Generalization in Healthcare AI: Evaluation of a Clinical Large Language Model

Advances in large language models (LLMs) provide new opportunities in healthcare for improved patient care, clinical decision-making, and enhancement of physician and administrator workflows. However, the potential of these models importantly depends on their ability to generalize effectively across clinical environments and populations, a challenge often underestimated in early development. To better understand reasons for these challenges and inform mitigation approaches, we evaluated ClinicLLM, an LLM trained on [HOSPITAL]'s clinical notes, analyzing its performance on 30-day all-cause readmission prediction focusing on variability across hospitals and patient characteristics. We found poorer generalization particularly in hospitals with fewer samples, among patients with government and unspecified insurance, the elderly, and those with high comorbidities. To understand reasons for lack of generalization, we investigated sample sizes for fine-tuning, note content (number of words per note), patient characteristics (comorbidity level, age, insurance type, borough), and health system aspects (hospital, all-cause 30-day readmission, and mortality rates). We used descriptive statistics and supervised classification to identify features. We found that, along with sample size, patient age, number of comorbidities, and the number of words in notes are all important factors related to generalization. Finally, we compared local fine-tuning (hospital specific), instance-based augmented fine-tuning and cluster-based fine-tuning for improving generalization. Among these, local fine-tuning proved most effective, increasing AUC by 0.25% to 11.74% (most helpful in settings with limited data). Overall, this study provides new insights for enhancing the deployment of large language models in the societally important domain of healthcare, and improving their performance for broader populations.

Socratis: Are large multimodal models emotionally aware?

Existing emotion prediction benchmarks contain coarse emotion labels which do not consider the diversity of emotions that an image and text can elicit in humans due to various reasons. Learning diverse reactions to multimodal content is important as intelligent machines take a central role in generating and delivering content to society. To address this gap, we propose Socratis, a societal reactions benchmark, where each image-caption (IC) pair is annotated with multiple emotions and the reasons for feeling them. Socratis contains 18K free-form reactions for 980 emotions on 2075 image-caption pairs from 5 widely-read news and image-caption (IC) datasets. We benchmark the capability of state-of-the-art multimodal large language models to generate the reasons for feeling an emotion given an IC pair. Based on a preliminary human study, we observe that humans prefer human-written reasons over 2 times more often than machine-generated ones. This shows our task is harder than standard generation tasks because it starkly contrasts recent findings where humans cannot tell apart machine vs human-written news articles, for instance. We further see that current captioning metrics based on large vision-language models also fail to correlate with human preferences. We hope that these findings and our benchmark will inspire further research on training emotionally aware models.

Mental-LLM: Leveraging Large Language Models for Mental Health Prediction via Online Text Data

Advances in large language models (LLMs) have empowered a variety of applications. However, there is still a significant gap in research when it comes to understanding and enhancing the capabilities of LLMs in the field of mental health. In this work, we present the first comprehensive evaluation of multiple LLMs, including Alpaca, Alpaca-LoRA, FLAN-T5, GPT-3.5, and GPT-4, on various mental health prediction tasks via online text data. We conduct a broad range of experiments, covering zero-shot prompting, few-shot prompting, and instruction fine-tuning. The results indicate a promising yet limited performance of LLMs with zero-shot and few-shot prompt designs for the mental health tasks. More importantly, our experiments show that instruction finetuning can significantly boost the performance of LLMs for all tasks simultaneously. Our best-finetuned models, Mental-Alpaca and Mental-FLAN-T5, outperform the best prompt design of GPT-3.5 (25 and 15 times bigger) by 10.9% on balanced accuracy and the best of GPT-4 (250 and 150 times bigger) by 4.8%. They further perform on par with the state-of-the-art task-specific language model. We also conduct an exploratory case study on LLMs' capability on the mental health reasoning tasks, illustrating the promising capability of certain models such as GPT-4. We summarize our findings into a set of action guidelines for potential methods to enhance LLMs' capability for mental health tasks. Meanwhile, we also emphasize the important limitations before achieving deployability in real-world mental health settings, such as known racial and gender bias. We highlight the important ethical risks accompanying this line of research.

NaturalAdversaries: Can Naturalistic Adversaries Be as Effective as Artificial Adversaries?

While a substantial body of prior work has explored adversarial example generation for natural language understanding tasks, these examples are often unrealistic and diverge from the real-world data distributions. In this work, we introduce a two-stage adversarial example generation framework (NaturalAdversaries), for designing adversaries that are effective at fooling a given classifier and demonstrate natural-looking failure cases that could plausibly occur during in-the-wild deployment of the models. At the first stage a token attribution method is used to summarize a given classifier's behaviour as a function of the key tokens in the input. In the second stage a generative model is conditioned on the key tokens from the first stage. NaturalAdversaries is adaptable to both black-box and white-box adversarial attacks based on the level of access to the model parameters. Our results indicate these adversaries generalize across domains, and offer insights for future research on improving robustness of neural text classification models.

ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection

Toxic language detection systems often falsely flag text that contains minority group mentions as toxic, as those groups are often the targets of online hate. Such over-reliance on spurious correlations also causes systems to struggle with detecting implicitly toxic language. To help mitigate these issues, we create ToxiGen, a new large-scale and machine-generated dataset of 274k toxic and benign statements about 13 minority groups. We develop a demonstration-based prompting framework and an adversarial classifier-in-the-loop decoding method to generate subtly toxic and benign text with a massive pretrained language model. Controlling machine generation in this way allows ToxiGen to cover implicitly toxic text at a larger scale, and about more demographic groups, than previous resources of human-written text. We conduct a human evaluation on a challenging subset of ToxiGen and find that annotators struggle to distinguish machine-generated text from human-written language. We also find that 94.5% of toxic examples are labeled as hate speech by human annotators. Using three publicly-available datasets, we show that finetuning a toxicity classifier on our data improves its performance on human-written data substantially. We also demonstrate that ToxiGen can be used to fight machine-generated toxicity as finetuning improves the classifier significantly on our evaluation subset.

Misinfo Belief Frames: A Case Study on Covid & Climate News

Prior beliefs of readers impact the way in which they project meaning onto news headlines. These beliefs can influence their perception of news reliability, as well as their reaction to news, and their likelihood of spreading the misinformation through social networks. However, most prior work focuses on fact-checking veracity of news or stylometry rather than measuring impact of misinformation. We propose Misinfo Belief Frames, a formalism for understanding how readers perceive the reliability of news and the impact of misinformation. We also introduce the Misinfo Belief Frames (MBF) corpus, a dataset of 66k inferences over 23.5k headlines. Misinformation frames use commonsense reasoning to uncover implications of real and fake news headlines focused on global crises: the Covid-19 pandemic and climate change. Our results using large-scale language modeling to predict misinformation frames show that machine-generated inferences can influence readers' trust in news headlines (readers' trust in news headlines was affected in 29.3% of cases). This demonstrates the potential effectiveness of using generated frames to counter misinformation.

Go Figure! A Meta Evaluation of Factuality in Summarization

Text generation models can generate factually inconsistent text containing distorted or fabricated facts about the source text. Recent work has focused on building evaluation models to verify the factual correctness of semantically constrained text generation tasks such as document summarization. While the field of factuality evaluation is growing fast, we don't have well-defined criteria for measuring the effectiveness, generalizability, reliability, or sensitivity of the factuality metrics. Focusing on these aspects, in this paper, we introduce a meta-evaluation framework for evaluating factual consistency metrics. We introduce five necessary, common-sense conditions for effective factuality metrics and experiment with nine recent factuality metrics using synthetic and human-labeled factuality data from short news, long news and dialogue summarization domains. Our framework enables assessing the efficiency of any new factual consistency metric on a variety of dimensions over multiple summarization domains and can be easily extended with new meta-evaluation criteria. We also present our conclusions towards standardizing the factuality evaluation metrics.

Paragraph-Level Commonsense Transformers with Recurrent Memory

Human understanding of narrative texts requires making commonsense inferences beyond what is stated in the text explicitly. A recent model, COMeT, can generate such inferences along several dimensions such as pre- and post-conditions, motivations, and mental-states of the participants. However, COMeT was trained on short phrases, and is therefore discourse-agnostic. When presented with each sentence of a multi-sentence narrative, it might generate inferences that are inconsistent with the rest of the narrative. We present the task of discourse-aware commonsense inference. Given a sentence within a narrative, the goal is to generate commonsense inferences along predefined dimensions, while maintaining coherence with the rest of the narrative. Such large-scale paragraph-level annotation is hard to get and costly, so we use available sentence-level annotations to efficiently and automatically construct a distantly supervised corpus. Using this corpus, we train PARA-COMeT, a discourse-aware model that incorporates paragraph-level information to generate coherent commonsense inferences from narratives. PARA-COMeT captures both semantic knowledge pertaining to prior world knowledge, and episodic knowledge involving how current events relate to prior and future events in a narrative. Our results confirm that PARA-COMeT outperforms the sentence-level baselines, particularly in generating inferences that are both coherent and novel.

Detecting and Tracking Communal Bird Roosts in Weather Radar Data

The US weather radar archive holds detailed information about biological phenomena in the atmosphere over the last 20 years. Communally roosting birds congregate in large numbers at nighttime roosting locations, and their morning exodus from the roost is often visible as a distinctive pattern in radar images. This paper describes a machine learning system to detect and track roost signatures in weather radar data. A significant challenge is that labels were collected opportunistically from previous research studies and there are systematic differences in labeling style. We contribute a latent variable model and EM algorithm to learn a detection model together with models of labeling styles for individual annotators. By properly accounting for these variations we learn a significantly more accurate detector. The resulting system detects previously unknown roosting locations and provides comprehensive spatio-temporal data about roosts across the US. This data will provide biologists important information about the poorly understood phenomena of broad-scale habitat use and movements of communally roosting birds during the non-breeding season.