitive natural language cohort discovery in the IDC. Text2Cohorts translates user input into IDC database queries using prompt engineering and autocorrection and returns the query's response to the user. Autocorrection resolves errors in queries by passing the errors back to the model for interpretation and correction. We evaluate Text2Cohort on 50 natural language user inputs ranging from information extraction to cohort discovery. The resulting queries and outputs were verified by two computer scientists to measure Text2Cohort's accuracy and F1 score. Text2Cohort successfully generated queries and their responses with an 88% accuracy and F1 score of 0.94. However, it failed to generate queries for 6/50 (12%) user inputs due to syntax and semantic errors. Our results indicate that Text2Cohort succeeded at generating queries with correct responses, but occasionally failed due to a lack of understanding of the data schema. Despite these shortcomings, Text2Cohort demonstrates the utility of LLMs to enable researchers to discover and curate cohorts using data hosted on IDC with high levels of accuracy using natural language in a more intuitive and user-friendly way.