Hiba Ahsan

Hiba Ahsan, Denis Jered McInerney, Jisoo Kim et al. · amazon-science, salesforce

Unstructured data in Electronic Health Records (EHRs) often contains critical information -- complementary to imaging -- that could inform radiologists' diagnoses. But the large volume of notes often associated with patients together with time constraints renders manually identifying relevant evidence practically infeasible. In this work we propose and evaluate a zero-shot strategy for using LLMs as a mechanism to efficiently retrieve and summarize unstructured evidence in patient EHR relevant to a given query. Our method entails tasking an LLM to infer whether a patient has, or is at risk of, a particular condition on the basis of associated notes; if so, we ask the model to summarize the supporting evidence. Under expert evaluation, we find that this LLM-based approach provides outputs consistently preferred to a pre-LLM information retrieval baseline. Manual evaluation is expensive, so we also propose and validate a method using an LLM to evaluate (other) LLM outputs for this task, allowing us to scale up evaluation. Our findings indicate the promise of LLMs as interfaces to EHR, but also highlight the outstanding challenge posed by "hallucinations". In this setting, however, we show that model confidence in outputs strongly correlates with faithful summaries, offering a practical means to limit confabulations.

Hiba Ahsan, Byron C. Wallace

LLMs are increasingly being used in healthcare. This promises to free physicians from drudgery, enabling better care to be delivered at scale. But the use of LLMs in this space also brings risks; for example, such models may worsen existing biases. How can we spot when LLMs are (spuriously) relying on patient race to inform predictions? In this work we assess the degree to which Sparse Autoencoders (SAEs) can reveal (and control) associations the model has made between race and stigmatizing concepts. We first identify SAE latents in Gemma-2 models which appear to correlate with Black individuals. We find that this latent activates on reasonable input sequences (e.g., "African American") but also problematic words like "incarceration". We then show that we can use this latent to steer models to generate outputs about Black patients, and further that this can induce problematic associations in model outputs as a result. For example, activating the Black latent increases the risk assigned to the probability that a patient will become "belligerent". We evaluate the degree to which such steering via latents might be useful for mitigating bias. We find that this offers improvements in simple settings, but is less successful for more realistic and complex clinical tasks. Overall, our results suggest that: SAEs may offer a useful tool in clinical applications of LLMs to identify problematic reliance on demographics but mitigating bias via SAE steering appears to be of marginal utility for realistic tasks.

Hiba Ahsan, Arnab Sen Sharma, Silvio Amir et al.

We know from prior work that LLMs encode social biases, and that this manifests in clinical tasks. In this work we adopt tools from mechanistic interpretability to unveil sociodemographic representations and biases within LLMs in the context of healthcare. Specifically, we ask: Can we identify activations within LLMs that encode sociodemographic information (e.g., gender, race)? We find that gender information is highly localized in MLP layers and can be reliably manipulated at inference time via patching. Such interventions can surgically alter generated clinical vignettes for specific conditions, and also influence downstream clinical predictions which correlate with gender, e.g., patient risk of depression. We find that representation of patient race is somewhat more distributed, but can also be intervened upon, to a degree. To our knowledge, this is the first application of mechanistic interpretability methods to LLMs for healthcare.

Hiba Ahsan, Nikita Bhalla, Daivat Bhatt et al.

One of the ways blind people understand their surroundings is by clicking images and relying on descriptions generated by image captioning systems. Current work on captioning images for the visually impaired do not use the textual data present in the image when generating captions. This problem is critical as many visual scenes contain text. Moreover, up to 21% of the questions asked by blind people about the images they click pertain to the text present in them. In this work, we propose altering AoANet, a state-of-the-art image captioning model, to leverage the text detected in the image as an input feature. In addition, we use a pointer-generator mechanism to copy the detected text to the caption when tokens need to be reproduced accurately. Our model outperforms AoANet on the benchmark dataset VizWiz, giving a 35% and 16.2% performance improvement on CIDEr and SPICE scores, respectively.

Hiba Ahsan, Rahul Agrawal

Sponsored Search is a major source of revenue for web search engines. Since sponsored search follows a pay-per-click model, showing relevant ads for receiving clicks is crucial. Matching categories of a query and its ad candidates have been explored in modeling relevance of query-ad pairs. The approach involves matching cached categories of queries seen in the past to categories of candidate ads. Since queries have a heavy tail distribution, the approach has limited coverage. In this work, we propose approximating categorical similarity of a query-ad pairs using neural networks, particularly CLSM. Embedding of a query (or document) is generated using its tri-letter representation which allows coverage of tail queries. Offline experiments of incorporating this feature as opposed to using the categories directly show a 5.23% improvement in AUC ROC. A/B testing results show an improvement of 8.2% in relevance.