In this paper, we consider the challenge of summarizing patients' medical progress notes in a limited data setting. For the Problem List Summarization (shared task 1A) at the BioNLP Workshop 2023, we demonstrate that Clinical-T5 fine-tuned to 765 medical clinic notes outperforms other extractive, abstractive and zero-shot baselines, yielding reasonable baseline systems for medical note summarization. Further, we introduce Hierarchical Ensemble of Summarization Models (HESM), consisting of token-level ensembles of diverse fine-tuned Clinical-T5 models, followed by Minimum Bayes Risk (MBR) decoding. Our HESM approach lead to a considerable summarization performance boost, and when evaluated on held-out challenge data achieved a ROUGE-L of 32.77, which was the best-performing system at the top of the shared task leaderboard.

The BioNLP Workshop 2023 initiated the launch of a shared task on Problem List Summarization (ProbSum) in January 2023. The aim of this shared task is to attract future research efforts in building NLP models for real-world diagnostic decision support applications, where a system generating relevant and accurate diagnoses will augment the healthcare providers decision-making process and improve the quality of care for patients. The goal for participants is to develop models that generated a list of diagnoses and problems using input from the daily care notes collected from the hospitalization of critically ill patients. Eight teams submitted their final systems to the shared task leaderboard. In this paper, we describe the tasks, datasets, evaluation metrics, and baseline systems. Additionally, the techniques and results of the evaluation of the different approaches tried by the participating teams are summarized.

Traditional robot task planning methods face challenges when dealing with highly unstructured environments and complex tasks. We propose a task planning method that combines human expertise with an LLM and have designed an LLM prompt template, Think_Net_Prompt, with stronger expressive power to represent structured professional knowledge. We further propose a method to progressively decompose tasks and generate a task tree to reduce the planning volume for each task, and we have designed a strategy to decouple robot task planning. By dividing different planning entities and separating the task from the actual machine binding process, the task planning process becomes more flexible. Research results show that our method performs well in handling specified code formats, understanding the relationship between tasks and subtasks, and extracting parameters from text descriptions. However, there are also problems such as limited complexity of task logic handling, ambiguity in the quantity of parts and the precise location of assembly. Improving the precision of task description and cognitive structure can bring certain improvements. https://github.com/NOMIzy/Think_Net_Prompt

The emergence of large-language models (LLMs) that excel at code generation and commercial products such as GitHub's Copilot has sparked interest in human-AI pair programming (referred to as "pAIr programming") where an AI system collaborates with a human programmer. While traditional pair programming between humans has been extensively studied, it remains uncertain whether its findings can be applied to human-AI pair programming. We compare human-human and human-AI pair programming, exploring their similarities and differences in interaction, measures, benefits, and challenges. We find that the effectiveness of both approaches is mixed in the literature (though the measures used for pAIr programming are not as comprehensive). We summarize moderating factors on the success of human-human pair programming, which provides opportunities for pAIr programming research. For example, mismatched expertise makes pair programming less productive, therefore well-designed AI programming assistants may adapt to differences in expertise levels.

Instruction tuning large language models (LLMs) remains a challenging task, owing to the complexity of hyperparameter selection and the difficulty involved in evaluating the tuned models. To determine the optimal hyperparameters, an automatic, robust, and reliable evaluation benchmark is essential. However, establishing such a benchmark is not a trivial task due to the challenges associated with evaluation accuracy and privacy protection. In response to these challenges, we introduce a judge large language model, named PandaLM, which is trained to distinguish the superior model given several LLMs. PandaLM's focus extends beyond just the objective correctness of responses, which is the main focus of traditional evaluation datasets. It addresses vital subjective factors such as relative conciseness, clarity, adherence to instructions, comprehensiveness, and formality. To ensure the reliability of PandaLM, we collect a diverse human-annotated test dataset, where all contexts are generated by humans and labels are aligned with human preferences. Our results indicate that PandaLM-7B achieves 93.75% of GPT-3.5's evaluation ability and 88.28% of GPT-4's in terms of F1-score on our test dataset. PandaLM enables the evaluation of LLM to be fairer but with less cost, evidenced by significant improvements achieved by models tuned through PandaLM compared to their counterparts trained with default Alpaca's hyperparameters. In addition, PandaLM does not depend on API-based evaluations, thus avoiding potential data leakage. All resources of PandaLM are released at https://github.com/WeOpenML/PandaLM.

A few benchmarking datasets have been released to evaluate the factual knowledge of pretrained language models. These benchmarks (e.g., LAMA, and ParaRel) are mainly developed in English and later are translated to form new multilingual versions (e.g., mLAMA, and mParaRel). Results on these multilingual benchmarks suggest that using English prompts to recall the facts from multilingual models usually yields significantly better and more consistent performance than using non-English prompts. Our analysis shows that mLAMA is biased toward facts from Western countries, which might affect the fairness of probing models. We propose a new framework for curating factual triples from Wikidata that are culturally diverse. A new benchmark DLAMA-v1 is built of factual triples from three pairs of contrasting cultures having a total of 78,259 triples from 20 relation predicates. The three pairs comprise facts representing the (Arab and Western), (Asian and Western), and (South American and Western) countries respectively. Having a more balanced benchmark (DLAMA-v1) supports that mBERT performs better on Western facts than non-Western ones, while monolingual Arabic, English, and Korean models tend to perform better on their culturally proximate facts. Moreover, both monolingual and multilingual models tend to make a prediction that is culturally or geographically relevant to the correct label, even if the prediction is wrong.

Tabular data is often hidden in text, particularly in medical diagnostic reports. Traditional machine learning (ML) models designed to work with tabular data, cannot effectively process information in such form. On the other hand, large language models (LLMs) which excel at textual tasks, are probably not the best tool for modeling tabular data. Therefore, we propose a novel, simple, and effective methodology for extracting structured tabular data from textual medical reports, called TEMED-LLM. Drawing upon the reasoning capabilities of LLMs, TEMED-LLM goes beyond traditional extraction techniques, accurately inferring tabular features, even when their names are not explicitly mentioned in the text. This is achieved by combining domain-specific reasoning guidelines with a proposed data validation and reasoning correction feedback loop. By applying interpretable ML models such as decision trees and logistic regression over the extracted and validated data, we obtain end-to-end interpretable predictions. We demonstrate that our approach significantly outperforms state-of-the-art text classification models in medical diagnostics. Given its predictive performance, simplicity, and interpretability, TEMED-LLM underscores the potential of leveraging LLMs to improve the performance and trustworthiness of ML models in medical applications.

This work introduces approaches to assessing phrase breaks in ESL learners' speech using pre-trained language models (PLMs) and large language models (LLMs). There are two tasks: overall assessment of phrase break for a speech clip and fine-grained assessment of every possible phrase break position. To leverage NLP models, speech input is first force-aligned with texts, and then pre-processed into a token sequence, including words and phrase break information. To utilize PLMs, we propose a pre-training and fine-tuning pipeline with the processed tokens. This process includes pre-training with a replaced break token detection module and fine-tuning with text classification and sequence labeling. To employ LLMs, we design prompts for ChatGPT. The experiments show that with the PLMs, the dependence on labeled training data has been greatly reduced, and the performance has improved. Meanwhile, we verify that ChatGPT, a renowned LLM, has potential for further advancement in this area.

Semantic matching is a mainstream paradigm of zero-shot relation extraction, which matches a given input with a corresponding label description. The entities in the input should exactly match their hypernyms in the description, while the irrelevant contexts should be ignored when matching. However, general matching methods lack explicit modeling of the above matching pattern. In this work, we propose a fine-grained semantic matching method tailored for zero-shot relation extraction. Following the above matching pattern, we decompose the sentence-level similarity score into entity and context matching scores. Due to the lack of explicit annotations of the redundant components, we design a feature distillation module to adaptively identify the relation-irrelevant features and reduce their negative impact on context matching. Experimental results show that our method achieves higher matching $F_1$ score and has an inference speed 10 times faster, when compared with the state-of-the-art methods.

The COVID-19 pandemic led to 1.1 million deaths in the United States, despite the explosion of coronavirus research. These new findings are slow to translate to clinical interventions, leading to poorer patient outcomes and unnecessary deaths. One reason is that clinicians, overwhelmed by patients, struggle to keep pace with the rate of new coronavirus literature. A potential solution is developing a tool for evaluating coronavirus literature using large language models (LLMs) -- neural networks that are deployed for natural language processing. LLMs can be used to summarize and extract user-specified information. The greater availability and advancement of LLMs and pre-processed coronavirus literature databases provide the opportunity to assist clinicians in evaluating coronavirus literature through a coronavirus literature specific LLM (covLLM), a tool that directly takes an inputted research article and a user query to return an answer. Using the COVID-19 Open Research Dataset (CORD-19), we produced two datasets: (1) synCovid, which uses a combination of handwritten prompts and synthetic prompts generated using OpenAI, and (2) real abstracts, which contains abstract and title pairs. covLLM was trained with LLaMA 7B as a baseline model to produce three models trained on (1) the Alpaca and synCovid datasets, (2) the synCovid dataset, and (3) the synCovid and real abstract datasets. These models were evaluated by two human evaluators and ChatGPT. Results demonstrate that training covLLM on the synCovid and abstract pairs datasets performs competitively with ChatGPT and outperforms covLLM trained primarily using the Alpaca dataset.