arXiv subjects

Rahul Thapa

Publications and source records attributed to Rahul Thapa.

25 records · Page 2Linked to original sources

arXiv2023

MedAlign: A Clinician-Generated Dataset for Instruction Following with Electronic Medical Records

The ability of large language models (LLMs) to follow natural language instructions with human-level fluency suggests many opportunities in healthcare to reduce administrative burden and improve quality of care. However, evaluating LLMs on realistic text generation tasks for healthcare remains challenging. Existing question answering datasets for electronic health record (EHR) data fail to capture the complexity of information needs and documentation burdens experienced by clinicians. To address these challenges, we introduce MedAlign, a benchmark dataset of 983 natural language instructions for EHR data. MedAlign is curated by 15 clinicians (7 specialities), includes clinician-written reference responses for 303 instructions, and provides 276 longitudinal EHRs for grounding instruction-response pairs. We used MedAlign to evaluate 6 general domain LLMs, having clinicians rank the accuracy and quality of each LLM response. We found high error rates, ranging from 35% (GPT-4) to 68% (MPT-7B-Instruct), and an 8.3% drop in accuracy moving from 32k to 2k context lengths for GPT-4. Finally, we report correlations between clinician rankings and automated natural language generation metrics as a way to rank LLMs without human review. We make MedAlign available under a research data use agreement to enable LLM evaluations on tasks aligned with clinician needs and preferences.

cs.CL↗

arXiv2023

EHRSHOT: An EHR Benchmark for Few-Shot Evaluation of Foundation Models

While the general machine learning (ML) community has benefited from public datasets, tasks, and models, the progress of ML in healthcare has been hampered by a lack of such shared assets. The success of foundation models creates new challenges for healthcare ML by requiring access to shared pretrained models to validate performance benefits. We help address these challenges through three contributions. First, we publish a new dataset, EHRSHOT, which contains deidentified structured data from the electronic health records (EHRs) of 6,739 patients from Stanford Medicine. Unlike MIMIC-III/IV and other popular EHR datasets, EHRSHOT is longitudinal and not restricted to ICU/ED patients. Second, we publish the weights of CLMBR-T-base, a 141M parameter clinical foundation model pretrained on the structured EHR data of 2.57M patients. We are one of the first to fully release such a model for coded EHR data; in contrast, most prior models released for clinical data (e.g. GatorTron, ClinicalBERT) only work with unstructured text and cannot process the rich, structured data within an EHR. We provide an end-to-end pipeline for the community to validate and build upon its performance. Third, we define 15 few-shot clinical prediction tasks, enabling evaluation of foundation models on benefits such as sample efficiency and task adaptation. Our model and dataset are available via a research data use agreement from our website: https://ehrshot.stanford.edu. Code to reproduce our results are available at our Github repo: https://github.com/som-shahlab/ehrshot-benchmark

cs.LG↗

arXiv2023

Evaluation of GPT-3.5 and GPT-4 for supporting real-world information needs in healthcare delivery

Despite growing interest in using large language models (LLMs) in healthcare, current explorations do not assess the real-world utility and safety of LLMs in clinical settings. Our objective was to determine whether two LLMs can serve information needs submitted by physicians as questions to an informatics consultation service in a safe and concordant manner. Sixty six questions from an informatics consult service were submitted to GPT-3.5 and GPT-4 via simple prompts. 12 physicians assessed the LLM responses' possibility of patient harm and concordance with existing reports from an informatics consultation service. Physician assessments were summarized based on majority vote. For no questions did a majority of physicians deem either LLM response as harmful. For GPT-3.5, responses to 8 questions were concordant with the informatics consult report, 20 discordant, and 9 were unable to be assessed. There were 29 responses with no majority on "Agree", "Disagree", and "Unable to assess". For GPT-4, responses to 13 questions were concordant, 15 discordant, and 3 were unable to be assessed. There were 35 responses with no majority. Responses from both LLMs were largely devoid of overt harm, but less than 20% of the responses agreed with an answer from an informatics consultation service, responses contained hallucinated references, and physicians were divided on what constitutes harm. These results suggest that while general purpose LLMs are able to provide safe and credible responses, they often do not meet the specific information need of a given question. A definitive evaluation of the usefulness of LLMs in healthcare settings will likely require additional research on prompt engineering, calibration, and custom-tailoring of general purpose models.

cs.AI↗

arXiv2023

The Shaky Foundations of Clinical Foundation Models: A Survey of Large Language Models and Foundation Models for EMRs

The successes of foundation models such as ChatGPT and AlphaFold have spurred significant interest in building similar models for electronic medical records (EMRs) to improve patient care and hospital operations. However, recent hype has obscured critical gaps in our understanding of these models' capabilities. We review over 80 foundation models trained on non-imaging EMR data (i.e. clinical text and/or structured data) and create a taxonomy delineating their architectures, training data, and potential use cases. We find that most models are trained on small, narrowly-scoped clinical datasets (e.g. MIMIC-III) or broad, public biomedical corpora (e.g. PubMed) and are evaluated on tasks that do not provide meaningful insights on their usefulness to health systems. In light of these findings, we propose an improved evaluation framework for measuring the benefits of clinical foundation models that is more closely grounded to metrics that matter in healthcare.

cs.LG↗

arXiv2023

DEPLOYR: A technical framework for deploying custom real-time machine learning models into the electronic medical record

Machine learning (ML) applications in healthcare are extensively researched, but successful translations to the bedside are scant. Healthcare institutions are establishing frameworks to govern and promote the implementation of accurate, actionable and reliable models that integrate with clinical workflow. Such governance frameworks require an accompanying technical framework to deploy models in a resource efficient manner. Here we present DEPLOYR, a technical framework for enabling real-time deployment and monitoring of researcher created clinical ML models into a widely used electronic medical record (EMR) system. We discuss core functionality and design decisions, including mechanisms to trigger inference based on actions within EMR software, modules that collect real-time data to make inferences, mechanisms that close-the-loop by displaying inferences back to end-users within their workflow, monitoring modules that track performance of deployed models over time, silent deployment capabilities, and mechanisms to prospectively evaluate a deployed model's impact. We demonstrate the use of DEPLOYR by silently deploying and prospectively evaluating twelve ML models triggered by clinician button-clicks in Stanford Health Care's production instance of Epic. Our study highlights the need and feasibility for such silent deployment, because prospectively measured performance varies from retrospective estimates. By describing DEPLOYR, we aim to inform ML deployment best practices and help bridge the model implementation gap.

cs.LG↗

arXiv2022

MoleHD: Ultra-Low-Cost Drug Discovery using Hyperdimensional Computing

Modern drug discovery is often time-consuming, complex and cost-ineffective due to the large volume of molecular data and complicated molecular properties. Recently, machine learning algorithms have shown promising results in virtual screening of automated drug discovery by predicting molecular properties. While emerging learning methods such as graph neural networks and recurrent neural networks exhibit high accuracy, they are also notoriously computation-intensive and memory-intensive with operations such as feature embeddings or deep convolutions. In this paper, we propose a viable alternative to existing learning methods by presenting MoleHD, a method based on brain-inspired hyperdimensional computing (HDC) for molecular property prediction. We develop HDC encoders to project SMILES representation of a molecule into high-dimensional vectors that are used for HDC training and inference. We perform an extensive evaluation using 29 classification tasks from 3 widely-used molecule datasets (Clintox, BBBP, SIDER) under three splits methods (random, scaffold, and stratified). By an comprehensive comparison with 8 existing learning models including SOTA graph/recurrent neural networks, we show that MoleHD is able to achieve highest ROC-AUC score on random and scaffold splits on average across 3 datasets and achieve second-highest on stratified split. Importantly, MoleHD achieves such performance with significantly reduced computing cost and training efforts. To the best of our knowledge, this is the first HDC-based method for drug discovery. The promising results presented in this paper can potentially lead to a novel path in drug discovery research.

cs.NE↗

arXiv2021

HDXplore: Automated Blackbox Testing of Brain-Inspired Hyperdimensional Computing

Inspired by the way human brain works, the emerging hyperdimensional computing (HDC) is getting more and more attention. HDC is an emerging computing scheme based on the working mechanism of brain that computes with deep and abstract patterns of neural activity instead of actual numbers. Compared with traditional ML algorithms such as DNN, HDC is more memory-centric, granting it advantages such as relatively smaller model size, less computation cost, and one-shot learning, making it a promising candidate in low-cost computing platforms. However, the robustness of HDC models have not been systematically studied. In this paper, we systematically expose the unexpected or incorrect behaviors of HDC models by developing HDXplore, a blackbox differential testing-based framework. We leverage multiple HDC models with similar functionality as cross-referencing oracles to avoid manual checking or labeling the original input. We also propose different perturbation mechanisms in HDXplore. HDXplore automatically finds thousands of incorrect corner case behaviors of the HDC model. We propose two retraining mechanisms and using the corner cases generated by HDXplore to retrain the HDC model, we can improve the model accuracy by up to 9%.

cs.NE↗