Sparse observations and coarse-resolution climate models limit effective regional decision-making, underscoring the need for robust downscaling. However, existing AI methods struggle with generalization across variables and geographies and are constrained by the quadratic complexity of Vision Transformer (ViT) self-attention. We introduce ORBIT-2, a scalable foundation model for global, hyper-resolution climate downscaling. ORBIT-2 incorporates two key innovations: (1) Residual Slim ViT (Reslim), a lightweight architecture with residual learning and Bayesian regularization for efficient, robust prediction; and (2) TILES, a tile-wise sequence scaling algorithm that reduces self-attention complexity from quadratic to linear, enabling long-sequence processing and massive parallelism. ORBIT-2 scales to 10 billion parameters across 65,536 GPUs, achieving up to 4.1 ExaFLOPS sustained throughput and 74–98% strong scaling efficiency. It supports downscaling to 0.9 km global resolution and processes sequences up to 4.2 billion tokens. On 7 km resolution benchmarks, ORBIT-2 achieves high accuracy with R2 scores in range of 0.98–0.99 against observation data.

The process of deciphering, extracting, and compiling information from texts dense with domain-specific terminology and technical jargon is a challenging endeavor. It demands considerable expertise and deep knowledge in the respective field, resulting in a labor-intensive process when executed by humans. Furthermore, the task of identifying multiple class labels in extensive texts presents a challenge due to intra- and inter-reader variability, making the process time-consuming and costly.We’re introducing a user-friendly graphical interface, fortified with a BERT model-powered decision support system. This advanced system aims to augment efficiency, curtail data collection time, and sustain high precision in data acquisition. It is instrumental in deciphering and synthesizing intricate texts teeming with a spectrum of expressions, even within similar mitigation categories. Such tasks traditionally demand substantial human effort and specialized knowledge in the domain.Our system is specifically engineered for the task of extracting environmental mitigation information to promote sustainable hydropower development from licenses issued by the Federal Energy Regulatory Commission (FERC). These license documents are comprehensive, each containing over 15,000 words and requiring the identification of 135 different class labels. We anticipate that our system will boost reading speed, improve the consistency of classification outputs among readers, and contribute to the development of a robust scientific database of environmental mitigations associated with the 2,000+ non-federal hydropower facilities licensed by FERC in the United States.

INTRODUCTION: Machine learning algorithms are expected to work side-by-side with humans in decision-making pipelines. Thus, the ability of classifiers to make reliable decisions is of paramount importance. Deep neural networks (DNNs) represent the state-of-the-art models to address real-world classification. Although the strength of activation in DNNs is often correlated with the network's confidence, in-depth analyses are needed to establish whether they are well calibrated. METHOD: In this paper, we demonstrate the use of DNN-based classification tools to benefit cancer registries by automating information extraction of disease at diagnosis and at surgery from electronic text pathology reports from the US National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) population-based cancer registries. In particular, we introduce multiple methods for selective classification to achieve a target level of accuracy on multiple classification tasks while minimizing the rejection amount-that is, the number of electronic pathology reports for which the model's predictions are unreliable. We evaluate the proposed methods by comparing our approach with the current in-house deep learning-based abstaining classifier. RESULTS: Overall, all the proposed selective classification methods effectively allow for achieving the targeted level of accuracy or higher in a trade-off analysis aimed to minimize the rejection rate. On in-distribution validation and holdout test data, with all the proposed methods, we achieve on all tasks the required target level of accuracy with a lower rejection rate than the deep abstaining classifier (DAC). Interpreting the results for the out-of-distribution test data is more complex; nevertheless, in this case as well, the rejection rate from the best among the proposed methods achieving 97% accuracy or higher is lower than the rejection rate based on the DAC. CONCLUSIONS: We show that although both approaches can flag those samples that should be manually reviewed and labeled by human annotators, the newly proposed methods retain a larger fraction and do so without retraining-thus offering a reduced computational cost compared with the in-house deep learning-based abstaining classifier.

The peculiar nature of whole slide imaging (WSI), digitizing conventional glass slides to obtain multiple high resolution images which capture microscopic details of a patient’s histopathological features, has garnered increased interest from the computer vision research community over the last two decades. Given the unique computational space and time complexity inherent to gigapixel-size whole slide image data, researchers have proposed novel machine learning algorithms to aid in the performance of diagnostic tasks in clinical pathology. One effective algorithm represents a Whole slide image as a bag of smaller image patches, which can be represented as low-dimension image patch embeddings. Weakly supervised deep-learning methods, such as cluster-constrained-attention multiple instance learning (CLAM), have shown promising results when combined with image patch embeddings. While traditional ensemble classifiers yield improved task performance, such methods come with a steep cost in model complexity. Through knowledge distillation, it is possible to retain some performance improvements from an ensemble, while minimizing costs to model complexity. In this work, we implement a weakly supervised ensemble using clustering-constrained-attention multiple-instance learners (CLAM), which uses attention and instance-level clustering to identify task salient regions and feature extraction in whole slides. By applying logit-based and attention-based knowledge distillation, we show it is possible to retain some performance improvements resulting from the ensemble at zero cost to model complexity.

In the last decade, the widespread adoption of electronic health record documentation has created huge opportunities for information mining. Natural language processing (NLP) techniques using machine and deep learning are becoming increasingly widespread for information extraction tasks from unstructured clinical notes. Disparities in performance when deploying machine learning models in the real world have recently received considerable attention. In the clinical NLP domain, the robustness of convolutional neural networks (CNNs) for classifying cancer pathology reports under natural distribution shifts remains understudied. In this research, we aim to quantify and improve the performance of the CNN for text classification on out-of-distribution (OOD) datasets resulting from the natural evolution of clinical text in pathology reports. We identified class imbalance due to different prevalence of cancer types as one of the sources of performance drop and analyzed the impact of previous methods for addressing class imbalance when deploying models in real-world domains. Our results show that our novel class-specialized ensemble technique outperforms other methods for the classification of rare cancer types in terms of macro F1 scores. We also found that traditional ensemble methods perform better in top classes, leading to higher micro F1 scores. Based on our findings, we formulate a series of recommendations for other ML practitioners on how to build robust models with extremely imbalanced datasets in biomedical NLP applications.

Recent applications ofdeep learning have shown promising results for classifying unstructured text in the healthcare domain. However, the reliability of models in production settings has been hindered by imbalanced data sets in which a small subset of the classes dominate. In the absence of adequate training data, rare classes necessitate additional model constraints for robust performance. Here, we present a strategy for incorporating short sequences of text (i.e. keywords) into training to boost model accuracy on rare classes. In our approach, we assemble a set of keywords, including short phrases, associated with each class. The keywords are then used as additional data during each batch of model training, resulting in a training loss that has contributions from both raw data and keywords. We evaluate our approach on classification of cancer pathology reports, which shows a substantial increase in model performance for rare classes. Furthermore, we analyze the impact of keywords on model output probabilities for bigrams, providing a straightforward method to identify model difficulties for limited training data.

Reusing data is difficult even within well-defined science communities and only gets worse when combining data from multiple communities and disciplines. Through the lens of current work on constructing an environmental epidemiological data set from multiple disciplinary sources, we demonstrate the need for a new tool ecosystem to support heterogeneous Big Data science. Extending existing community standards for schemas and/or data formats through human auditing and wrangling of the data is not feasible at scale. This work therefore suggests new approaches for the multi-disciplinary communities to build a shared tool ecosystem for big data. We discuss both the larger context of data wrangling of epidemiological data sets for novel artificial intelligence algorithms and the specific lessons from working with these multi-disciplinary data sets. Adopting a more model-driven, automatable approach promises not only better efficiency but also removes key sources of human-generated errors and promotes reuse and reproducibility of science data.

BACKGROUND: Automated text classification has many important applications in the clinical setting; however, obtaining labelled data for training machine learning and deep learning models is often difficult and expensive. Active learning techniques may mitigate this challenge by reducing the amount of labelled data required to effectively train a model. In this study, we analyze the effectiveness of 11 active learning algorithms on classifying subsite and histology from cancer pathology reports using a Convolutional Neural Network as the text classification model. RESULTS: We compare the performance of each active learning strategy using two differently sized datasets and two different classification tasks. Our results show that on all tasks and dataset sizes, all active learning strategies except diversity-sampling strategies outperformed random sampling, i.e., no active learning. On our large dataset (15K initial labelled samples, adding 15K additional labelled samples each iteration of active learning), there was no clear winner between the different active learning strategies. On our small dataset (1K initial labelled samples, adding 1K additional labelled samples each iteration of active learning), marginal and ratio uncertainty sampling performed better than all other active learning techniques. We found that compared to random sampling, active learning strongly helps performance on rare classes by focusing on underrepresented classes. CONCLUSIONS: Active learning can save annotation cost by helping human annotators efficiently and intelligently select which samples to label. Our results show that a dataset constructed using effective active learning techniques requires less than half the amount of labelled data to achieve the same performance as a dataset constructed using random sampling.

Bidirectional Encoder Representations from Transformers (BERT) and BERT-based approaches are the current state-of-the-art in many natural language processing (NLP) tasks; however, their application to document classification on long clinical texts is limited. In this work, we introduce four methods to scale BERT, which by default can only handle input sequences up to approximately 400 words long, to perform document classification on clinical texts several thousand words long. We compare these methods against two much simpler architectures - a word-level convolutional neural network and a hierarchical self-attention network - and show that BERT often cannot beat these simpler baselines when classifying MIMIC-III discharge summaries and SEER cancer pathology reports. In our analysis, we show that two key components of BERT - pretraining and WordPiece tokenization - may actually be inhibiting BERT's performance on clinical text classification tasks where the input document is several thousand words long and where correctly identifying labels may depend more on identifying a few key words or phrases rather than understanding the contextual meaning of sequences of text.

OBJECTIVE: We implement 2 different multitask learning (MTL) techniques, hard parameter sharing and cross-stitch, to train a word-level convolutional neural network (CNN) specifically designed for automatic extraction of cancer data from unstructured text in pathology reports. We show the importance of learning related information extraction (IE) tasks leveraging shared representations across the tasks to achieve state-of-the-art performance in classification accuracy and computational efficiency. MATERIALS AND METHODS: Multitask CNN (MTCNN) attempts to tackle document information extraction by learning to extract multiple key cancer characteristics simultaneously. We trained our MTCNN to perform 5 information extraction tasks: (1) primary cancer site (65 classes), (2) laterality (4 classes), (3) behavior (3 classes), (4) histological type (63 classes), and (5) histological grade (5 classes). We evaluated the performance on a corpus of 95 231 pathology documents (71 223 unique tumors) obtained from the Louisiana Tumor Registry. We compared the performance of the MTCNN models against single-task CNN models and 2 traditional machine learning approaches, namely support vector machine (SVM) and random forest classifier (RFC). RESULTS: MTCNNs offered superior performance across all 5 tasks in terms of classification accuracy as compared with the other machine learning models. Based on retrospective evaluation, the hard parameter sharing and cross-stitch MTCNN models correctly classified 59.04% and 57.93% of the pathology reports respectively across all 5 tasks. The baseline models achieved 53.68% (CNN), 46.37% (RFC), and 36.75% (SVM). Based on prospective evaluation, the percentages of correctly classified cases across the 5 tasks were 60.11% (hard parameter sharing), 58.13% (cross-stitch), 51.30% (single-task CNN), 42.07% (RFC), and 35.16% (SVM). Moreover, hard parameter sharing MTCNNs outperformed the other models in computational efficiency by using about the same number of trainable parameters as a single-task CNN. CONCLUSIONS: The hard parameter sharing MTCNN offers superior classification accuracy for automated coding support of pathology documents across a wide range of cancers and multiple information extraction tasks while maintaining similar training and inference time as those of a single task-specific model.

Graph-of-words is a flexible and efficient text representation which addresses well-known challenges, such as word ordering and variation of expressions, to natural language processing. In this paper, we consider the latest graph-based convolutional neural network technique, the Text GraphConvolutional Network (Text GCN), in the context of performingclassification tasks on free-form natural language texts. To do this, we designed a study of multi-task information extraction from medical text documents. We implemented multi-task learning in the Text GCN, performed hyperparameter optimization, and measured the clinical task performances. We evaluated micro and macro-F1 scores of four information extraction tasks,including subsite, laterality, behavior, and histological grades from cancer pathology reports. The scores for the Text GCN significantly outperformed our previous studies with convolutional neural networks, suggesting that the Text GCN model is superior to traditional models in task performance.

We introduce a deep learning architecture, hierarchical self-attention networks (HiSANs), designed for classifying pathology reports and show how its unique architecture leads to a new state-of-the-art in accuracy, faster training, and clear interpretability. We evaluate performance on a corpus of 374,899 pathology reports obtained from the National Cancer Institute's (NCI) Surveillance, Epidemiology, and End Results (SEER) program. Each pathology report is associated with five clinical classification tasks - site, laterality, behavior, histology, and grade. We compare the performance of the HiSAN against other machine learning and deep learning approaches commonly used on medical text data - Naive Bayes, logistic regression, convolutional neural networks, and hierarchical attention networks (the previous state-of-the-art). We show that HiSANs are superior to other machine learning and deep learning text classifiers in both accuracy and macro F-score across all five classification tasks. Compared to the previous state-of-the-art, hierarchical attention networks, HiSANs not only are an order of magnitude faster to train, but also achieve about 1% better relative accuracy and 5% better relative macro F-score.

BACKGROUND: Deep Learning (DL) has advanced the state-of-the-art capabilities in bioinformatics applications which has resulted in trends of increasingly sophisticated and computationally demanding models trained by larger and larger data sets. This vastly increased computational demand challenges the feasibility of conducting cutting-edge research. One solution is to distribute the vast computational workload across multiple computing cluster nodes with data parallelism algorithms. In this study, we used a High-Performance Computing environment and implemented the Downpour Stochastic Gradient Descent algorithm for data parallelism to train a Convolutional Neural Network (CNN) for the natural language processing task of information extraction from a massive dataset of cancer pathology reports. We evaluated the scalability improvements using data parallelism training and the Titan supercomputer at Oak Ridge Leadership Computing Facility. To evaluate scalability, we used different numbers of worker nodes and performed a set of experiments comparing the effects of different training batch sizes and optimizer functions. RESULTS: We found that Adadelta would consistently converge at a lower validation loss, though requiring over twice as many training epochs as the fastest converging optimizer, RMSProp. The Adam optimizer consistently achieved a close 2nd place minimum validation loss significantly faster; using a batch size of 16 and 32 allowed the network to converge in only 4.5 training epochs. CONCLUSIONS: We demonstrated that the networked training process is scalable across multiple compute nodes communicating with message passing interface while achieving higher classification accuracy compared to a traditional machine learning algorithm.

OBJECTIVE: We explored how a deep learning (DL) approach based on hierarchical attention networks (HANs) can improve model performance for multiple information extraction tasks from unstructured cancer pathology reports compared to conventional methods that do not sufficiently capture syntactic and semantic contexts from free-text documents. MATERIALS AND METHODS: Data for our analyses were obtained from 942 deidentified pathology reports collected by the National Cancer Institute Surveillance, Epidemiology, and End Results program. The HAN was implemented for 2 information extraction tasks: (1) primary site, matched to 12 International Classification of Diseases for Oncology topography codes (7 breast, 5 lung primary sites), and (2) histological grade classification, matched to G1-G4. Model performance metrics were compared to conventional machine learning (ML) approaches including naive Bayes, logistic regression, support vector machine, random forest, and extreme gradient boosting, and other DL models, including a recurrent neural network (RNN), a recurrent neural network with attention (RNN w/A), and a convolutional neural network. RESULTS: Our results demonstrate that for both information tasks, HAN performed significantly better compared to the conventional ML and DL techniques. In particular, across the 2 tasks, the mean micro and macro F-scores for the HAN with pretraining were (0.852,0.708), compared to naive Bayes (0.518, 0.213), logistic regression (0.682, 0.453), support vector machine (0.634, 0.434), random forest (0.698, 0.508), extreme gradient boosting (0.696, 0.522), RNN (0.505, 0.301), RNN w/A (0.637, 0.471), and convolutional neural network (0.714, 0.460). CONCLUSIONS: HAN-based DL models show promise in information abstraction tasks within unstructured clinical pathology reports.

Pathology reports are a primary source of information for cancer registries which process high volumes of free-text reports annually. Information extraction and coding is a manual, labor-intensive process. In this study, we investigated deep learning and a convolutional neural network (CNN), for extracting ICD-O-3 topographic codes from a corpus of breast and lung cancer pathology reports. We performed two experiments, using a CNN and a more conventional term frequency vector approach, to assess the effects of class prevalence and inter-class transfer learning. The experiments were based on a set of 942 pathology reports with human expert annotations as the gold standard. CNN performance was compared against a more conventional term frequency vector space approach. We observed that the deep learning models consistently outperformed the conventional approaches in the class prevalence experiment, resulting in micro- and macro-F score increases of up to 0.132 and 0.226, respectively, when class labels were well populated. Specifically, the best performing CNN achieved a micro-F score of 0.722 over 12 ICD-O-3 topography codes. Transfer learning provided a consistent but modest performance boost for the deep learning methods but trends were contingent on the CNN method and cancer site. These encouraging results demonstrate the potential of deep learning for automated abstraction of pathology reports.

With large deep neural networks (DNNs) necessary to solve complex and data-intensive problems, energy efficiency is a key bottleneck for effectively deploying DL in the real world. Deep spiking NNs have gained much research attention recently due to the interest in building biological neural networks and the availability of neuromorphic platforms, which can be orders of magnitude more energy efficient compared to CPUs and GPUs. Although spiking NNs have proven to be an efficient technique for solving many machine learning and computer vision problems, to the best of our knowledge, this is the first attempt to adapt spiking NNs to sparse datasets. In this paper, we study the behaviour of spiking NNs in handling NLP datasets and the sparsity in their data representation. Then, we propose a novel framework for spiking NN using the concept of stochastic computing. Specifically, instead of generating spike trains with firing rates proportional to the intensity of each value in the feature set separately, the whole feature set is treated as a distribution function and a stochastic spiking train that follow this distribution is generated. This framework reduces the connectivity between NN layers from O(N) to O(log N). Also, it encodes input data differently and make suitable to handle sparse datasets. Finally, the framework achieves high energy efficiency since it uses Integrate and Fire neurons same as conventional spiking NNs. The results show that our proposed stochastic-based SNN achieves nearly the same accuracy as the original DNN on MNIST dataset, and it has better performance than state-of-the-art SNN. Besides that stochastic-based SNN is energy efficient, where the fully connected DNN, the conventional SNN, and the data normalized SNN consume 38.24, 1.83, and 1.85-times more energy than the stochastic-based SNN, respectively. For sparse datasets, including IMDb and In-House clinical datasets, stochastic-based SNN achieves performance comparable to that of the conventional DNN. However, the conventional spiking NN has a significant decline in classification accuracy.

BACKGROUND: The World Wide Web has emerged as a powerful data source for epidemiological studies related to infectious disease surveillance. However, its potential for cancer-related epidemiological discoveries is largely unexplored. METHODS: Using advanced web crawling and tailored information extraction procedures, the authors automatically collected and analyzed the text content of 79 394 online obituary articles published between 1998 and 2014. The collected data included 51 911 cancer (27 330 breast; 9470 lung; 6496 pancreatic; 6342 ovarian; 2273 colon) and 27 483 non-cancer cases. With the derived information, the authors replicated a case-control study design to investigate the association between parity (i.e., childbearing) and cancer risk. Age-adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for each cancer type and compared to those reported in large-scale epidemiological studies. RESULTS: Parity was found to be associated with a significantly reduced risk of breast cancer (OR = 0.78, 95% CI, 0.75-0.82), pancreatic cancer (OR = 0.78, 95% CI, 0.72-0.83), colon cancer (OR = 0.67, 95% CI, 0.60-0.74), and ovarian cancer (OR = 0.58, 95% CI, 0.54-0.62). Marginal association was found for lung cancer risk (OR = 0.87, 95% CI, 0.81-0.92). The linear trend between increased parity and reduced cancer risk was dramatically more pronounced for breast and ovarian cancer than the other cancers included in the analysis. CONCLUSION: This large web-mining study on parity and cancer risk produced findings very similar to those reported with traditional observational studies. It may be used as a promising strategy to generate study hypotheses for guiding and prioritizing future epidemiological studies.

MOTIVATION: Life stories of diseased and healthy individuals are abundantly available on the Internet. Collecting and mining such online content can offer many valuable insights into patients' physical and emotional states throughout the pre-diagnosis, diagnosis, treatment and post-treatment stages of the disease compared with those of healthy subjects. However, such content is widely dispersed across the web. Using traditional query-based search engines to manually collect relevant materials is rather labor intensive and often incomplete due to resource constraints in terms of human query composition and result parsing efforts. The alternative option, blindly crawling the whole web, has proven inefficient and unaffordable for e-health researchers. RESULTS: We propose a user-oriented web crawler that adaptively acquires user-desired content on the Internet to meet the specific online data source acquisition needs of e-health researchers. Experimental results on two cancer-related case studies show that the new crawler can substantially accelerate the acquisition of highly relevant online content compared with the existing state-of-the-art adaptive web crawling technology. For the breast cancer case study using the full training set, the new method achieves a cumulative precision between 74.7 and 79.4% after 5 h of execution till the end of the 20-h long crawling session as compared with the cumulative precision between 32.8 and 37.0% using the peer method for the same time period. For the lung cancer case study using the full training set, the new method achieves a cumulative precision between 56.7 and 61.2% after 5 h of execution till the end of the 20-h long crawling session as compared with the cumulative precision between 29.3 and 32.4% using the peer method. Using the reduced training set in the breast cancer case study, the cumulative precision of our method is between 44.6 and 54.9%, whereas the cumulative precision of the peer method is between 24.3 and 26.3%; for the lung cancer case study using the reduced training set, the cumulative precisions of our method and the peer method are, respectively, between 35.7 and 46.7% versus between 24.1 and 29.6%. These numbers clearly show a consistently superior accuracy of our method in discovering and acquiring user-desired online content for e-health research. AVAILABILITY AND IMPLEMENTATION: The implementation of our user-oriented web crawler is freely available to non-commercial users via the following Web site: http://bsec.ornl.gov/AdaptiveCrawler.shtml. The Web site provides a step-by-step guide on how to execute the web crawler implementation. In addition, the Web site provides the two study datasets including manually labeled ground truth, initial seeds and the crawling results reported in this article.

Computer-aided Gleason grading of prostate cancer tissue images has been in rapid development during the past decade. Automated classifiers using features derived from multiwavelet transform, fractal dimension and other measurements, and using texton forests have shown considerable successes. This paper presents our study on application of cardinal multiridgelet transform (CMRT) to prostate cancer images to extract texture features in the transform domain. CMRT can provide cardinality, approximate translation invariance and rotation invariance simultaneously. With 32 images of Gleason grade 3 and grade 4 as a training set and using texture features extracted therefrom, a support vector machine with Gaussian kernel has been trained to classify grade 3 and grade 4. The leave-one-out cross-validation showed its accuracy of 93.75% and AUC of 0.9651. 10 test images of grade 4 showed 100% accuracy.