Xindi Wang

Xindi Wang, Yufei Wang, Can Xu et al. · microsoft-research

Large language models (LLMs) have shown remarkable capacity for in-context learning (ICL), where learning a new task from just a few training examples is done without being explicitly pre-trained. However, despite the success of LLMs, there has been little understanding of how ICL learns the knowledge from the given prompts. In this paper, to make progress toward understanding the learning behaviour of ICL, we train the same LLMs with the same demonstration examples via ICL and supervised learning (SL), respectively, and investigate their performance under label perturbations (i.e., noisy labels and label imbalance) on a range of classification tasks. First, via extensive experiments, we find that gold labels have significant impacts on the downstream in-context performance, especially for large language models; however, imbalanced labels matter little to ICL across all model sizes. Second, when comparing with SL, we show empirically that ICL is less sensitive to label perturbations than SL, and ICL gradually attains comparable performance to SL as the model size increases.

Xindi Wang, Robert E. Mercer, Frank Rudzicz

Currently, Medical Subject Headings (MeSH) are manually assigned to every biomedical article published and subsequently recorded in the PubMed database to facilitate retrieving relevant information. With the rapid growth of the PubMed database, large-scale biomedical document indexing becomes increasingly important. MeSH indexing is a challenging task for machine learning, as it needs to assign multiple labels to each article from an extremely large hierachically organized collection. To address this challenge, we propose KenMeSH, an end-to-end model that combines new text features and a dynamic \textbf{K}nowledge-\textbf{en}hanced mask attention that integrates document features with MeSH label hierarchy and journal correlation features to index MeSH terms. Experimental results show the proposed method achieves state-of-the-art performance on a number of measures.

Xindi Wang, Robert E. Mercer, Frank Rudzicz

Medical Subject Heading (MeSH) indexing refers to the problem of assigning a given biomedical document with the most relevant labels from an extremely large set of MeSH terms. Currently, the vast number of biomedical articles in the PubMed database are manually annotated by human curators, which is time consuming and costly; therefore, a computational system that can assist the indexing is highly valuable. When developing supervised MeSH indexing systems, the availability of a large-scale annotated text corpus is desirable. A publicly available, large corpus that permits robust evaluation and comparison of various systems is important to the research community. We release a large scale annotated MeSH indexing corpus, MeSHup, which contains 1,342,667 full text articles in English, together with the associated MeSH labels and metadata, authors, and publication venues that are collected from the MEDLINE database. We train an end-to-end model that combines features from documents and their associated labels on our corpus and report the new baseline.

Yishu Wei, Xindi Wang, Hanley Ong et al.

Despite significant progress in applying large language models (LLMs) to the medical domain, several limitations still prevent them from practical applications. Among these are the constraints on model size and the lack of cohort-specific labeled datasets. In this work, we investigated the potential of improving a lightweight LLM, such as Llama 3.1-8B, through fine-tuning with datasets using synthetic labels. Two tasks are jointly trained by combining their respective instruction datasets. When the quality of the task-specific synthetic labels is relatively high (e.g., generated by GPT4- o), Llama 3.1-8B achieves satisfactory performance on the open-ended disease detection task, with a micro F1 score of 0.91. Conversely, when the quality of the task-relevant synthetic labels is relatively low (e.g., from the MIMIC-CXR dataset), fine-tuned Llama 3.1-8B is able to surpass its noisy teacher labels (micro F1 score of 0.67 v.s. 0.63) when calibrated against curated labels, indicating the strong inherent underlying capability of the model. These findings demonstrate the potential of fine-tuning LLMs with synthetic labels, offering a promising direction for future research on LLM specialization in the medical domain.

Mengfan Shen, Kangqi Song, Xindi Wang et al.

Structured information extraction from police incident announcements is crucial for timely and accurate data processing, yet presents considerable challenges due to the variability and informal nature of textual sources such as social media posts. To address these challenges, we developed a domain-adapted extraction pipeline that leverages targeted prompt engineering with parameter-efficient fine-tuning of the Qwen2.5-7B model using Low-Rank Adaptation (LoRA). This approach enables the model to handle noisy, heterogeneous text while reliably extracting 15 key fields, including location, event characteristics, and impact assessment, from a high-quality, manually annotated dataset of 4,933 instances derived from 27,822 police briefing posts on Chinese Weibo (2019-2020). Experimental results demonstrated that LoRA-based fine-tuning significantly improved performance over both the base and instruction-tuned models, achieving an accuracy exceeding 98.36% for mortality detection and Exact Match Rates of 95.31% for fatality counts and 95.54% for province-level location extraction. The proposed pipeline thus provides a validated and efficient solution for multi-task structured information extraction in specialized domains, offering a practical framework for transforming unstructured text into reliable structured data in social science research.

John Chen, Ian Berlot-Attwell, Safwan Hossain et al.

Clinical machine learning is increasingly multimodal, collected in both structured tabular formats and unstructured forms such as freetext. We propose a novel task of exploring fairness on a multimodal clinical dataset, adopting equalized odds for the downstream medical prediction tasks. To this end, we investigate a modality-agnostic fairness algorithm - equalized odds post processing - and compare it to a text-specific fairness algorithm: debiased clinical word embeddings. Despite the fact that debiased word embeddings do not explicitly address equalized odds of protected groups, we show that a text-specific approach to fairness may simultaneously achieve a good balance of performance and classical notions of fairness. We hope that our paper inspires future contributions at the critical intersection of clinical NLP and fairness. The full source code is available here: https://github.com/johntiger1/multimodal_fairness

Xindi Wang, Haining Li, Tao Ding et al.

This paper investigates the multi-UAV multi-task coordination problem in infrastructure-less emergency scenarios, where UAVs collaboratively are required to jointly perform aerial image acquisition and ground-user communication. To tackle the challenge of balancing heterogeneous tasks within dynamic environments, we propose a hierarchical dynamic weighting Deep Reinforcement Learning (DRL) framework. Specifically, an episode-level module is introduced to capture global task preferences, while a step-level module adaptively adjusts the objective weights according to real-time system conditions. By integrating global and instantaneous weights, the proposed framework improves decision stability and responsiveness during task execution. Simulation results demonstrate that the proposed method achieves faster convergence, more stable training, and higher task completion efficiency than conventional works.

Xindi Wang, Mahsa Salmani, Parsa Omidi et al.

Recently, large language models (LLMs) have shown remarkable capabilities including understanding context, engaging in logical reasoning, and generating responses. However, this is achieved at the expense of stringent computational and memory requirements, hindering their ability to effectively support long input sequences. This survey provides an inclusive review of the recent techniques and methods devised to extend the sequence length in LLMs, thereby enhancing their capacity for long-context understanding. In particular, we review and categorize a wide range of techniques including architectural modifications, such as modified positional encoding and altered attention mechanisms, which are designed to enhance the processing of longer sequences while avoiding a proportional increase in computational requirements. The diverse methodologies investigated in this study can be leveraged across different phases of LLMs, i.e., training, fine-tuning and inference. This enables LLMs to efficiently process extended sequences. The limitations of the current methodologies is discussed in the last section along with the suggestions for future research directions, underscoring the importance of sequence length in the continued advancement of LLMs.

Cong Sun, Kurt Teichman, Yiliang Zhou et al.

In this retrospective study, a dataset was constructed with two parts. The first part included 1,656 synthetic chest radiology reports generated by GPT-4 using specified prompts, with 828 being error-free synthetic reports and 828 containing errors. The second part included 614 reports: 307 error-free reports between 2011 and 2016 from the MIMIC-CXR database and 307 corresponding synthetic reports with errors generated by GPT-4 on the basis of these MIMIC-CXR reports and specified prompts. All errors were categorized into four types: negation, left/right, interval change, and transcription errors. Then, several models, including Llama-3, GPT-4, and BiomedBERT, were refined using zero-shot prompting, few-shot prompting, or fine-tuning strategies. Finally, the performance of these models was evaluated using the F1 score, 95\% confidence interval (CI) and paired-sample t-tests on our constructed dataset, with the prediction results further assessed by radiologists. Using zero-shot prompting, the fine-tuned Llama-3-70B-Instruct model achieved the best performance with the following F1 scores: 0.769 for negation errors, 0.772 for left/right errors, 0.750 for interval change errors, 0.828 for transcription errors, and 0.780 overall. In the real-world evaluation phase, two radiologists reviewed 200 randomly selected reports output by the model. Of these, 99 were confirmed to contain errors detected by the models by both radiologists, and 163 were confirmed to contain model-detected errors by at least one radiologist. Generative LLMs, fine-tuned on synthetic and MIMIC-CXR radiology reports, greatly enhanced error detection in radiology reports.

Yinuo Wang, Baiyang Wang, Robert E. Mercer et al.

Trustworthiness in healthcare question-answering (QA) systems is important for ensuring patient safety, clinical effectiveness, and user confidence. As large language models (LLMs) become increasingly integrated into medical settings, the reliability of their responses directly influences clinical decision-making and patient outcomes. However, achieving comprehensive trustworthiness in medical QA poses significant challenges due to the inherent complexity of healthcare data, the critical nature of clinical scenarios, and the multifaceted dimensions of trustworthy AI. In this survey, we systematically examine six key dimensions of trustworthiness in medical QA, i.e., Factuality, Robustness, Fairness, Safety, Explainability, and Calibration. We review how each dimension is evaluated in existing LLM-based medical QA systems. We compile and compare major benchmarks designed to assess these dimensions and analyze evaluation-guided techniques that drive model improvements, such as retrieval-augmented grounding, adversarial fine-tuning, and safety alignment. Finally, we identify open challenges-such as scalable expert evaluation, integrated multi-dimensional metrics, and real-world deployment studies-and propose future research directions to advance the safe, reliable, and transparent deployment of LLM-powered medical QA.

Xuri Ge, Chunhao Wang, Xindi Wang et al.

Composed Image Retrieval (CIR) aims to retrieve target images based on a reference image and modified texts. However, existing methods often struggle to extract the correct semantic cues from the reference image that best reflect the user's intent under textual modification prompts, resulting in interference from irrelevant visual noise. In this paper, we propose a novel Multi-level Vision Selection by Multi-modal Chain-of-Thought Reasoning (MCoT-MVS) for CIR, integrating attention-aware multi-level vision features guided by reasoning cues from a multi-modal large language model (MLLM). Specifically, we leverage an MLLM to perform chain-of-thought reasoning on the multimodal composed input, generating the retained, removed, and target-inferred texts. These textual cues subsequently guide two reference visual attention selection modules to selectively extract discriminative patch-level and instance-level semantics from the reference image. Finally, to effectively fuse these multi-granular visual cues with the modified text and the imagined target description, we design a weighted hierarchical combination module to align the composed query with target images in a unified embedding space. Extensive experiments on two CIR benchmarks, namely CIRR and FashionIQ, demonstrate that our approach consistently outperforms existing methods and achieves new state-of-the-art performance. Code and trained models are publicly released.

John Giorgi, Xindi Wang, Nicola Sahar et al.

Named entity recognition (NER) and relation extraction (RE) are two important tasks in information extraction and retrieval (IE \& IR). Recent work has demonstrated that it is beneficial to learn these tasks jointly, which avoids the propagation of error inherent in pipeline-based systems and improves performance. However, state-of-the-art joint models typically rely on external natural language processing (NLP) tools, such as dependency parsers, limiting their usefulness to domains (e.g. news) where those tools perform well. The few neural, end-to-end models that have been proposed are trained almost completely from scratch. In this paper, we propose a neural, end-to-end model for jointly extracting entities and their relations which does not rely on external NLP tools and which integrates a large, pre-trained language model. Because the bulk of our model's parameters are pre-trained and we eschew recurrence for self-attention, our model is fast to train. On 5 datasets across 3 domains, our model matches or exceeds state-of-the-art performance, sometimes by a large margin.

Xindi Wang, Onur Varol, Tina Eliassi-Rad

Many real-world prediction tasks have outcome variables that have characteristic heavy-tail distributions. Examples include copies of books sold, auction prices of art pieces, demand for commodities in warehouses, etc. By learning heavy-tailed distributions, "big and rare" instances (e.g., the best-sellers) will have accurate predictions. Most existing approaches are not dedicated to learning heavy-tailed distribution; thus, they heavily under-predict such instances. To tackle this problem, we introduce Learning to Place (L2P), which exploits the pairwise relationships between instances for learning. In its training phase, L2P learns a pairwise preference classifier: is instance A > instance B? In its placing phase, L2P obtains a prediction by placing the new instance among the known instances. Based on its placement, the new instance is then assigned a value for its outcome variable. Experiments on real data show that L2P outperforms competing approaches in terms of accuracy and ability to reproduce heavy-tailed outcome distribution. In addition, L2P provides an interpretable model by placing each predicted instance in relation to its comparable neighbors. Interpretable models are highly desirable when lives and treasure are at stake.

Xindi Wang, Syed Arefinul Haque

In this paper we cluster 330 classical music pieces collected from MusicNet database based on their musical note sequence. We use shingling and chord trajectory matrices to create signature for each music piece and performed spectral clustering to find the clusters. Based on different resolution, the output clusters distinctively indicate composition from different classical music era and different composing style of the musicians.