UNASSIGNED: Electronic health record textual sources such as medication signeturs (sigs) contain valuable information that is not always available in structured form. Commonly processed through manual annotation, this repetitive and time-consuming task could be fully automated using large language models (LLMs). While most sigs include simple instructions, some include complex patterns.
UNASSIGNED: We aimed to compare the performance of GPT-3.5 and GPT-4 with smaller fine-tuned models (ClinicalBERT, BlueBERT) in extracting the average daily dose of 2 immunomodulating medications with frequent complex sigs: hydroxychloroquine, and prednisone.
UNASSIGNED: Using manually annotated sigs as the gold standard, we compared the performance of these models in 702 hydroxychloroquine and 22 104 prednisone prescriptions.
UNASSIGNED: GPT-4 vastly outperformed all other models for this task at any level of in-context learning. With 100 in-context examples, the model correctly annotates 94% of hydroxychloroquine and 95% of prednisone sigs to within 1 significant digit. Error analysis conducted by 2 additional manual annotators on annotator-model disagreements suggests that the vast majority of disagreements are model errors. Many model errors relate to ambiguous sigs on which there was also frequent annotator disagreement.
UNASSIGNED: Paired with minimal manual annotation, GPT-4 achieved excellent performance for language regression of complex medication sigs and vastly outperforms GPT-3.5, ClinicalBERT, and BlueBERT. However, the number of in-context examples needed to reach maximum performance was similar to GPT-3.5.
UNASSIGNED: LLMs show great potential to rapidly extract structured data from sigs in no-code fashion for clinical and research applications.

In recent years, with the rise of various machine learning methods, the Ultraviolet and Near Infrared (UV-NIR) spectral analysis has been impressive in the determination of intricate systems. However, the UV-NIR spectral analysis based on traditional machine learning requires independent training with tedious parameter tuning for different samples or tasks. As a result, training a high-quality model is often complicated and time-consuming. Large language model (LLM) is one of the cutting-edge achievements in deep learning, with the parameter size of the order of billion. LLM can extract abstract information from input and use it effectively. Even without any additional training, using only simple natural language prompts, LLM can accomplish tasks that have never been seen before in completely new domains. We look forward to utilizing this capability in spectral analysis to reduce the time-consuming and operational difficulties. In this study, we used UV-NIR spectral analysis to predict the concentration of Chemical Oxygen Demand (COD) in three different water samples, including a complex wastewater. By extracting the characteristic bands in the spectrum, we input them into LLM for concentration prediction. We compared the COD prediction results of different models on water samples and discussed the effects of different experiments setting on LLM. The results show that even with brief prompts, the prediction of LLM in wastewater achieved the best performance, with R2 and RMSE equal to 0.931 and 10.966, which exceed the best results of traditional models, where R2 and RMSE correspond to 0.920 and 11.854. This result indicates that LLM, with simpler operation and less time-consuming, has ability to approach or even surpass traditional machine learning models in UV-NIR spectral analysis. In conclusion, our study proposed a new method for the UV-NIR spectral analysis based on LLM and preliminary demonstrated the potential of LLM for application.

Amblyopia is a neurodevelopmental visual disorder that affects approximately 3-5% of children globally and it can lead to vision loss if it is not diagnosed and treated early. Traditional diagnostic methods, which rely on subjective assessments and expert interpretation of eye movement recordings presents challenges in resource-limited eye care centers. This study introduces a new approach that integrates the Gemini large language model (LLM) with eye-tracking data to develop a classification tool for diagnosis of patients with amblyopia. The study demonstrates: (1) LLMs can be successfully applied to the analysis of fixation eye movement data to diagnose patients with amblyopia; and (2) Input of medical subject matter expertise, introduced in this study in the form of medical expert augmented generation (MEAG), is an effective adaption of the generic retrieval augmented generation (RAG) approach for medical applications using LLMs. This study introduces a new multi-view prompting framework for ophthalmology applications that incorporates fine granularity feedback from pediatric ophthalmologist together with in-context learning to report an accuracy of 80% in diagnosing patients with amblyopia. In addition to the binary classification task, the classification tool is generalizable to specific subpopulations of amblyopic patients based on severity of amblyopia, type of amblyopia, and with or without nystagmus. The model reports an accuracy of: (1) 83% in classifying patients with moderate or severe amblyopia, (2) 81% in classifying patients with mild or treated amblyopia; and (3) 85% accuracy in classifying patients with nystagmus. To the best of our knowledge, this is the first study that defines a multi-view prompting framework with MEAG to analyze eye tracking data for the diagnosis of amblyopic patients.

In recent years, there has been a surge in the publication of clinical trial reports, making it challenging to conduct systematic reviews. Automatically extracting Population, Intervention, Comparator, and Outcome (PICO) from clinical trial studies can alleviate the traditionally time-consuming process of manually scrutinizing systematic reviews. Existing approaches of PICO frame extraction involves supervised approach that relies on the existence of manually annotated data points in the form of BIO label tagging. Recent approaches, such as In-Context Learning (ICL), which has been shown to be effective for a number of downstream NLP tasks, require the use of labeled examples. In this work, we adopt ICL strategy by employing the pretrained knowledge of Large Language Models (LLMs), gathered during the pretraining phase of an LLM, to automatically extract the PICO-related terminologies from clinical trial documents in unsupervised set up to bypass the availability of large number of annotated data instances. Additionally, to showcase the highest effectiveness of LLM in oracle scenario where large number of annotated samples are available, we adopt the instruction tuning strategy by employing Low Rank Adaptation (LORA) to conduct the training of gigantic model in low resource environment for the PICO frame extraction task. More specifically, both of the proposed frameworks utilize AlpaCare as base LLM which employs both few-shot in-context learning and instruction tuning techniques to extract PICO-related terms from the clinical trial reports. We applied these approaches to the widely used coarse-grained datasets such as EBM-NLP, EBM-COMET and fine-grained datasets such as EBM-NLPrev and EBM-NLPh. Our empirical results show that our proposed ICL-based framework produces comparable results on all the version of EBM-NLP datasets and the proposed instruction tuned version of our framework produces state-of-the-art results on all the different EBM-NLP datasets. Our project is available at https://github.com/shrimonmuke0202/AlpaPICO.git.

BACKGROUND: Large language models (LLMs) have shown remarkable capabilities in natural language processing (NLP), especially in domains where labeled data are scarce or expensive, such as the clinical domain. However, to unlock the clinical knowledge hidden in these LLMs, we need to design effective prompts that can guide them to perform specific clinical NLP tasks without any task-specific training data. This is known as in-context learning, which is an art and science that requires understanding the strengths and weaknesses of different LLMs and prompt engineering approaches.
OBJECTIVE: The objective of this study is to assess the effectiveness of various prompt engineering techniques, including 2 newly introduced types-heuristic and ensemble prompts, for zero-shot and few-shot clinical information extraction using pretrained language models.
METHODS: This comprehensive experimental study evaluated different prompt types (simple prefix, simple cloze, chain of thought, anticipatory, heuristic, and ensemble) across 5 clinical NLP tasks: clinical sense disambiguation, biomedical evidence extraction, coreference resolution, medication status extraction, and medication attribute extraction. The performance of these prompts was assessed using 3 state-of-the-art language models: GPT-3.5 (OpenAI), Gemini (Google), and LLaMA-2 (Meta). The study contrasted zero-shot with few-shot prompting and explored the effectiveness of ensemble approaches.
RESULTS: The study revealed that task-specific prompt tailoring is vital for the high performance of LLMs for zero-shot clinical NLP. In clinical sense disambiguation, GPT-3.5 achieved an accuracy of 0.96 with heuristic prompts and 0.94 in biomedical evidence extraction. Heuristic prompts, alongside chain of thought prompts, were highly effective across tasks. Few-shot prompting improved performance in complex scenarios, and ensemble approaches capitalized on multiple prompt strengths. GPT-3.5 consistently outperformed Gemini and LLaMA-2 across tasks and prompt types.
CONCLUSIONS: This study provides a rigorous evaluation of prompt engineering methodologies and introduces innovative techniques for clinical information extraction, demonstrating the potential of in-context learning in the clinical domain. These findings offer clear guidelines for future prompt-based clinical NLP research, facilitating engagement by non-NLP experts in clinical NLP advancements. To the best of our knowledge, this is one of the first works on the empirical evaluation of different prompt engineering approaches for clinical NLP in this era of generative artificial intelligence, and we hope that it will inspire and inform future research in this area.

Human learning is sensitive to rule-like structure and the curriculum of examples used for training. In tasks governed by succinct rules, learning is more robust when related examples are blocked across trials, but in the absence of such rules, interleaving is more effective. To date, no neural model has simultaneously captured these seemingly contradictory effects. Here we show that this same tradeoff spontaneously emerges with \"in-context learning\" (ICL) both in neural networks trained with metalearning and in large language models (LLMs). ICL is the ability to learn new tasks \"in context\" - without weight changes - via an inner-loop algorithm implemented in activation dynamics. Experiments with pretrained LLMs and metalearning transformers show that ICL exhibits the blocking advantage demonstrated in humans on a task involving rule-like structure, and conversely, that concurrent in-weight learning reproduces the interleaving advantage observed in humans on tasks lacking such structure.

This paper introduces the paradigm of \"in-context operator learning\" and the corresponding model \"In-Context Operator Networks\" to simultaneously learn operators from the prompted data and apply it to new questions during the inference stage, without any weight update. Existing methods are limited to using a neural network to approximate a specific equation solution or a specific operator, requiring retraining when switching to a new problem with different equations. By training a single neural network as an operator learner, rather than a solution/operator approximator, we can not only get rid of retraining (even fine-tuning) the neural network for new problems but also leverage the commonalities shared across operators so that only a few examples in the prompt are needed when learning a new operator. Our numerical results show the capability of a single neural network as a few-shot operator learner for a diversified type of differential equation problems, including forward and inverse problems of ordinary differential equations, partial differential equations, and mean-field control problems, and also show that it can generalize its learning capability to operators beyond the training distribution.