Yu Wang

Yifan Hu, Yu Wang

The development of IoT technology enables a variety of sensors can be integrated into mobile devices. Human Activity Recognition (HAR) based on sensor data has become an active research topic in the field of machine learning and ubiquitous computing. However, due to the inconsistent frequency of human activities, the amount of data for each activity in the human activity dataset is imbalanced. Considering the limited sensor resources and the high cost of manually labeled sensor data, human activity recognition is facing the challenge of highly imbalanced activity datasets. In this paper, we propose Balancing Sensor Data Generative Adversarial Networks (BSDGAN) to generate sensor data for minority human activities. The proposed BSDGAN consists of a generator model and a discriminator model. Considering the extreme imbalance of human activity dataset, an autoencoder is employed to initialize the training process of BSDGAN, ensure the data features of each activity can be learned. The generated activity data is combined with the original dataset to balance the amount of activity data across human activity classes. We deployed multiple human activity recognition models on two publicly available imbalanced human activity datasets, WISDM and UNIMIB. Experimental results show that the proposed BSDGAN can effectively capture the data features of real human activity sensor data, and generate realistic synthetic sensor data. Meanwhile, the balanced activity dataset can effectively help the activity recognition model to improve the recognition accuracy.

Zexue He, Yu Wang, An Yan et al.

Curated datasets for healthcare are often limited due to the need of human annotations from experts. In this paper, we present MedEval, a multi-level, multi-task, and multi-domain medical benchmark to facilitate the development of language models for healthcare. MedEval is comprehensive and consists of data from several healthcare systems and spans 35 human body regions from 8 examination modalities. With 22,779 collected sentences and 21,228 reports, we provide expert annotations at multiple levels, offering a granular potential usage of the data and supporting a wide range of tasks. Moreover, we systematically evaluated 10 generic and domain-specific language models under zero-shot and finetuning settings, from domain-adapted baselines in healthcare to general-purposed state-of-the-art large language models (e.g., ChatGPT). Our evaluations reveal varying effectiveness of the two categories of language models across different tasks, from which we notice the importance of instruction tuning for few-shot usage of large language models. Our investigation paves the way toward benchmarking language models for healthcare and provides valuable insights into the strengths and limitations of adopting large language models in medical domains, informing their practical applications and future advancements.

Xintong Li, Junda Wu, Tong Yu et al.

Instruction tuning in multimodal large language models (MLLMs) generally involves cooperative learning between a backbone LLM and a feature encoder of non-text input modalities. The major challenge is how to efficiently find the synergy between the two modules so that LLMs can adapt their reasoning abilities to downstream tasks while feature encoders can adjust to provide more task-specific information about its modality. In this paper, we analyze the MLLM instruction tuning from both theoretical and empirical perspectives, where we find the unbalanced learning between the feature encoder and the LLM can cause problems of oscillation and biased learning that lead to sub-optimal convergence. Inspired by our findings, we propose a Multimodal Balance Coefficient that enables quantitative measurement of the balance of learning. Based on this, we further design a dynamic learning scheduler that better coordinates the learning between the LLM and feature encoder, alleviating the problems of oscillation and biased learning. In addition, we introduce an auxiliary regularization on the gradient to promote updating with larger step sizes, which potentially allows for a more accurate estimation of the proposed MultiModal Balance Coefficient and further improves the training sufficiency. Our proposed approach is agnostic to the architecture of LLM and feature encoder, so it can be generically integrated with various MLLMs. We conduct experiments on multiple downstream tasks with various MLLMs, demonstrating that the proposed method is more effective than the baselines in MLLM instruction tuning.

Pingjie Wang, Ziqing Fan, Shengchao Hu et al.

Structured pruning is a promising hardware-friendly compression technique for large language models (LLMs), which is expected to be retraining-free to avoid the enormous retraining cost. This retraining-free paradigm involves (1) pruning criteria to define the architecture and (2) distortion reconstruction to restore performance. However, existing methods often emphasize pruning criteria while using reconstruction techniques that are specific to certain modules or criteria, resulting in limited generalizability. To address this, we introduce the Linear Interpolation-based Adaptive Reconstruction (LIAR) framework, which is both efficient and effective. LIAR does not require back-propagation or retraining and is compatible with various pruning criteria and modules. By applying linear interpolation to the preserved weights, LIAR minimizes reconstruction error and effectively reconstructs the pruned output. Our evaluations on benchmarks such as GLUE, SQuAD, WikiText, and common sense reasoning show that LIAR enables a BERT model to maintain 98% accuracy even after removing 50% of its parameters and achieves top performance for LLaMA in just a few minutes.

Yu Wang, Zexue He, Zhankui He et al.

Understanding and accurately explaining compatibility relationships between fashion items is a challenging problem in the burgeoning domain of AI-driven outfit recommendations. Present models, while making strides in this area, still occasionally fall short, offering explanations that can be elementary and repetitive. This work aims to address these shortcomings by introducing the Pair Fashion Explanation (PFE) dataset, a unique resource that has been curated to illuminate these compatibility relationships. Furthermore, we propose an innovative two-stage pipeline model that leverages this dataset. This fine-tuning allows the model to generate explanations that convey the compatibility relationships between items. Our experiments showcase the model's potential in crafting descriptions that are knowledgeable, aligned with ground-truth matching correlations, and that produce understandable and informative descriptions, as assessed by both automatic metrics and human evaluation. Our code and data are released at https://github.com/wangyu-ustc/PairFashionExplanation

Yu Wang, Pengchong Qiao, Chang Liu et al.

Recent advances in robust semi-supervised learning (SSL) typically filter out-of-distribution (OOD) information at the sample level. We argue that an overlooked problem of robust SSL is its corrupted information on semantic level, practically limiting the development of the field. In this paper, we take an initial step to explore and propose a unified framework termed OOD Semantic Pruning (OSP), which aims at pruning OOD semantics out from in-distribution (ID) features. Specifically, (i) we propose an aliasing OOD matching module to pair each ID sample with an OOD sample with semantic overlap. (ii) We design a soft orthogonality regularization, which first transforms each ID feature by suppressing its semantic component that is collinear with paired OOD sample. It then forces the predictions before and after soft orthogonality decomposition to be consistent. Being practically simple, our method shows a strong performance in OOD detection and ID classification on challenging benchmarks. In particular, OSP surpasses the previous state-of-the-art by 13.7% on accuracy for ID classification and 5.9% on AUROC for OOD detection on TinyImageNet dataset. The source codes are publicly available at https://github.com/rain305f/OSP.

Yu Wang, Charu Aggarwal, Tyler Derr

Recent years have witnessed the significant success of applying graph neural networks (GNNs) in learning effective node representations for classification. However, current GNNs are mostly built under the balanced data-splitting, which is inconsistent with many real-world networks where the number of training nodes can be extremely imbalanced among the classes. Thus, directly utilizing current GNNs on imbalanced data would generate coarse representations of nodes in minority classes and ultimately compromise the classification performance. This therefore portends the importance of developing effective GNNs for handling imbalanced graph data. In this work, we propose a novel Distance-wise Prototypical Graph Neural Network (DPGNN), which proposes a class prototype-driven training to balance the training loss between majority and minority classes and then leverages distance metric learning to differentiate the contributions of different dimensions of representations and fully encode the relative position of each node to each class prototype. Moreover, we design a new imbalanced label propagation mechanism to derive extra supervision from unlabeled nodes and employ self-supervised learning to smooth representations of adjacent nodes while separating inter-class prototypes. Comprehensive node classification experiments and parameter analysis on multiple networks are conducted and the proposed DPGNN almost always significantly outperforms all other baselines, which demonstrates its effectiveness in imbalanced node classification. The implementation of DPGNN is available at \url{https://github.com/YuWVandy/DPGNN}.

Ahmed Elnaggar, Michael Heinzinger, Christian Dallago et al.

Computational biology and bioinformatics provide vast data gold-mines from protein sequences, ideal for Language Models taken from NLP. These LMs reach for new prediction frontiers at low inference costs. Here, we trained two auto-regressive models (Transformer-XL, XLNet) and four auto-encoder models (BERT, Albert, Electra, T5) on data from UniRef and BFD containing up to 393 billion amino acids. The LMs were trained on the Summit supercomputer using 5616 GPUs and TPU Pod up-to 1024 cores. Dimensionality reduction revealed that the raw protein LM-embeddings from unlabeled data captured some biophysical features of protein sequences. We validated the advantage of using the embeddings as exclusive input for several subsequent tasks. The first was a per-residue prediction of protein secondary structure (3-state accuracy Q3=81%-87%); the second were per-protein predictions of protein sub-cellular localization (ten-state accuracy: Q10=81%) and membrane vs. water-soluble (2-state accuracy Q2=91%). For the per-residue predictions the transfer of the most informative embeddings (ProtT5) for the first time outperformed the state-of-the-art without using evolutionary information thereby bypassing expensive database searches. Taken together, the results implied that protein LMs learned some of the grammar of the language of life. To facilitate future work, we released our models at https://github.com/agemagician/ProtTrans.

Yusheng Liao, Shuyang Jiang, Yu Wang et al.

Large language models like ChatGPT have shown substantial progress in natural language understanding and generation, proving valuable across various disciplines, including the medical field. Despite advancements, challenges persist due to the complexity and diversity inherent in medical tasks which often require multi-task learning capabilities. Previous approaches, although beneficial, fall short in real-world applications because they necessitate task-specific annotations at inference time, limiting broader generalization. This paper introduces MING-MOE, a novel Mixture-of-Expert~(MOE)-based medical large language model designed to manage diverse and complex medical tasks without requiring task-specific annotations, thus enhancing its usability across extensive datasets. MING-MOE employs a Mixture of Low-Rank Adaptation (MoLoRA) technique, allowing for efficient parameter usage by maintaining base model parameters static while adapting through a minimal set of trainable parameters. We demonstrate that MING-MOE achieves state-of-the-art (SOTA) performance on over 20 medical tasks, illustrating a significant improvement over existing models. This approach not only extends the capabilities of medical language models but also improves inference efficiency.

Yu Wang, Ziqiao Guan, Wei Hou et al.

Chronic kidney disease (CKD) has a poor prognosis due to excessive risk factors and comorbidities associated with it. The early detection of CKD faces challenges of insufficient medical histories of positive patients and complicated risk factors. In this paper, we propose the TRACE (Transformer-RNN Autoencoder-enhanced CKD Detector) framework, an end-to-end prediction model using patients' medical history data, to deal with these challenges. TRACE presents a comprehensive medical history representation with a novel key component: a Transformer-RNN autoencoder. The autoencoder jointly learns a medical concept embedding via Transformer for each hospital visit, and a latent representation which summarizes a patient's medical history across all the visits. We compared TRACE with multiple state-of-the-art methods on a dataset derived from real-world medical records. Our model has achieved 0.5708 AUPRC with a 2.31% relative improvement over the best-performing method. We also validated the clinical meaning of the learned embeddings through visualizations and a case study, showing the potential of TRACE to serve as a general disease prediction model.

Yu Wang, JunPeng Bao, JianQiang Du et al.

The incidence of Acute Kidney Injury (AKI) commonly happens in the Intensive Care Unit (ICU) patients, especially in the adults, which is an independent risk factor affecting short-term and long-term mortality. Though researchers in recent years highlight the early prediction of AKI, the performance of existing models are not precise enough. The objective of this research is to precisely predict AKI by means of Convolutional Neural Network on Electronic Health Record (EHR) data. The data sets used in this research are two public Electronic Health Record (EHR) databases: MIMIC-III and eICU database. In this study, we take several Convolutional Neural Network models to train and test our AKI predictor, which can precisely predict whether a certain patient will suffer from AKI after admission in ICU according to the last measurements of the 16 blood gas and demographic features. The research is based on Kidney Disease Improving Global Outcomes (KDIGO) criteria for AKI definition. Our work greatly improves the AKI prediction precision, and the best AUROC is up to 0.988 on MIMIC-III data set and 0.936 on eICU data set, both of which outperform the state-of-art predictors. And the dimension of the input vector used in this predictor is much fewer than that used in other existing researches. Compared with the existing AKI predictors, the predictor in this work greatly improves the precision of early prediction of AKI by using the Convolutional Neural Network architecture and a more concise input vector. Early and precise prediction of AKI will bring much benefit to the decision of treatment, so it is believed that our work is a very helpful clinical application.

Yu Wang

In this paper, we present a fast and reliable method based on PCA to select the number of dimensions for word embeddings. First, we train one embedding with a generous upper bound (e.g. 1,000) of dimensions. Then we transform the embeddings using PCA and incrementally remove the lesser dimensions one at a time while recording the embeddings' performance on language tasks. Lastly, we select the number of dimensions while balancing model size and accuracy. Experiments using various datasets and language tasks demonstrate that we are able to train 10 times fewer sets of embeddings while retaining optimal performance. Researchers interested in training the best-performing embeddings for downstream tasks, such as sentiment analysis, question answering and hypernym extraction, as well as those interested in embedding compression should find the method helpful.