Search

The quality of the powder bed is critical in the laser powder bed fusion (LPBF) process, and defects in the powder bed likely affect the quality of the final part. With the development of artificial intelligence technology, machine learning methods have been widely applied in powder-bed defect detection. However, to achieve high-precision defect detection, it is often necessary to construct complex network models and use high-resolution powder bed image data. To address these issues, this study used an off-axis industrial camera to capture layer-wise powder bed image data and proposed a defect detection model based on YOLOv7x and channel pruning to achieve defect identification and localization of powder bed patch images. Furthermore, an end-to-end defect detection pipeline based on image processing methods was proposed to detect defects in layer-wise powder bed images. Finally, the gradient-based class activation map technique (Grad CAM++) was used to analyze the interpretability of the detection results of the model. The results indicated that the proposed model was more lightweight than other models (YOLOv7x, Faster R-CNN, and SSD), with a model size of only 12.4MB. The average time for detecting powder bed image patches was significantly reduced to only 3.4 ms, and the average detection accuracy was as high as 97.4%. This demonstrates that the proposed detection method has the advantages of faster detection speed, higher detection accuracy, and simpler models, providing a reference for the real-time online detection of powder bed defects.

Federated Split Learning (FSL) is a promising distributed learning paradigm in practice, which gathers the strengths of both Federated Learning (FL) and Split Learning (SL) paradigms, to ensure model privacy while diminishing the resource overhead of each client, especially on large transformer models in a resource-constrained environment, e.g., Internet of Things (IoT). However, almost all works merely investigate the performance with simple neural network models in FSL. Despite the minor efforts focusing on incorporating Vision Transformers (ViT) as model architectures, they train ViT from scratch, thereby leading to enormous training overhead in each device with limited resources. Therefore, in this paper, we harness Pre-trained Image Transformers (PITs) as the initial model, coined FedV, to accelerate the training process and improve model robustness. Furthermore, we propose FedVZ to hinder the gradient inversion attack, especially having the capability compatible with black-box scenarios, where the gradient information is unavailable. Concretely, FedVZ approximates the server gradient by utilizing a zeroth-order (ZO) optimization, which replaces the backward propagation with just one forward process. Empirically, we are the first to provide a systematic evaluation of FSL methods with PITs in real-world datasets, different partial device participations, and heterogeneous data splits. Our experiments verify the effectiveness of our algorithms.

Long-form video understanding represents a significant challenge within computer vision, demanding a model capable of reasoning over long multi-modal sequences. Motivated by the human cognitive process for long-form video understanding, we emphasize interactive reasoning and planning over the ability to process lengthy visual inputs. We introduce a novel agent-based system, VideoAgent, that employs a large language model as a central agent to iteratively identify and compile crucial information to answer a question, with vision-language foundation models serving as tools to translate and retrieve visual information. Evaluated on the challenging EgoSchema and NExT-QA benchmarks, VideoAgent achieves 54.1% and 71.3% zero-shot accuracy with only 8.4 and 8.2 frames used on average. These results demonstrate superior effectiveness and efficiency of our method over the current state-of-the-art methods, highlighting the potential of agent-based approaches in advancing long-form video understanding.

Building cross-modal applications is challenging due to limited paired multi-modal data. Recent works have shown that leveraging a pre-trained multi-modal contrastive representation space enables cross-modal tasks to be learned from uni-modal data. This is based on the assumption that contrastive optimization makes embeddings from different modalities interchangeable. However, this assumption is under-explored due to the poorly understood geometry of the multi-modal contrastive space, where a modality gap exists. In our study, we provide a theoretical explanation of this space's geometry and introduce a three-step method, $C^3$ (Connect, Collapse, Corrupt), to bridge the modality gap, enhancing the interchangeability of embeddings. Our $C^3$ method significantly improves cross-modal learning from uni-modal data, achieving state-of-the-art results on zero-shot image / audio / video captioning and text-to-image generation., Comment: Published at ICLR 2024

Image classification is one of the most fundamental capabilities of machine vision intelligence. In this work, we revisit the image classification task using visually-grounded language models (VLMs) such as GPT-4V and LLaVA. We find that existing proprietary and public VLMs, despite often using CLIP as a vision encoder and having many more parameters, significantly underperform CLIP on standard image classification benchmarks like ImageNet. To understand the reason, we explore several hypotheses concerning the inference algorithms, training objectives, and data processing in VLMs. Our analysis reveals that the primary cause is data-related: critical information for image classification is encoded in the VLM's latent space but can only be effectively decoded with enough training data. Specifically, there is a strong correlation between the frequency of class exposure during VLM training and instruction-tuning and the VLM's performance in those classes; when trained with sufficient data, VLMs can match the accuracy of state-of-the-art classification models. Based on these findings, we enhance a VLM by integrating classification-focused datasets into its training, and demonstrate that the enhanced classification performance of the VLM transfers to its general capabilities, resulting in an improvement of 11.8% on the newly collected ImageWikiQA dataset.

AimsHeart failure (HF) remains a major public health problem with increasing prevalence in China. This study evaluated the clinical performance and quality measures for HF management to identify gaps in the standardization of care for patients hospitalized for HF in China.Methods and resultsFollowing the results of China-HF stage I (2012-2015), the second stage of the China-HF was launched in 2017. Among 113 hospitals with ≥100 cases, the China-HF Stage II assessed the quality of care measures for HF and compared results with previous data in China and the US-based Get with The Guidelines-Heart Failure (GWTG-HF) registries. In total, 34 938 patients hospitalized with HF were enrolled from January 2017 to October 2020. Echocardiographic left ventricular function and natriuretic peptide test were performed in 93.7% and 93.0% of the cases, respectively. Adherence to standardized guidelines in China-HF stage II was higher than that in the China-HF stage I, but generally lower than GWTG-HF registry with 78.2% of eligible patients was prescribed oral diuretics, 78.7% renin-angiotensin-system inhibitors, and 82.2% beta-blockers. Implantable cardioverter-defibrillators and cardiac resynchronization devices were implanted in 3.9% and 14.6%, respectively. In contrast, the proportion of eligible patients discharged with spironolactone (87.8%) was higher than GWTG-HF. The median length of hospitalization was 9 (6, 12) days, and 938 (2.8%) patients died or withdrew from treatment during hospitalization.ConclusionsDespite significant improvements in the use of guideline-recommended testing and therapy, there remain major gaps in quality of care for patients hospitalized for HF in China that are generally larger than gaps observed in the United States.

A Clear Zone (CZ), the unobstructed roadside area with highly compacted soil, naturally accumulates high concentrations of pollutants from traffic activities. These pollutants are washed off by road runoff and enter waterways. In situ, treatment of polluted runoff from the CZ could not only protect water resources but also provide an opportunity to recharge groundwater. However, the soil in the CZ requires compaction, which limits the natural infiltration and treatment of road runoff. In this study, we examine whether and how amending the soil in the CZ with sand, a common bulking agent used in road design, and Expanded Shale, Clay, and Slate (ESCS) aggregates, a novel light-weight engineered bulking agent, could help treat stormwater in situ. ESCS-amended soil media infiltrated 220% more water than sand-amended soil under compaction, indicating that the addition of ESCS would make the CZ better at treating road runoff generated during high-intensity rainfall. Compared to sand-amended soil in the CZ, ESCS-amended soil provided 58% more plant-available water during prolonged drying, indicating that ESCS addition would help maintain vegetation, thereby minimizing maintenance needs. Finally, replacing sand with ESCS improved the soil capacity in the CZ to remove pollutants, including heavy metals and E. coli, indicating the performance life of ESCS-amended soil would be longer than that of sand-amended soil in the CZ. Collectively, these results indicate that the addition of ESCS as an alternative bulking agent to sand in compacted soil in the CZ could potentially treat road runoff in situ and prevent pollution originating from road infrastructure.

Yuyi Chen,1 Xuemei Zhao,1 Lin Liang,1 Pengchao Tian,1 Jiayu Feng,1 Liyan Huang,1 Boping Huang,1 Yihang Wu,1 Jing Wang,1 Jingyuan Guan,1 Xinqing Li,1 Jian Zhang,1,2 Yuhui Zhang1 1Heart Failure Center, State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, People’s Republic of China; 2Key Laboratory of Clinical Research for Cardiovascular Medications, National Health Committee, Beijing, People’s Republic of ChinaCorrespondence: Yuhui Zhang, Heart Failure Center, State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, People’s Republic of China, Tel +8615901314243, Email yuhuizhangjoy@163.com Jian Zhang, Heart Failure Center, State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, People’s Republic of China, Tel +8613911102015, Email fwzhangjian62@126.comObjective: To identify biomarkers with independent prognostic value and investigate the prognostic value of multiple biomarkers in combination in patients hospitalized with heart failure.Methods: A total of 884 consecutive patients hospitalized with heart failure from 2015 to 2017 were enrolled. Twelve biomarkers were measured on admission, and the relationships between biomarkers and outcomes were assessed.Results: During the median follow-up of 913 days, 291 patients (32.9%) suffered from primary endpoint events. Soluble suppression of tumorigenicity-2 (sST2) (per log [unit] increase, adjusted HR [95% CI]: 1.39 [1.13,1.72], P = 0.002) and big endothelin-1 (big ET-1) (per log [unit] increase, adjusted HR [95% CI]: 1.56 [1.23,1.97], P < 0.001) remained independent predictors of primary endpoint event after adjust

We consider the scenario of deep clustering, in which the available prior knowledge is limited. In this scenario, few existing state-of-the-art deep clustering methods can perform well for both non-complex topology and complex topology datasets. To address the problem, we propose a constraint utilizing symmetric InfoNCE, which helps an objective of deep clustering method in the scenario train the model so as to be efficient for not only non-complex topology but also complex topology datasets. Additionally, we provide several theoretical explanations of the reason why the constraint can enhances performance of deep clustering methods. To confirm the effectiveness of the proposed constraint, we introduce a deep clustering method named MIST, which is a combination of an existing deep clustering method and our constraint. Our numerical experiments via MIST demonstrate that the constraint is effective. In addition, MIST outperforms other state-of-the-art deep clustering methods for most of the commonly used ten benchmark datasets., Comment: 48 pages, 6 figures

The inevitable feature deviation of synthetic aperture radar (SAR) image due to the special imaging principle (depression angle variation) leads to poor recognition accuracy, especially in few-shot learning (FSL). To deal with this problem, we propose a dense graph prototype network (DGP-Net) to eliminate the feature deviation by learning potential features, and classify by learning feature distribution. The role of the prototype in this model is to solve the problem of large distance between congeneric samples taken due to the contingency of single sampling in FSL, and enhance the robustness of the model. Experimental results on the MSTAR dataset show that the DGP-Net has good classification results for SAR images with different depression angles and the recognition accuracy of it is higher than typical FSL methods.

Given the prevalence of 3D medical imaging technologies such as MRI and CT that are widely used in diagnosing and treating diverse diseases, 3D segmentation is one of the fundamental tasks of medical image analysis. Recently, Transformer-based models have started to achieve state-of-the-art performances across many vision tasks, through pre-training on large-scale natural image benchmark datasets. While works on medical image analysis have also begun to explore Transformer-based models, there is currently no optimal strategy to effectively leverage pre-trained Transformers, primarily due to the difference in dimensionality between 2D natural images and 3D medical images. Existing solutions either split 3D images into 2D slices and predict each slice independently, thereby losing crucial depth-wise information, or modify the Transformer architecture to support 3D inputs without leveraging pre-trained weights. In this work, we use a simple yet effective weight inflation strategy to adapt pre-trained Transformers from 2D to 3D, retaining the benefit of both transfer learning and depth information. We further investigate the effectiveness of transfer from different pre-training sources and objectives. Our approach achieves state-of-the-art performances across a broad range of 3D medical image datasets, and can become a standard strategy easily utilized by all work on Transformer-based models for 3D medical images, to maximize performance., Comment: Published at ML4H 2022

Recent multi-modal contrastive learning models have demonstrated the ability to learn an embedding space suitable for building strong vision classifiers, by leveraging the rich information in large-scale image-caption datasets. Our work highlights a distinct advantage of this multi-modal embedding space: the ability to diagnose vision classifiers through natural language. The traditional process of diagnosing model behaviors in deployment settings involves labor-intensive data acquisition and annotation. Our proposed method can discover high-error data slices, identify influential attributes and further rectify undesirable model behaviors, without requiring any visual data. Through a combination of theoretical explanation and empirical verification, we present conditions under which classifiers trained on embeddings from one modality can be equivalently applied to embeddings from another modality. On a range of image datasets with known error slices, we demonstrate that our method can effectively identify the error slices and influential attributes, and can further use language to rectify failure modes of the classifier., Comment: Published at ICLR 2023

This paper introduces FedSecurity, an end-to-end benchmark designed to simulate adversarial attacks and corresponding defense mechanisms in Federated Learning (FL). FedSecurity comprises two pivotal components: FedAttacker, which facilitates the simulation of a variety of attacks during FL training, and FedDefender, which implements defensive mechanisms to counteract these attacks. As an open-source library, FedSecurity enhances its usability compared to from-scratch implementations that focus on specific attack/defense scenarios based on the following features: i) It offers extensive customization options to accommodate a broad range of machine learning models (e.g., Logistic Regression, ResNet, and GAN) and FL optimizers (e.g., FedAVG, FedOPT, and FedNOVA); ii) it enables exploring the variability in the effectiveness of attacks and defenses across different datasets and models; and iii) it supports flexible configuration and customization through a configuration file and some provided APIs. We further demonstrate FedSecurity's utility and adaptability through federated training of Large Language Models (LLMs), showcasing its potential to impact a wide range of complex applications.

Language models have been shown to exhibit positive scaling, where performance improves as models are scaled up in terms of size, compute, or data. In this work, we introduce NeQA, a dataset consisting of questions with negation in which language models do not exhibit straightforward positive scaling. We show that this task can exhibit inverse scaling, U-shaped scaling, or positive scaling, and the three scaling trends shift in this order as we use more powerful prompting methods or model families. We hypothesize that solving NeQA depends on two subtasks: question answering (task 1) and negation understanding (task 2). We find that task 1 has linear scaling, while task 2 has sigmoid-shaped scaling with an emergent transition point, and composing these two scaling trends yields the final scaling trend of NeQA. Our work reveals and provides a way to analyze the complex scaling trends of language models., Comment: Published at ACL 2023 Findings

Representation learning has been increasing its impact on the research and practice of machine learning, since it enables to learn representations that can apply to various downstream tasks efficiently. However, recent works pay little attention to the fact that real-world datasets used during the stage of representation learning are commonly contaminated by noise, which can degrade the quality of learned representations. This paper tackles the problem to learn robust representations against noise in a raw dataset. To this end, inspired by recent works on denoising and the success of the cosine-similarity-based objective functions in representation learning, we propose the denoising Cosine-Similarity (dCS) loss. The dCS loss is a modified cosine-similarity loss and incorporates a denoising property, which is supported by both our theoretical and empirical findings. To make the dCS loss implementable, we also construct the estimators of the dCS loss with statistical guarantees. Finally, we empirically show the efficiency of the dCS loss over the baseline objective functions in vision and speech domains.

Contrastive learning is an efficient approach to self-supervised representation learning. Although recent studies have made progress in the theoretical understanding of contrastive learning, the investigation of how to characterize the clusters of the learned representations is still limited. In this paper, we aim to elucidate the characterization from theoretical perspectives. To this end, we consider a kernel-based contrastive learning framework termed Kernel Contrastive Learning (KCL), where kernel functions play an important role when applying our theoretical results to other frameworks. We introduce a formulation of the similarity structure of learned representations by utilizing a statistical dependency viewpoint. We investigate the theoretical properties of the kernel-based contrastive loss via this formulation. We first prove that the formulation characterizes the structure of representations learned with the kernel-based contrastive learning framework. We show a new upper bound of the classification error of a downstream task, which explains that our theory is consistent with the empirical success of contrastive learning. We also establish a generalization error bound of KCL. Finally, we show a guarantee for the generalization ability of KCL to the downstream classification task via a surrogate bound.

How do two sets of images differ? Discerning set-level differences is crucial for understanding model behaviors and analyzing datasets, yet manually sifting through thousands of images is impractical. To aid in this discovery process, we explore the task of automatically describing the differences between two $\textbf{sets}$ of images, which we term Set Difference Captioning. This task takes in image sets $D_A$ and $D_B$, and outputs a description that is more often true on $D_A$ than $D_B$. We outline a two-stage approach that first proposes candidate difference descriptions from image sets and then re-ranks the candidates by checking how well they can differentiate the two sets. We introduce VisDiff, which first captions the images and prompts a language model to propose candidate descriptions, then re-ranks these descriptions using CLIP. To evaluate VisDiff, we collect VisDiffBench, a dataset with 187 paired image sets with ground truth difference descriptions. We apply VisDiff to various domains, such as comparing datasets (e.g., ImageNet vs. ImageNetV2), comparing classification models (e.g., zero-shot CLIP vs. supervised ResNet), summarizing model failure modes (supervised ResNet), characterizing differences between generative models (e.g., StableDiffusionV1 and V2), and discovering what makes images memorable. Using VisDiff, we are able to find interesting and previously unknown differences in datasets and models, demonstrating its utility in revealing nuanced insights., Comment: CVPR 2024 Oral

Recent advances in image tokenizers, such as VQ-VAE, have enabled text-to-image generation using auto-regressive methods, similar to language modeling. However, these methods have yet to leverage pre-trained language models, despite their adaptability to various downstream tasks. In this work, we explore this gap by adapting a pre-trained language model for auto-regressive text-to-image generation, and find that pre-trained language models offer limited help. We provide a two-fold explanation by analyzing tokens from each modality. First, we demonstrate that image tokens possess significantly different semantics compared to text tokens, rendering pre-trained language models no more effective in modeling them than randomly initialized ones. Second, the text tokens in the image-text datasets are too simple compared to normal language model pre-training data, which causes the catastrophic degradation of language models' capability.

Human commonsense understanding of the physical and social world is organized around intuitive theories. These theories support making causal and moral judgments. When something bad happens, we naturally ask: who did what, and why? A rich literature in cognitive science has studied people's causal and moral intuitions. This work has revealed a number of factors that systematically influence people's judgments, such as the violation of norms and whether the harm is avoidable or inevitable. We collected a dataset of stories from 24 cognitive science papers and developed a system to annotate each story with the factors they investigated. Using this dataset, we test whether large language models (LLMs) make causal and moral judgments about text-based scenarios that align with those of human participants. On the aggregate level, alignment has improved with more recent LLMs. However, using statistical analyses, we find that LLMs weigh the different factors quite differently from human participants. These results show how curated, challenge datasets combined with insights from cognitive science can help us go beyond comparisons based merely on aggregate metrics: we uncover LLMs implicit tendencies and show to what extent these align with human intuitions., Comment: 34 pages, 7 figures. NeurIPS 2023

Expert feedback lays the foundation of rigorous research. However, the rapid growth of scholarly production and intricate knowledge specialization challenge the conventional scientific feedback mechanisms. High-quality peer reviews are increasingly difficult to obtain. Researchers who are more junior or from under-resourced settings have especially hard times getting timely feedback. With the breakthrough of large language models (LLM) such as GPT-4, there is growing interest in using LLMs to generate scientific feedback on research manuscripts. However, the utility of LLM-generated feedback has not been systematically studied. To address this gap, we created an automated pipeline using GPT-4 to provide comments on the full PDFs of scientific papers. We evaluated the quality of GPT-4's feedback through two large-scale studies. We first quantitatively compared GPT-4's generated feedback with human peer reviewer feedback in 15 Nature family journals (3,096 papers in total) and the ICLR machine learning conference (1,709 papers). The overlap in the points raised by GPT-4 and by human reviewers (average overlap 30.85% for Nature journals, 39.23% for ICLR) is comparable to the overlap between two human reviewers (average overlap 28.58% for Nature journals, 35.25% for ICLR). The overlap between GPT-4 and human reviewers is larger for the weaker papers. We then conducted a prospective user study with 308 researchers from 110 US institutions in the field of AI and computational biology to understand how researchers perceive feedback generated by our GPT-4 system on their own papers. Overall, more than half (57.4%) of the users found GPT-4 generated feedback helpful/very helpful and 82.4% found it more beneficial than feedback from at least some human reviewers. While our findings show that LLM-generated feedback can help researchers, we also identify several limitations.

Work on scaling laws has found that large language models (LMs) show predictable improvements to overall loss with increased scale (model size, training data, and compute). Here, we present evidence for the claim that LMs may show inverse scaling, or worse task performance with increased scale, e.g., due to flaws in the training objective and data. We present empirical evidence of inverse scaling on 11 datasets collected by running a public contest, the Inverse Scaling Prize, with a substantial prize pool. Through analysis of the datasets, along with other examples found in the literature, we identify four potential causes of inverse scaling: (i) preference to repeat memorized sequences over following in-context instructions, (ii) imitation of undesirable patterns in the training data, (iii) tasks containing an easy distractor task which LMs could focus on, rather than the harder real task, and (iv) correct but misleading few-shot demonstrations of the task. We release the winning datasets at https://inversescaling.com/data to allow for further investigation of inverse scaling. Our tasks have helped drive the discovery of U-shaped and inverted-U scaling trends, where an initial trend reverses, suggesting that scaling trends are less reliable at predicting the behavior of larger-scale models than previously understood. Overall, our results suggest that there are tasks for which increased model scale alone may not lead to progress, and that more careful thought needs to go into the data and objectives for training language models., Comment: Published in TMLR (2023), 39 pages

Language models (LMs) are becoming the foundation for almost all major language technologies, but their capabilities, limitations, and risks are not well understood. We present Holistic Evaluation of Language Models (HELM) to improve the transparency of language models. First, we taxonomize the vast space of potential scenarios (i.e. use cases) and metrics (i.e. desiderata) that are of interest for LMs. Then we select a broad subset based on coverage and feasibility, noting what's missing or underrepresented (e.g. question answering for neglected English dialects, metrics for trustworthiness). Second, we adopt a multi-metric approach: We measure 7 metrics (accuracy, calibration, robustness, fairness, bias, toxicity, and efficiency) for each of 16 core scenarios when possible (87.5% of the time). This ensures metrics beyond accuracy don't fall to the wayside, and that trade-offs are clearly exposed. We also perform 7 targeted evaluations, based on 26 targeted scenarios, to analyze specific aspects (e.g. reasoning, disinformation). Third, we conduct a large-scale evaluation of 30 prominent language models (spanning open, limited-access, and closed models) on all 42 scenarios, 21 of which were not previously used in mainstream LM evaluation. Prior to HELM, models on average were evaluated on just 17.9% of the core HELM scenarios, with some prominent models not sharing a single scenario in common. We improve this to 96.0%: now all 30 models have been densely benchmarked on the same core scenarios and metrics under standardized conditions. Our evaluation surfaces 25 top-level findings. For full transparency, we release all raw model prompts and completions publicly for further analysis, as well as a general modular toolkit. We intend for HELM to be a living benchmark for the community, continuously updated with new scenarios, metrics, and models., Comment: Authored by the Center for Research on Foundation Models (CRFM) at the Stanford Institute for Human-Centered Artificial Intelligence (HAI). Project page: https://crfm.stanford.edu/helm/v1.0

We present modality gap, an intriguing geometric phenomenon of the representation space of multi-modal models. Specifically, we show that different data modalities (e.g. images and text) are embedded at arm's length in their shared representation in multi-modal models such as CLIP. Our systematic analysis demonstrates that this gap is caused by a combination of model initialization and contrastive learning optimization. In model initialization, we show empirically and theoretically that the representation of a common deep neural network is restricted to a narrow cone. As a consequence, in a multi-modal model with two encoders, the representations of the two modalities are clearly apart when the model is initialized. During optimization, contrastive learning keeps the different modalities separate by a certain distance, which is influenced by the temperature parameter in the loss function. Our experiments further demonstrate that varying the modality gap distance has a significant impact in improving the model's downstream zero-shot classification performance and fairness. Our code and data are available at https://modalitygap.readthedocs.io, Comment: Published at NeurIPS 2022. Code and data are available at https://modalitygap.readthedocs.io

AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks. We call these models foundation models to underscore their critically central yet incomplete character. This report provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical principles(e.g., model architectures, training procedures, data, systems, security, evaluation, theory) to their applications (e.g., law, healthcare, education) and societal impact (e.g., inequity, misuse, economic and environmental impact, legal and ethical considerations). Though foundation models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities,and their effectiveness across so many tasks incentivizes homogenization. Homogenization provides powerful leverage but demands caution, as the defects of the foundation model are inherited by all the adapted models downstream. Despite the impending widespread deployment of foundation models, we currently lack a clear understanding of how they work, when they fail, and what they are even capable of due to their emergent properties. To tackle these questions, we believe much of the critical research on foundation models will require deep interdisciplinary collaboration commensurate with their fundamentally sociotechnical nature., Comment: Authored by the Center for Research on Foundation Models (CRFM) at the Stanford Institute for Human-Centered Artificial Intelligence (HAI). Report page with citation guidelines: https://crfm.stanford.edu/report.html

We introduce biomedical and clinical English model packages for the Stanza Python NLP library. These packages offer accurate syntactic analysis and named entity recognition capabilities for biomedical and clinical text, by combining Stanza's fully neural architecture with a wide variety of open datasets as well as large-scale unsupervised biomedical and clinical text data. We show via extensive experiments that our packages achieve syntactic analysis and named entity recognition performance that is on par with or surpasses state-of-the-art results. We further show that these models do not compromise speed compared to existing toolkits when GPU acceleration is available, and are made easy to download and use with Stanza's Python interface. A demonstration of our packages is available at: http://stanza.run/bio., Comment: Website: https://stanfordnlp.github.io/stanza/; demo page: http://stanza.run/bio

We introduce Stanza, an open-source Python natural language processing toolkit supporting 66 human languages. Compared to existing widely used toolkits, Stanza features a language-agnostic fully neural pipeline for text analysis, including tokenization, multi-word token expansion, lemmatization, part-of-speech and morphological feature tagging, dependency parsing, and named entity recognition. We have trained Stanza on a total of 112 datasets, including the Universal Dependencies treebanks and other multilingual corpora, and show that the same neural architecture generalizes well and achieves competitive performance on all languages tested. Additionally, Stanza includes a native Python interface to the widely used Java Stanford CoreNLP software, which further extends its functionality to cover other tasks such as coreference resolution and relation extraction. Source code, documentation, and pretrained models for 66 languages are available at https://stanfordnlp.github.io/stanza., Comment: ACL2020 System Demonstration. First two authors contribute equally. Website: https://stanfordnlp.github.io/stanza

Zhendong Lv,* Linyu Jin,* Kun Wang,* Zhi Chen, Fengning Li, Yuhui Zhang, Lifeng Lao, Chun Zhou, Xinfeng Li, Hongxing Shen Department of Spine Surgery, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, People’s Republic of China*These authors contributed equally to this workCorrespondence: Xinfeng Li; Hongxing Shen No. 160 Pujian Road, Shanghai 200127, People’s Republic of ChinaTel +86-18835991449; +86-13601814912Fax +86-021-58395057Email lxfrenji@126.com; shenhxgk@126.comPurpose: Although degenerative lumbar spinal stenosis (LSS) is increasingly being diagnosed in older people, there is much uncertainty about the appropriate operative treatment options. The objective of this study was to compare the outcome of percutaneous endoscopic lumbar decompression (PELD) versus fenestration for lumbar lateral recess stenosis (LRS) in geriatric patients over 75 years old.Materials and methods: This prospective controlled study was performed on 46 consecutive over aged patients with lateral recess stenosis who underwent either PELD or fenestration. Clinical data were recorded before, 1 week, 3 months and 1.5 years after surgery using visual analog scale (VAS), Japanese Orthopaedic Association Score (JOA), The Short-Form-36 (SF-36), and the modified Macnab evaluation criteria.Results: The patients’ mean age was 82.7 years (aged 75–93 years) in PELD group and 79.1 years (aged 75–88 years) in fenestration group. No statistical difference was found between PELD group and fenestration group with regards to VAS-back pain, VAS-leg pain, JOA and at 3 months and 1.5-year follow-up. However, the PELD group had a lower mean VAS for back pain at 1 week postoperatively (P<0.05). The quality of life in PELD group achieved the same remarkable improvement as fenestration group (P>0.05). Operative time (min) was similar between two groups (p>0.05), while the PELD techniques brought advantages in b

AIMS Heart failure oral therapies (HFOTs), including beta-blockers (BB), renin-angiotensin system inhibitors (RASi) and mineralocorticoid receptor antagonists, administered before hospital discharge after acute heart failure (AHF) might improve outcome. However, concerns have been raised because early administration of HFOTs may worsen patient's condition. We hypothesized that HFOTs at hospital discharge might be associated with better post-discharge survival. METHODS AND RESULTS The study population was composed of 19 980 AHF patients from the GREAT registry. The primary and secondary outcomes were 90-day and 1-year all-cause mortality, respectively. Survival was estimated with univariate and covariate-adjusted Cox proportional hazards regression models for the whole population and after propensity-score matching. HFOTs at discharge were consistently associated with no excess mortality in the unadjusted and adjusted analyses of the whole and matched cohorts. In the matched cohort, BB and RASi at discharge were associated with lower 90-day mortality risks compared to the respective untreated groups [hazard ratio (HR) 0.56, 95% confidence interval (CI) 0.46-0.69; and HR 0.53, 95% CI 0.42-0.66, respectively]. The favourable associations of BB and RASi at discharge with 90-day mortality were present in many subgroups including patients with reduced or preserved left ventricular ejection fraction and persisted up to 1 year after discharge. The combination of RASi and BB was associated with an even lower risk of death than RASi or BB alone. CONCLUSIONS Administration of HFOTs at hospital discharge is associated with better survival of AHF patients.

Supervised learning is limited both by the quantity and quality of the labeled data. In the field of medical record tagging, writing styles between hospitals vary drastically. The knowledge learned from one hospital might not transfer well to another. This problem is amplified in veterinary medicine domain because veterinary clinics rarely apply medical codes to their records. We proposed and trained the first large-scale generative modeling algorithm in automated disease coding. We demonstrate that generative modeling can learn discriminative features when additionally trained with supervised fine-tuning. We systematically ablate and evaluate the effect of generative modeling on the final system's performance. We compare the performance of our model with several baselines in a challenging cross-hospital setting with substantial domain shift. We outperform competitive baselines by a large margin. In addition, we provide interpretation for what is learned by our model., Comment: Machine Learning for Health (ML4H) Workshop at NeurIPS 2018 arXiv:1811.07216

Studies have been carried out of the NaCl-CaCl2-NaClO3 leaching behaviour of lead and silver from lead sulphate residues from a zinc hydrometallurgical plant in Yunnan, China. The results indicate that Cl- concentration, NaClO3 concentration, leaching time and temperature had significant effects on lead and silver leaching, whilst agitation speed and the liquid-solid ratio had a secondary effect on silver leaching but little effect on lead leaching. The Ca2+ concentration had an important effect on SO2(4-) removal but little effect on lead and silver leaching. The optimum parameters for lead and silver leaching were: 207.5 g/L initial Cl- concentration; 12 g/L NaClO3 concentration; Ca2+ concentration of 1.3 times of stoichiometric quantity; temperature 90 +/-2 degrees C; a time of 1 hour; liquid-solid ratio of 8:1 to 10:1; and agitation speed of 450 rpm. The lead and silver leaching were over 98% and 95%, respectively. The removal ratio of SO2(4-) was greater than 98%. The leaching kinetics of lead follow the shrinking core model of mixed control, and the apparent activation energy was 13.4 kJ/mol. (Authors.), Studies have been carried out of the NaCl-CaCl2-NaClO3 leaching behaviour of lead and silver from lead sulphate residues from a zinc hydrometallurgical plant in Yunnan, China. The results indicate that Cl- concentration, NaClO3 concentration, leaching time and temperature had significant effects on lead and silver leaching, whilst agitation speed and the liquid-solid ratio had a secondary effect on silver leaching but little effect on lead leaching. The Ca2+ concentration had an important effect on SO2(4-) removal but little effect on lead and silver leaching. The optimum parameters for lead and silver leaching were: 207.5 g/L initial Cl- concentration; 12 g/L NaClO3 concentration; Ca2+ concentration of 1.3 times of stoichiometric quantity; temperature 90 +/-2 degrees C; a time of 1 hour; liquid-solid ratio of 8:1 to 10:1; and agitation speed of 450 rpm. The lead and silver leaching were over 98% and 95%, respectively. The removal ratio of SO2(4-) was greater than 98%. The leaching kinetics of lead follow the shrinking core model of mixed control, and the apparent activation energy was 13.4 kJ/mol. (Authors.)