Katie Matton

Katie Matton, Purvaja Balaji, Hamzeh Ghasemzadeh et al.

Phonotrauma refers to vocal fold tissue damage resulting from exposure to forces during voicing. It occurs on a continuum from mild to severe, and treatment options can vary based on severity. Assessment of severity involves a clinician's expert judgment, which is costly and can vary widely in reliability. In this work, we present the first method for automatically classifying phonotrauma severity from vocal fold images. To account for the ordinal nature of the labels, we adopt a widely used ordinal regression framework. To account for label uncertainty, we propose a novel modification to ordinal regression loss functions that enables them to operate on soft labels reflecting annotator rating distributions. Our proposed soft ordinal regression method achieves predictive performance approaching that of clinical experts, while producing well-calibrated uncertainty estimates. By providing an automated tool for phonotrauma severity assessment, our work can enable large-scale studies of phonotrauma, ultimately leading to improved clinical understanding and patient care.

Robert Lewis, Katie Matton, Rosalind W. Picard et al.

Self-supervised pre-training with contrastive learning is a powerful method for learning from sparsely labeled data. However, performance can drop considerably when there is a shift in the distribution of data from training to test time. We study this phenomenon in a setting in which the training data come from multiple domains, and the test data come from a domain not seen at training that is subject to significant covariate shift. We present a new method for contrastive learning that incorporates domain labels to increase the domain invariance of learned representations, leading to improved out-of-distribution generalization. Our method adjusts the temperature parameter in the InfoNCE loss -- which controls the relative weighting of negative pairs -- using the probability that a negative sample comes from the same domain as the anchor. This upweights pairs from more similar domains, encouraging the model to discriminate samples based on domain-invariant attributes. Through experiments on a variant of the MNIST dataset, we demonstrate that our method yields better out-of-distribution performance than domain generalization baselines. Furthermore, our method maintains strong in-distribution task performance, substantially outperforming baselines on this measure.

Katie Matton, Robert Osazuwa Ness, John Guttag et al.

Large language models (LLMs) are capable of generating plausible explanations of how they arrived at an answer to a question. However, these explanations can misrepresent the model's "reasoning" process, i.e., they can be unfaithful. This, in turn, can lead to over-trust and misuse. We introduce a new approach for measuring the faithfulness of LLM explanations. First, we provide a rigorous definition of faithfulness. Since LLM explanations mimic human explanations, they often reference high-level concepts in the input question that purportedly influenced the model. We define faithfulness in terms of the difference between the set of concepts that LLM explanations imply are influential and the set that truly are. Second, we present a novel method for estimating faithfulness that is based on: (1) using an auxiliary LLM to modify the values of concepts within model inputs to create realistic counterfactuals, and (2) using a Bayesian hierarchical model to quantify the causal effects of concepts at both the example- and dataset-level. Our experiments show that our method can be used to quantify and discover interpretable patterns of unfaithfulness. On a social bias task, we uncover cases where LLM explanations hide the influence of social bias. On a medical question answering task, we uncover cases where LLM explanations provide misleading claims about which pieces of evidence influenced the model's decisions.

Robert Osazuwa Ness, Katie Matton, Hayden Helm et al.

Large language models (LLM) have achieved impressive performance on medical question-answering benchmarks. However, high benchmark accuracy does not imply that the performance generalizes to real-world clinical settings. Medical question-answering benchmarks rely on assumptions consistent with quantifying LLM performance but that may not hold in the open world of the clinic. Yet LLMs learn broad knowledge that can help the LLM generalize to practical conditions regardless of unrealistic assumptions in celebrated benchmarks. We seek to quantify how well LLM medical question-answering benchmark performance generalizes when benchmark assumptions are violated. Specifically, we present an adversarial method that we call MedFuzz (for medical fuzzing). MedFuzz attempts to modify benchmark questions in ways aimed at confounding the LLM. We demonstrate the approach by targeting strong assumptions about patient characteristics presented in the MedQA benchmark. Successful "attacks" modify a benchmark item in ways that would be unlikely to fool a medical expert but nonetheless "trick" the LLM into changing from a correct to an incorrect answer. Further, we present a permutation test technique that can ensure a successful attack is statistically significant. We show how to use performance on a "MedFuzzed" benchmark, as well as individual successful attacks. The methods show promise at providing insights into the ability of an LLM to operate robustly in more realistic settings.