Trevor Cohen

Yue Guo, Wei Qiu, Gondy Leroy et al. · uw

Recent lay language generation systems have used Transformer models trained on a parallel corpus to increase health information accessibility. However, the applicability of these models is constrained by the limited size and topical breadth of available corpora. We introduce CELLS, the largest (63k pairs) and broadest-ranging (12 journals) parallel corpus for lay language generation. The abstract and the corresponding lay language summary are written by domain experts, assuring the quality of our dataset. Furthermore, qualitative evaluation of expert-authored plain language summaries has revealed background explanation as a key strategy to increase accessibility. Such explanation is challenging for neural models to generate because it goes beyond simplification by adding content absent from the source. We derive two specialized paired corpora from CELLS to address key challenges in lay language generation: generating background explanations and simplifying the original abstract. We adopt retrieval-augmented models as an intuitive fit for the task of background explanation generation, and show improvements in summary quality and simplicity while maintaining factual correctness. Taken together, this work presents the first comprehensive study of background explanation for lay language generation, paving the path for disseminating scientific knowledge to a broader audience. CELLS is publicly available at: https://github.com/LinguisticAnomalies/pls_retrieval.

Changye Li, Weizhe Xu, Trevor Cohen et al. · uw

The evidence is growing that machine and deep learning methods can learn the subtle differences between the language produced by people with various forms of cognitive impairment such as dementia and cognitively healthy individuals. Valuable public data repositories such as TalkBank have made it possible for researchers in the computational community to join forces and learn from each other to make significant advances in this area. However, due to variability in approaches and data selection strategies used by various researchers, results obtained by different groups have been difficult to compare directly. In this paper, we present TRESTLE (\textbf{T}oolkit for \textbf{R}eproducible \textbf{E}xecution of \textbf{S}peech \textbf{T}ext and \textbf{L}anguage \textbf{E}xperiments), an open source platform that focuses on two datasets from the TalkBank repository with dementia detection as an illustrative domain. Successfully deployed in the hackallenge (Hackathon/Challenge) of the International Workshop on Health Intelligence at AAAI 2022, TRESTLE provides a precise digital blueprint of the data pre-processing and selection strategies that can be reused via TRESTLE by other researchers seeking comparable results with their peers and current state-of-the-art (SOTA) approaches.

Changye Li, David Knopman, Weizhe Xu et al. · uw

Deep learning (DL) techniques involving fine-tuning large numbers of model parameters have delivered impressive performance on the task of discriminating between language produced by cognitively healthy individuals, and those with Alzheimer's disease (AD). However, questions remain about their ability to generalize beyond the small reference sets that are publicly available for research. As an alternative to fitting model parameters directly, we propose a novel method by which a Transformer DL model (GPT-2) pre-trained on general English text is paired with an artificially degraded version of itself (GPT-D), to compute the ratio between these two models' \textit{perplexities} on language from cognitively healthy and impaired individuals. This technique approaches state-of-the-art performance on text data from a widely used "Cookie Theft" picture description task, and unlike established alternatives also generalizes well to spontaneous conversations. Furthermore, GPT-D generates text with characteristics known to be associated with AD, demonstrating the induction of dementia-related linguistic anomalies. Our study is a step toward better understanding of the relationships between the inner workings of generative neural language models, the language that they produce, and the deleterious effects of dementia on human speech and language characteristics.

Yue Guo, Joseph Chee Chang, Maria Antoniak et al. · allen-ai, uw

Scientific jargon can impede researchers when they read materials from other domains. Current methods of jargon identification mainly use corpus-level familiarity indicators (e.g., Simple Wikipedia represents plain language). However, researchers' familiarity of a term can vary greatly based on their own background. We collect a dataset of over 10K term familiarity annotations from 11 computer science researchers for terms drawn from 100 paper abstracts. Analysis of this data reveals that jargon familiarity and information needs vary widely across annotators, even within the same sub-domain (e.g., NLP). We investigate features representing individual, sub-domain, and domain knowledge to predict individual jargon familiarity. We compare supervised and prompt-based approaches, finding that prompt-based methods including personal publications yields the highest accuracy, though zero-shot prompting provides a strong baseline. This research offers insight into features and methods to integrate personal data into scientific jargon identification.

Changye Li, Trevor Cohen, Serguei Pakhomov · uw

Linguistic anomalies detectable in spontaneous speech have shown promise for various clinical applications including screening for dementia and other forms of cognitive impairment. The feasibility of deploying automated tools that can classify language samples obtained from speech in large-scale clinical settings depends on the ability to capture and automatically transcribe the speech for subsequent analysis. However, the impressive performance of self-supervised learning (SSL) automatic speech recognition (ASR) models with curated speech data is not apparent with challenging speech samples from clinical settings. One of the key questions for successfully applying ASR models for clinical applications is whether imperfect transcripts they generate provide sufficient information for downstream tasks to operate at an acceptable level of accuracy. In this study, we examine the relationship between the errors produced by several deep learning ASR systems and their impact on the downstream task of dementia classification. One of our key findings is that, paradoxically, ASR systems with relatively high error rates can produce transcripts that result in better downstream classification accuracy than classification based on verbatim transcripts.

Daniil Filienko, Yinzhou Wang, Caroline El Jazmi et al. · uw

While Large Language Models (LLMs) are being quickly adapted to many domains, including healthcare, their strengths and pitfalls remain under-explored. In our study, we examine the effects of prompt engineering to guide Large Language Models (LLMs) in delivering parts of a Problem-Solving Therapy (PST) session via text, particularly during the symptom identification and assessment phase for personalized goal setting. We present evaluation results of the models' performances by automatic metrics and experienced medical professionals. We demonstrate that the models' capability to deliver protocolized therapy can be improved with the proper use of prompt engineering methods, albeit with limitations. To our knowledge, this study is among the first to assess the effects of various prompting techniques in enhancing a generalist model's ability to deliver psychotherapy, focusing on overall quality, consistency, and empathy. Exploring LLMs' potential in delivering psychotherapy holds promise with the current shortage of mental health professionals amid significant needs, enhancing the potential utility of AI-based and AI-enhanced care services.

Xiruo Ding, Zhecheng Sheng, Meliha Yetişgen et al.

Natural Language Processing (NLP) methods have been broadly applied to clinical tasks. Machine learning and deep learning approaches have been used to improve the performance of clinical NLP. However, these approaches require sufficiently large datasets for training, and trained models have been shown to transfer poorly across sites. These issues have led to the promotion of data collection and integration across different institutions for accurate and portable models. However, this can introduce a form of bias called confounding by provenance. When source-specific data distributions differ at deployment, this may harm model performance. To address this issue, we evaluate the utility of backdoor adjustment for text classification in a multi-site dataset of clinical notes annotated for mentions of substance abuse. Using an evaluation framework devised to measure robustness to distributional shifts, we assess the utility of backdoor adjustment. Our results indicate that backdoor adjustment can effectively mitigate for confounding shift.

Feng Chen, Justin Tauscher, Changye Li et al.

Speech monologues recorded in naturalistic settings provide opportunities to characterize mental illness phenomenology and detect symptom exacerbation. Large language models (LLMs) offer new possibilities for automating this process, as they require annotated data primarily for evaluation rather than training. In this paper, we present a novel automated, multi-agent LLM pipeline for the fine-grained, multi-label extraction of language suggestive of delusional beliefs, associated affective responses, and behavioral responses from transcripts of naturalistic audio diaries collected from people with moderate persecutory ideation. Evaluating an ensemble of three foundation models, we demonstrate that detailed diagnostic prompt instructions successfully reduce false positives for delusional theme classification, but also constrain the interpretation of affective or behavioral responses. Furthermore, comparing multi-agent adjudication frameworks shows that complex conversational debate between agents diminishes accuracy on clinically ambiguous text by inducing premature consensus. Instead, majority voting establishes robust performance (Micro F1 of 0.872 and 0.779 for delusion detection and classification respectively). This work provides a validated and scalable pipeline for the automated detection and characterization of content suggesting delusional beliefs in naturalistic speech.

Changye Li, Trevor Cohen, Serguei Pakhomov · uw

Automatic speech recognition (ASR) models trained on large amounts of audio data are now widely used to convert speech to written text in a variety of applications from video captioning to automated assistants used in healthcare and other domains. As such, it is important that ASR models and their use is fair and equitable. Prior work examining the performance of commercial ASR systems on the Corpus of Regional African American Language (CORAAL) demonstrated significantly worse ASR performance on African American English (AAE). The current study seeks to understand the factors underlying this disparity by examining the performance of the current state-of-the-art neural network based ASR system (Whisper, OpenAI) on the CORAAL dataset. Two key findings have been identified as a result of the current study. The first confirms prior findings of significant dialectal variation even across neighboring communities, and worse ASR performance on AAE that can be improved to some extent with fine-tuning of ASR models. The second is a novel finding not discussed in prior work on CORAAL: differences in audio recording practices within the dataset have a significant impact on ASR accuracy resulting in a ``confounding by provenance'' effect in which both language use and recording quality differ by study location. These findings highlight the need for further systematic investigation to disentangle the effects of recording quality and inherent linguistic diversity when examining the fairness and bias present in neural ASR models, as any bias in ASR accuracy may have negative downstream effects on disparities in various domains of life in which ASR technology is used.

Xinyang Ren, Hannah A Burkhardt, Patricia A Areán et al.

Prior work has shown that analyzing the use of first-person singular pronouns can provide insight into individuals' mental status, especially depression symptom severity. These findings were generated by counting frequencies of first-person singular pronouns in text data. However, counting doesn't capture how these pronouns are used. Recent advances in neural language modeling have leveraged methods generating contextual embeddings. In this study, we sought to utilize the embeddings of first-person pronouns obtained from contextualized language representation models to capture ways these pronouns are used, to analyze mental status. De-identified text messages sent during online psychotherapy with weekly assessment of depression severity were used for evaluation. Results indicate the advantage of contextualized first-person pronoun embeddings over standard classification token embeddings and frequency-based pronoun analysis results in predicting depression symptom severity. This suggests contextual representations of first-person pronouns can enhance the predictive utility of language used by people with depression symptoms.

Feng Chen, Manas Satish Bedmutha, Ray-Yuan Chung et al.

Implicit bias can impede patient-provider interactions and lead to inequities in care. Raising awareness is key to reducing such bias, but its manifestations in the social dynamics of patient-provider communication are difficult to detect. In this study, we used automated speech recognition (ASR) and natural language processing (NLP) to identify social signals in patient-provider interactions. We built an automated pipeline to predict social signals from audio recordings of 782 primary care visits that achieved 90.1% average accuracy across codes, and exhibited fairness in its predictions for white and non-white patients. Applying this pipeline, we identified statistically significant differences in provider communication behavior toward white versus non-white patients. In particular, providers expressed more patient-centered behaviors towards white patients including more warmth, engagement, and attentiveness. Our study underscores the potential of automated tools in identifying subtle communication signals that may be linked with bias and impact healthcare quality and equity.

Feng Chen, Manas Bedmutha, Janice Sabin et al.

Depression is underdiagnosed in primary care, yet timely identification remains critical. Recorded clinical encounters, increasingly common with digital scribing technologies, present an opportunity to detect depression from naturalistic dialogue. We investigated automated depression detection from 1,108 audio-recorded primary care encounters in the Establishing Focus study, with depression defined by PHQ-9 (n=253 depressed, n=855 non-depressed). We compared three supervised approaches, Sentence-BERT + Logistic Regression (LR), LIWC+LR and ModernBERT, against a zero-shot GPT-OSS. GPT-OSS achieved the strongest performance (AUPRC=0.510, AUROC=0.774), with LIWC+LR competitive among supervised models (AUPRC=0.500, AUROC=0.742). Combined dyadic transcripts outperformed single-speaker configurations, with providers linguistically mirroring patients in depression encounters, an additive signal not captured by either speaker alone. Meaningful detection is achievable from the first 128 patient tokens (AUPRC=0.356, AUROC=0.675), supporting in-the-moment clinical decision support. These findings argue for passively collected clinical audio as a low-burden complement to existing screening workflows.

Yongsen Tan, Zhecheng Sheng, Xiruo Ding et al.

Despite the burgeoning body of work on distribution shifts, provenance shift-where the relationship between data source and label changes at deployment-remains poorly understood and under-addressed. In this paper, we establish a formal connection between provenance shift, counterfactual invariance, and invariant learning to derive a learning objective for robustness. We then introduce \textsc{DeconDTN-Toolkit}, a specialized evaluation and remediation suite designed to simulate provenance shifts of varying degrees while maintaining the training protocol and the infrastructure of existing benchmarks. We reveal the vulnerability of Empirical Risk Minimization under provenance shift, introduce a robust out-of-distribution performance indicator, and conduct a comprehensive evaluation on existing algorithms. Our work provides both the theoretical grounding and the practical tools necessary to characterize the problem of confounding by provenance, and implementations of methods to mitigate it.

Yue Guo, Tal August, Gondy Leroy et al.

While there has been significant development of models for Plain Language Summarization (PLS), evaluation remains a challenge. PLS lacks a dedicated assessment metric, and the suitability of text generation evaluation metrics is unclear due to the unique transformations involved (e.g., adding background explanations, removing jargon). To address these questions, our study introduces a granular meta-evaluation testbed, APPLS, designed to evaluate metrics for PLS. We identify four PLS criteria from previous work -- informativeness, simplification, coherence, and faithfulness -- and define a set of perturbations corresponding to these criteria that sensitive metrics should be able to detect. We apply these perturbations to extractive hypotheses for two PLS datasets to form our testbed. Using APPLS, we assess performance of 14 metrics, including automated scores, lexical features, and LLM prompt-based evaluations. Our analysis reveals that while some current metrics show sensitivity to specific criteria, no single method captures all four criteria simultaneously. We therefore recommend a suite of automated metrics be used to capture PLS quality along all relevant criteria. This work contributes the first meta-evaluation testbed for PLS and a comprehensive evaluation of existing metrics. APPLS and our evaluation code is available at https://github.com/LinguisticAnomalies/APPLS.

Zhecheng Sheng, Yongsen Tan, Xiruo Ding et al.

In classification tasks, models may rely on confounding variables to achieve strong in-distribution performance, capturing spurious features that fail under distribution shift. This shortcut behavior leads to substantial degradation in out-of-distribution settings. Task arithmetic offers a potential solution by removing unwanted signals via subtraction of secondary model updates, but it typically requires full fine-tuning, which is computationally expensive. Prompt tuning provides a parameter-efficient alternative by adapting models through a small set of trainable virtual tokens. Task arithmetic on the resulting prompts presents an appealing alternative to operations on entire models, but the extent to which this approach can limit reliance on spurious features remains to be established. In this work, we study whether composing soft prompts through task arithmetic improves robustness to confounding shifts. We propose Hybrid Prompt Arithmetic (HyPA), which combines task prompts with linearized confounder prompts to counteract spurious correlations. Across multiple benchmarks, HyPA consistently improves the robustness-performance trade-off relative to prompt-arithmetic baselines under distribution shift. We further analyze how HyPA affects hidden representations and find evidence consistent with it mitigating confounding either by reducing the influence of confounder signals on predictions or by suppressing them in the representation. These results establish HyPA as a parameter-efficient and promising approach for improving robustness under confounding shifts in the evaluated setting.

Changye Li, Weizhe Xu, Trevor Cohen et al. · uw

\textbf{Objectives}: We aimed to investigate how errors from automatic speech recognition (ASR) systems affect dementia classification accuracy, specifically in the ``Cookie Theft'' picture description task. We aimed to assess whether imperfect ASR-generated transcripts could provide valuable information for distinguishing between language samples from cognitively healthy individuals and those with Alzheimer's disease (AD). \textbf{Methods}: We conducted experiments using various ASR models, refining their transcripts with post-editing techniques. Both these imperfect ASR transcripts and manually transcribed ones were used as inputs for the downstream dementia classification. We conducted comprehensive error analysis to compare model performance and assess ASR-generated transcript effectiveness in dementia classification. \textbf{Results}: Imperfect ASR-generated transcripts surprisingly outperformed manual transcription for distinguishing between individuals with AD and those without in the ``Cookie Theft'' task. These ASR-based models surpassed the previous state-of-the-art approach, indicating that ASR errors may contain valuable cues related to dementia. The synergy between ASR and classification models improved overall accuracy in dementia classification. \textbf{Conclusion}: Imperfect ASR transcripts effectively capture linguistic anomalies linked to dementia, improving accuracy in classification tasks. This synergy between ASR and classification models underscores ASR's potential as a valuable tool in assessing cognitive impairment and related clinical applications.

William R. Kearns, Jessica Bertram, Myra Divina et al. · uw

Despite the high prevalence and burden of mental health conditions, there is a global shortage of mental health providers. Artificial Intelligence (AI) methods have been proposed as a way to address this shortage, by supporting providers with less extensive training as they deliver care. To this end, we developed the AI-Assisted Provider Platform (A2P2), a text-based virtual therapy interface that includes a response suggestion feature, which supports providers in delivering protocolized therapies empathetically. We studied providers with and without expertise in mental health treatment delivering a therapy session using the platform with (intervention) and without (control) AI-assistance features. Upon evaluation, the AI-assisted system significantly decreased response times by 29.34% (p=0.002), tripled empathic response accuracy (p=0.0001), and increased goal recommendation accuracy by 66.67% (p=0.001) across both user groups compared to the control. Both groups rated the system as having excellent usability.

Yue Guo, Jae Ho Sohn, Gondy Leroy et al. · uw

Plain language summaries (PLSs) are essential for facilitating effective communication between clinicians and patients by making complex medical information easier for laypeople to understand and act upon. Large language models (LLMs) have recently shown promise in automating PLS generation, but their effectiveness in supporting health information comprehension remains unclear. Prior evaluations have generally relied on automated scores that do not measure understandability directly, or subjective Likert-scale ratings from convenience samples with limited generalizability. To address these gaps, we conducted a large-scale crowdsourced evaluation of LLM-generated PLSs using Amazon Mechanical Turk with 150 participants. We assessed PLS quality through subjective Likert-scale ratings focusing on simplicity, informativeness, coherence, and faithfulness; and objective multiple-choice comprehension and recall measures of reader understanding. Additionally, we examined the alignment between 10 automated evaluation metrics and human judgments. Our findings indicate that while LLMs can generate PLSs that appear indistinguishable from human-written ones in subjective evaluations, human-written PLSs lead to significantly better comprehension. Furthermore, automated evaluation metrics fail to reflect human judgment, calling into question their suitability for evaluating PLSs. This is the first study to systematically evaluate LLM-generated PLSs based on both reader preferences and comprehension outcomes. Our findings highlight the need for evaluation frameworks that move beyond surface-level quality and for generation methods that explicitly optimize for layperson comprehension.

Manas Satish Bedmutha, Feng Chen, Andrea Hartzler et al.

Effective communication between providers and their patients influences health and care outcomes. The effectiveness of such conversations has been linked not only to the exchange of clinical information, but also to a range of interpersonal behaviors; commonly referred to as social signals, which are often conveyed through non-verbal cues and shape the quality of the patient-provider relationship. Recent advances in large language models (LLMs) have demonstrated an increasing ability to infer emotional and social behaviors even when analyzing only textual information. As automation increases also in clinical settings, such as for transcription of patient-provider conversations, there is growing potential for LLMs to automatically analyze and extract social behaviors from these interactions. To explore the foundational capabilities of LLMs in tracking social signals in clinical dialogue, we designed task-specific prompts and evaluated model performance across multiple architectures and prompting styles using a highly imbalanced, annotated dataset spanning 20 distinct social signals such as provider dominance, patient warmth, etc. We present the first system capable of tracking all these 20 coded signals, and uncover patterns in LLM behavior. Further analysis of model configurations and clinical context provides insights for enhancing LLM performance on social signal processing tasks in healthcare settings.

Feng Chen, Dror Ben-Zeev, Gillian Sparks et al.

Post-Traumatic Stress Disorder (PTSD) remains underdiagnosed in clinical settings, presenting opportunities for automated detection to identify patients. This study evaluates natural language processing approaches for detecting PTSD from clinical interview transcripts. We compared general and mental health-specific transformer models (BERT/RoBERTa), embedding-based methods (SentenceBERT/LLaMA), and large language model prompting strategies (zero-shot/few-shot/chain-of-thought) using the DAIC-WOZ dataset. Domain-specific end-to-end models significantly outperformed general models (Mental-RoBERTa AUPRC=0.675+/-0.084 vs. RoBERTa-base 0.599+/-0.145). SentenceBERT embeddings with neural networks achieved the highest overall performance (AUPRC=0.758+/-0.128). Few-shot prompting using DSM-5 criteria yielded competitive results with two examples (AUPRC=0.737). Performance varied significantly across symptom severity and comorbidity status with depression, with higher accuracy for severe PTSD cases and patients with comorbid depression. Our findings highlight the potential of domain-adapted embeddings and LLMs for scalable screening while underscoring the need for improved detection of nuanced presentations and offering insights for developing clinically viable AI tools for PTSD assessment.

Feng Chen, Weizhe Xu, Changye Li et al.

Formal thought disorder (FTD), a hallmark of schizophrenia spectrum disorders, manifests as incoherent speech and poses challenges for clinical assessment. Traditional clinical rating scales, though validated, are resource-intensive and lack scalability. Automated speech analysis with automatic speech recognition (ASR) allows for objective quantification of linguistic and temporal features of speech, offering scalable alternatives. The use of utterance timestamps in ASR captures pause dynamics, which are thought to reflect the cognitive processes underlying speech production. However, the utility of integrating these ASR-derived features for assessing FTD severity requires further evaluation. This study integrates pause features with semantic coherence metrics across three datasets: naturalistic self-recorded diaries (AVH, n = 140), structured picture descriptions (TOPSY, n = 72), and dream narratives (PsyCL, n = 43). We evaluated pause related features alongside established coherence measures, using support vector regression (SVR) to predict clinical FTD scores. Key findings demonstrate that pause features alone robustly predict the severity of FTD. Integrating pause features with semantic coherence metrics enhanced predictive performance compared to semantic-only models, with integration of independent models achieving correlations up to \r{ho} = 0.649 and AUC = 83.71% for severe cases detection (TOPSY, with best \r{ho} = 0.584 and AUC = 79.23% for semantic-only models). The performance gains from semantic and pause features integration held consistently across all contexts, though the nature of pause patterns was dataset-dependent. These findings suggest that frameworks combining temporal and semantic analyses provide a roadmap for refining the assessment of disorganized speech and advance automated speech analysis in psychosis.

Zhecheng Sheng, Xiruo Ding, Brian Hur et al. · uw

Deep transformer models have been used to detect linguistic anomalies in patient transcripts for early Alzheimer's disease (AD) screening. While pre-trained neural language models (LMs) fine-tuned on AD transcripts perform well, little research has explored the effects of the gender of the speakers represented by these transcripts. This work addresses gender confounding in dementia detection and proposes two methods: the $\textit{Extended Confounding Filter}$ and the $\textit{Dual Filter}$, which isolate and ablate weights associated with gender. We evaluate these methods on dementia datasets with first-person narratives from patients with cognitive impairment and healthy controls. Our results show transformer models tend to overfit to training data distributions. Disrupting gender-related weights results in a deconfounded dementia classifier, with the trade-off of slightly reduced dementia detection performance.

Changye Li, Zhecheng Sheng, Trevor Cohen et al. · uw

Alzheimer's Disease (AD) dementia is a progressive neurodegenerative disease that negatively impacts patients' cognitive ability. Previous studies have demonstrated that changes in naturalistic language samples can be useful for early screening of AD dementia. However, the nature of language deficits often requires test administrators to use various speech elicitation techniques during spontaneous language assessments to obtain enough propositional utterances from dementia patients. This could lead to the ``observer's effect'' on the downstream analysis that has not been fully investigated. Our study seeks to quantify the influence of test administrators on linguistic features in dementia assessment with two English corpora the ``Cookie Theft'' picture description datasets collected at different locations and test administrators show different levels of administrator involvement. Our results show that the level of test administrator involvement significantly impacts observed linguistic features in patient speech. These results suggest that many of significant linguistic features in the downstream classification task may be partially attributable to differences in the test administration practices rather than solely to participants' cognitive status. The variations in test administrator behavior can lead to systematic biases in linguistic data, potentially confounding research outcomes and clinical assessments. Our study suggests that there is a need for a more standardized test administration protocol in the development of responsible clinical speech analytics frameworks.

Changye Li, Weizhe Xu, Serguei Pakhomov et al. · uw

Disorganized thinking is a key diagnostic indicator of schizophrenia-spectrum disorders. Recently, clinical estimates of the severity of disorganized thinking have been shown to correlate with measures of how difficult speech transcripts would be for large language models (LLMs) to predict. However, LLMs' deployment challenges -- including privacy concerns, computational and financial costs, and lack of transparency of training data -- limit their clinical utility. We investigate whether smaller neural language models can serve as effective alternatives for detecting positive formal thought disorder, using the same sliding window based perplexity measurements that proved effective with larger models. Surprisingly, our results show that smaller models are more sensitive to linguistic differences associated with formal thought disorder than their larger counterparts. Detection capability declines beyond a certain model size and context length, challenging the common assumption of ``bigger is better'' for LLM-based applications. Our findings generalize across audio diaries and clinical interview speech samples from individuals with psychotic symptoms, suggesting a promising direction for developing efficient, cost-effective, and privacy-preserving screening tools that can be deployed in both clinical and naturalistic settings.

Changye Li, Zhecheng Sheng, Trevor Cohen et al.

As artificial neural networks grow in complexity, understanding their inner workings becomes increasingly challenging, which is particularly important in healthcare applications. The intrinsic evaluation metrics of autoregressive neural language models (NLMs), perplexity (PPL), can reflect how "surprised" an NLM model is at novel input. PPL has been widely used to understand the behavior of NLMs. Previous findings show that changes in PPL when masking attention layers in pre-trained transformer-based NLMs reflect linguistic anomalies associated with Alzheimer's disease dementia. Building upon this, we explore a novel bidirectional attention head ablation method that exhibits properties attributed to the concepts of cognitive and brain reserve in human brain studies, which postulate that people with more neurons in the brain and more efficient processing are more resilient to neurodegeneration. Our results show that larger GPT-2 models require a disproportionately larger share of attention heads to be masked/ablated to display degradation of similar magnitude to masking in smaller models. These results suggest that the attention mechanism in transformer models may present an analogue to the notions of cognitive and brain reserve and could potentially be used to model certain aspects of the progression of neurodegenerative disorders and aging.

Xiruo Ding, Zhecheng Sheng, Brian Hur et al.

Foundation models are a current focus of attention in both industry and academia. While they have shown their capabilities in a variety of tasks, in-depth research is required to determine their robustness to distribution shift when used as a basis for supervised machine learning. This is especially important in the context of clinical data, with particular limitations related to data accessibility, lack of pretraining materials, and limited availability of high-quality annotations. In this work, we examine the stability of models based on representations from foundation models under distribution shift. We focus on confounding by provenance, a form of distribution shift that emerges in the context of multi-institutional datasets when there are differences in source-specific language use and class distributions. Using a sampling strategy that synthetically induces varying degrees of distribution shift, we evaluate the extent to which representations from foundation models result in predictions that are inherently robust to confounding by provenance. Additionally, we examine the effectiveness of a straightforward confounding adjustment method inspired by Pearl's conception of backdoor adjustment. Results indicate that while foundation models do show some out-of-the-box robustness to confounding-by-provenance related distribution shifts, this can be considerably improved through adjustment. These findings suggest a need for deliberate adjustment of predictive models using representations from foundation models in the context of source-specific distributional differences.

Dominic Widdows, Kristen Howell, Trevor Cohen

Vector representations have become a central element in semantic language modelling, leading to mathematical overlaps with many fields including quantum theory. Compositionality is a core goal for such representations: given representations for 'wet' and 'fish', how should the concept 'wet fish' be represented? This position paper surveys this question from two points of view. The first considers the question of whether an explicit mathematical representation can be successful using only tools from within linear algebra, or whether other mathematical tools are needed. The second considers whether semantic vector composition should be explicitly described mathematically, or whether it can be a model-internal side-effect of training a neural network. A third and newer question is whether a compositional model can be implemented on a quantum computer. Given the fundamentally linear nature of quantum mechanics, we propose that these questions are related, and that this survey may help to highlight candidate operations for future quantum implementation.

Dominic Widdows, Kirsty Kitto, Trevor Cohen

In the decade since 2010, successes in artificial intelligence have been at the forefront of computer science and technology, and vector space models have solidified a position at the forefront of artificial intelligence. At the same time, quantum computers have become much more powerful, and announcements of major advances are frequently in the news. The mathematical techniques underlying both these areas have more in common than is sometimes realized. Vector spaces took a position at the axiomatic heart of quantum mechanics in the 1930s, and this adoption was a key motivation for the derivation of logic and probability from the linear geometry of vector spaces. Quantum interactions between particles are modelled using the tensor product, which is also used to express objects and operations in artificial neural networks. This paper describes some of these common mathematical areas, including examples of how they are used in artificial intelligence (AI), particularly in automated reasoning and natural language processing (NLP). Techniques discussed include vector spaces, scalar products, subspaces and implication, orthogonal projection and negation, dual vectors, density matrices, positive operators, and tensor products. Application areas include information retrieval, categorization and implication, modelling word-senses and disambiguation, inference in knowledge bases, and semantic composition. Some of these approaches can potentially be implemented on quantum hardware. Many of the practical steps in this implementation are in early stages, and some are already realized. Explaining some of the common mathematical tools can help researchers in both AI and quantum computing further exploit these overlaps, recognizing and exploring new directions along the way.

Yue Guo, Wei Qiu, Yizhong Wang et al.

Health literacy has emerged as a crucial factor in making appropriate health decisions and ensuring treatment outcomes. However, medical jargon and the complex structure of professional language in this domain make health information especially hard to interpret. Thus, there is an urgent unmet need for automated methods to enhance the accessibility of the biomedical literature to the general population. This problem can be framed as a type of translation problem between the language of healthcare professionals, and that of the general public. In this paper, we introduce the novel task of automated generation of lay language summaries of biomedical scientific reviews, and construct a dataset to support the development and evaluation of automated methods through which to enhance the accessibility of the biomedical literature. We conduct analyses of the various challenges in solving this task, including not only summarization of the key points but also explanation of background knowledge and simplification of professional language. We experiment with state-of-the-art summarization models as well as several data augmentation techniques, and evaluate their performance using both automated metrics and human assessment. Results indicate that automatically generated summaries produced using contemporary neural architectures can achieve promising quality and readability as compared with reference summaries developed for the lay public by experts (best ROUGE-L of 50.24 and Flesch-Kincaid readability score of 13.30). We also discuss the limitations of the current attempt, providing insights and directions for future work.

Xiruo Ding, Trevor Cohen

Adverse drug events (ADE) are prevalent and costly. Clinical trials are constrained in their ability to identify potential ADEs, motivating the development of spontaneous reporting systems for post-market surveillance. Statistical methods provide a convenient way to detect signals from these reports but have limitations in leveraging relationships between drugs and ADEs given their discrete count-based nature. A previously proposed method, aer2vec, generates distributed vector representations of ADE report entities that capture patterns of similarity but cannot utilize lexical knowledge. We address this limitation by retrofitting aer2vec drug embeddings to knowledge from RxNorm and developing a novel retrofitting variant using vector rescaling to preserve magnitude. When evaluated in the context of a pharmacovigilance signal detection task, aer2vec with retrofitting consistently outperforms disproportionality metrics when trained on minimally preprocessed data. Retrofitting with rescaling results in further improvements in the larger and more challenging of two pharmacovigilance reference sets used for evaluation.

Trevor Cohen, Serguei Pakhomov

In recent years there has been a burgeoning interest in the use of computational methods to distinguish between elicited speech samples produced by patients with dementia, and those from healthy controls. The difference between perplexity estimates from two neural language models (LMs) - one trained on transcripts of speech produced by healthy participants and the other trained on transcripts from patients with dementia - as a single feature for diagnostic classification of unseen transcripts has been shown to produce state-of-the-art performance. However, little is known about why this approach is effective, and on account of the lack of case/control matching in the most widely-used evaluation set of transcripts (DementiaBank), it is unclear if these approaches are truly diagnostic, or are sensitive to other variables. In this paper, we interrogate neural LMs trained on participants with and without dementia using synthetic narratives previously developed to simulate progressive semantic dementia by manipulating lexical frequency. We find that perplexity of neural LMs is strongly and differentially associated with lexical frequency, and that a mixture model resulting from interpolating control and dementia LMs improves upon the current state-of-the-art for models trained on transcript text exclusively.

Zhiguo Yu, Byron C. Wallace, Todd Johnson et al.

Estimation of semantic similarity and relatedness between biomedical concepts has utility for many informatics applications. Automated methods fall into two categories: methods based on distributional statistics drawn from text corpora, and methods using the structure of existing knowledge resources. Methods in the former category disregard taxonomic structure, while those in the latter fail to consider semantically relevant empirical information. In this paper, we present a method that retrofits distributional context vector representations of biomedical concepts using structural information from the UMLS Metathesaurus, such that the similarity between vector representations of linked concepts is augmented. We evaluated it on the UMNSRS benchmark. Our results demonstrate that retrofitting of concept vector representations leads to better correlation with human raters for both similarity and relatedness, surpassing the best results reported to date. They also demonstrate a clear improvement in performance on this reference standard for retrofitted vector representations, as compared to those without retrofitting.