Showing papers by "Vipin Kumar published in 2020"

Integrating Physics-Based Modeling with Machine Learning: A Survey

Abstract: In this manuscript, we provide a structured and comprehensive overview of techniques to integrate machine learning with physics-based modeling. First, we provide a summary of application areas for which these approaches have been applied. Then, we describe classes of methodologies used to construct physics-guided machine learning models and hybrid physics-machine learning frameworks from a machine learning standpoint. With this foundation, we then provide a systematic organization of these existing techniques and discuss ideas for future research.

Physics-Guided Machine Learning for Scientific Discovery: An Application in Simulating Lake Temperature Profiles

Abstract: Physics-based models of dynamical systems are often used to study engineering and environmental systems. Despite their extensive use, these models have several well-known limitations due to simplified representations of the physical processes being modeled or challenges in selecting appropriate parameters. While-state-of-the-art machine learning models can sometimes outperform physics-based models given ample amount of training data, they can produce results that are physically inconsistent. This paper proposes a physics-guided recurrent neural network model (PGRNN) that combines RNNs and physics-based models to leverage their complementary strengths and improves the modeling of physical processes. Specifically, we show that a PGRNN can improve prediction accuracy over that of physics-based models, while generating outputs consistent with physical laws. An important aspect of our PGRNN approach lies in its ability to incorporate the knowledge encoded in physics-based models. This allows training the PGRNN model using very few true observed data while also ensuring high prediction accuracy. Although we present and evaluate this methodology in the context of modeling the dynamics of temperature in lakes, it is applicable more widely to a range of scientific and engineering disciplines where physics-based (also known as mechanistic) models are used, e.g., climate science, materials science, computational chemistry, and biomedicine.

High throughput detection and genetic epidemiology of SARS-CoV-2 using COVIDSeq next generation sequencing

Abstract: The rapid emergence of coronavirus disease 2019 (COVID-19) as a global pandemic affecting millions of individuals globally has necessitated sensitive and high-throughput approaches for the diagnosis, surveillance and for determining the genetic epidemiology of SARS-CoV-2. In the present study, we used the COVIDSeq protocol, which involves multiplex-PCR, barcoding and sequencing of samples for high-throughput detection and deciphering the genetic epidemiology of SARS-CoV-2. We used the approach on 752 clinical samples in duplicates, amounting to a total of 1536 samples which could be sequenced on a single S4 sequencing flow cell on NovaSeq 6000. Our analysis suggests a high concordance between technical duplicates and a high concordance of detection of SARS-CoV-2 between the COVIDSeq as well as RT-PCR approaches. An in-depth analysis revealed a total of six samples in which COVIDSeq detected SARS-CoV-2 in high confidence which were negative in RT-PCR. Additionally, the assay could detect SARS-CoV-2 in 21 samples and 16 samples which were classified inconclusive and pan-sarbeco positive respectively suggesting that COVIDSeq could be used as a confirmatory test. The sequencing approach also enabled insights into the evolution and genetic epidemiology of the SARS-CoV-2 samples. The samples were classified into a total of 3 clades. This study reports two lineages B.1.112 and B.1.99 for the first time in India. This study also revealed 1,143 unique single nucleotide variants and added a total of 73 novel variants identified for the first time. To the best of our knowledge, this is the first report of the COVIDSeq approach for detection and genetic epidemiology of SARS-CoV-2. Our analysis suggests that COVIDSeq could be a potential high sensitivity assay for detection of SARS-CoV-2, with an additional advantage of enabling genetic epidemiology of SARS-CoV-2.

Integrating Scientific Knowledge with Machine Learning for Engineering and Environmental Systems

Abstract: There is a growing consensus that solutions to complex science and engineering problems require novel methodologies that are able to integrate traditional physics-based modeling approaches with state-of-the-art machine learning (ML) techniques. This paper provides a structured overview of such techniques. Application-centric objective areas for which these approaches have been applied are summarized, and then classes of methodologies used to construct physics-guided ML models and hybrid physics-ML frameworks are described. We then provide a taxonomy of these existing techniques, which uncovers knowledge gaps and potential crossovers of methods between disciplines that can serve as ideas for future research.

Satellite-based remote sensing data set of global surface water storage change from 1992 to 2018

Abstract: . The recent availability of freely and openly available satellite remote sensing products has enabled the implementation of global surface water monitoring at a level not previously possible. Here we present a global set of satellite-derived time series of surface water storage variations for lakes and reservoirs for a period that covers the satellite altimetry era. Our goals are to promote the use of satellite-derived products for the study of large inland water bodies and to set the stage for the expected availability of products from the Surface Water and Ocean Topography (SWOT) mission, which will vastly expand the spatial coverage of such products, expected from 2021 on. Our general strategy is to estimate global surface water storage changes ( ΔV ) in large lakes and reservoirs using a combination of paired water surface elevation (WSE) and water surface area (WSA) extent products. Specifically, we use data produced by multiple satellite altimetry missions (TOPEX/Poseidon, Jason-1, Jason-2, Jason-3, and Envisat) from 1992 on, with surface extent estimated from Terra/Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) from 2000 on. We leverage relationships between elevation and surface area (i.e., hypsometry) to produce estimates of ΔV even during periods when either of the variables was not available. This approach is successful provided that there are strong relationships between the two variables during an overlapping period. Our target is to produce time series of ΔV as well as of WSE and WSA for a set of 347 lakes and reservoirs globally for the 1992–2018 period. The data sets presented and their respective algorithm theoretical basis documents are publicly available and distributed via the Physical Oceanography Distributed Active Archive Center (PO DAAC; https://podaac.jpl.nasa.gov/ , last access: 13 May 2020) of NASA's Jet Propulsion Laboratory. Specifically, the WSE data set is available at https://doi.org/10.5067/UCLRS-GREV2 (Birkett et al., 2019), the WSA data set is available at https://doi.org/10.5067/UCLRS-AREV2 (Khandelwal and Kumar, 2019), and the ΔV data set is available at https://doi.org/10.5067/UCLRS-STOV2 (Tortini et al., 2019). The records we describe represent the most complete global surface water time series available from the launch of TOPEX/Poseidon in 1992 (beginning of the satellite altimetry era) to the near present. The production of long-term, consistent, and calibrated records of surface water cycle variables such as in the data set presented here is of fundamental importance to baseline future SWOT products.

Predicting Water Temperature Dynamics of Unmonitored Lakes with Meta Transfer Learning.

Abstract: Most environmental data come from a minority of well-observed sites. An ongoing challenge in the environmental sciences is transferring knowledge from monitored sites to unobserved sites. Here, we demonstrate a novel transfer learning framework that accurately predicts temperature in unobserved lakes (targets) by borrowing models from highly observed lakes (sources). This method, Meta Transfer Learning (MTL), builds a meta-learning model to predict transfer performance from candidate source models to targets using lake attributes and candidates' past performance. We constructed source models at 145 well-observed lakes using calibrated process-based modeling (PB) and a recently developed approach called process-guided deep learning (PGDL). We applied MTL to either PB or PGDL source models (PB-MTL or PGDL-MTL, respectively) to predict temperatures in 305 target lakes treated as unobserved in the Upper Midwestern United States. We show significantly improved performance relative to the uncalibrated process-based General Lake Model, where the median RMSE for the target lakes is $2.52^{\circ}C$. PB-MTL yielded a median RMSE of $2.43^{\circ}C$; PGDL-MTL yielded $2.16^{\circ}C$; and a PGDL-MTL ensemble of nine sources per target yielded $1.88^{\circ}C$. For sparsely observed target lakes, PGDL-MTL often outperformed PGDL models trained on the target lakes themselves. Differences in maximum depth between the source and target were consistently the most important predictors. Our approach readily scales to thousands of lakes in the Midwestern United States, demonstrating that MTL with meaningful predictor variables and high-quality source models is a promising approach for many kinds of unmonitored systems and environmental variables.

Physics Guided Machine Learning Methods for Hydrology.

Abstract: Streamflow prediction is one of the key challenges in the field of hydrology due to the complex interplay between multiple non-linear physical mechanisms behind streamflow generation. While physically-based models are rooted in rich understanding of the physical processes, a significant performance gap still remains which can be potentially addressed by leveraging the recent advances in machine learning. The goal of this work is to incorporate our understanding of physical processes and constraints in hydrology into machine learning algorithms, and thus bridge the performance gap while reducing the need for large amounts of data compared to traditional data-driven approaches. In particular, we propose an LSTM based deep learning architecture that is coupled with SWAT (Soil and Water Assessment Tool), an hydrology model that is in wide use today. The key idea of the approach is to model auxiliary intermediate processes that connect weather drivers to streamflow, rather than directly mapping runoff from weather variables which is what a deep learning architecture without physical insight will do. The efficacy of the approach is being analyzed on several small catchments located in the South Branch of the Root River Watershed in southeast Minnesota. Apart from observation data on runoff, the approach also leverages a 200-year synthetic dataset generated by SWAT to improve the performance while reducing convergence time. In the early phases of this study, simpler versions of the physics guided deep learning architectures are being used to achieve a system understanding of the coupling of physics and machine learning. As more complexity is introduced into the present implementation, the framework will be able to generalize to more sophisticated cases where spatial heterogeneity is present.

Personalized Image Retrieval with Sparse Graph Representation Learning

Abstract: Personalization is essential for enhancing the customer experience in retrieval tasks. In this paper, we develop a novel method CA-GCN for personalized image retrieval in the Adobe Stock image system. The proposed method CA-GCN leverages user behavior data in a Graph Convolutional Neural Network (GCN) model to learn user and image embeddings simultaneously. Standard GCN performs poorly on sparse user-image interaction graphs due to the limited knowledge gain from less representative neighbors. To address this challenge, we propose to augment the sparse user-image interaction data by considering the similarities among images. Specifically, we detect clusters of similar images and introduce a set of hidden super-nodes in the graph to represent clusters. We show that such an augmented graph structure can significantly improve the retrieval performance on real-world data collected from Adobe Stock service. In particular, when testing the proposed method on real users' stock image retrieval sessions, we get promoted average click position from 70 to 51.

Semi-supervised Classification using Attention-based Regularization on Coarse-resolution Data.

Abstract: Many real-world phenomena are observed at multiple resolutions. Predictive models designed to predict these phenomena typically consider different resolutions separately. This approach might be limiting in applications where predictions are desired at fine resolutions but available training data is scarce. In this paper, we propose classification algorithms that leverage supervision from coarser resolutions to help train models on finer resolutions. The different resolutions are modeled as different views of the data in a multi-view framework that exploits the complementarity of features across different views to improve models on both views. Unlike traditional multi-view learning problems, the key challenge in our case is that there is no one-to-one correspondence between instances across different views in our case, which requires explicit modeling of the correspondence of instances across resolutions. We propose to use the features of instances at different resolutions to learn the correspondence between instances across resolutions using an attention mechanism.Experiments on the real-world application of mapping urban areas using satellite observations and sentiment classification on text data show the effectiveness of the proposed methods.

Predicting diabetes clinical outcomes using longitudinal risk factor trajectories

Abstract: The ubiquity of electronic health records (EHR) offers an opportunity to observe trajectories of laboratory results and vital signs over long periods of time. This study assessed the value of risk factor trajectories available in the electronic health record to predict incident type 2 diabetes. Analysis was based on a large 13-year retrospective cohort of 71,545 adult, non-diabetic patients with baseline in 2005 and median follow-up time of 8 years. The trajectories of fasting plasma glucose, lipids, BMI and blood pressure were computed over three time frames (2000–2001, 2002–2003, 2004) before baseline. A novel method, Cumulative Exposure (CE), was developed and evaluated using Cox proportional hazards regression to assess risk of incident type 2 diabetes. We used the Framingham Diabetes Risk Scoring (FDRS) Model as control. The new model outperformed the FDRS Model (.802 vs .660; p-values <2e-16). Cumulative exposure measured over different periods showed that even short episodes of hyperglycemia increase the risk of developing diabetes. Returning to normoglycemia moderates the risk, but does not fully eliminate it. The longer an individual maintains glycemic control after a hyperglycemic episode, the lower the subsequent risk of diabetes. Incorporating risk factor trajectories substantially increases the ability of clinical decision support risk models to predict onset of type 2 diabetes and provides information about how risk changes over time.

Physics-Guided Recurrent Graph Networks for Predicting Flow and Temperature in River Networks

Abstract: This paper proposes a physics-guided machine learning approach that combines advanced machine learning models and physics-based models to improve the prediction of water flow and temperature in river networks. We first build a recurrent graph network model to capture the interactions among multiple segments in the river network. Then we present a pre-training technique which transfers knowledge from physics-based models to initialize the machine learning model and learn the physics of streamflow and thermodynamics. We also propose a new loss function that balances the performance over different river segments. We demonstrate the effectiveness of the proposed method in predicting temperature and streamflow in a subset of the Delaware River Basin. In particular, we show that the proposed method brings a 33\%/14\% improvement over the state-of-the-art physics-based model and 24\%/14\% over traditional machine learning models (e.g., Long-Short Term Memory Neural Network) in temperature/streamflow prediction using very sparse (0.1\%) observation data for training. The proposed method has also been shown to produce better performance when generalized to different seasons or river segments with different streamflow ranges.

Regularized graph convolutional networks for short text classification

Abstract: Short text classification is a fundamental problem in natural language processing, social network analysis, and e-commerce. The lack of structure in short text sequences limits the success of popular NLP methods based on deep learning. Simpler methods that rely on bag-of-words representations tend to perform on par with complex deep learning methods. To tackle the limitations of textual features in short text, we propose a Graph-regularized Graph Convolution Network (GR-GCN), which augments graph convolution networks by incorporating label dependencies in the output space. Our model achieves state-of-the-art results on both proprietary and external datasets, outperforming several baseline methods by up to 6% . Furthermore, we show that compared to baseline methods, GR-GCN is more robust to noise in textual features.

Mining Novel Multivariate Relationships in Time Series Data Using Correlation Networks

Abstract: In many domains, there is significant interest in capturing novel relationships between time series that represent activities recorded at different nodes of a highly complex system. In this paper, we introduce multipoles, a novel class of linear relationships between more than two time series. A multipole is a set of time series that have strong linear dependence among themselves, with the requirement that each time series makes a significant contribution to the linear dependence. We demonstrate that most interesting multipoles can be identified as cliques of negative correlations in a correlation network. Such cliques are typically rare in a real-world correlation network, which allows us to find almost all multipoles efficiently using a clique-enumeration approach. Using our proposed framework, we demonstrate the utility of multipoles in discovering new physical phenomena in two scientific domains: climate science and neuroscience. In particular, we discovered several multipole relationships that are reproducible in multiple other independent datasets and lead to novel domain insights.

Teaching deep learning causal effects improves predictive performance.

Abstract: Causal inference is a powerful statistical methodology for explanatory analysis and individualized treatment effect (ITE) estimation, a prominent causal inference task that has become a fundamental research problem. ITE estimation, when performed naively, tends to produce biased estimates. To obtain unbiased estimates, counterfactual information is needed, which is not directly observable from data. Based on mature domain knowledge, reliable traditional methods to estimate ITE exist. In recent years, neural networks have been widely used in clinical studies. Specifically, recurrent neural networks (RNN) have been applied to temporal Electronic Health Records (EHR) data analysis. However, RNNs are not guaranteed to automatically discover causal knowledge, correctly estimate counterfactual information, and thus correctly estimate the ITE. This lack of correct ITE estimates can hinder the performance of the model. In this work we study whether RNNs can be guided to correctly incorporate ITE-related knowledge and whether this improves predictive performance. Specifically, we first describe a Causal-Temporal Structure for temporal EHR data; then based on this structure, we estimate sequential ITE along the timeline, using sequential Propensity Score Matching (PSM); and finally, we propose a knowledge-guided neural network methodology to incorporate estimated ITE. We demonstrate on real-world and synthetic data (where the actual ITEs are known) that the proposed methodology can significantly improve the prediction performance of RNN.

Unlocking Inverse Problems Using Deep Learning: Breaking Symmetries in Phase Retrieval

Abstract: In many physical systems, inputs related by intrinsic system symmetries generate the same output. So when inverting such systems, an input is mapped to multiple symmetry-related outputs. This causes fundamental difficulties for tackling these inverse problems by the emerging end-to-end deep learning approach. Taking phase retrieval as an illustrative example, we show that careful symmetry breaking on the training data can help get rid of the difficulties and significantly improve learning performance in real data experiments. We also extract and highlight the underlying mathematical principle of the proposed solution, which is directly applicable to other inverse problems.

Deep Learning Initialized Phase Retrieval

Abstract: Phase retrieval (PR) consists of estimating 2D or 3D objects from their Fourier magnitudes and takes a central place in scientific imaging. At present, most iterative methods for PR work well only when the initialization is close enough to the solution and fail otherwise. But there has been no general way of obtaining desired initialization. In this paper, we show that a carefully designed deep learning pipeline can consistently generate reliable initialization, so that the subsequent iterative methods can solve the PR problem and produce high-quality solutions. Technically, PR is an inverse problem containing three forward symmetries, and naive deployment of end-to-end deep learning for PR yields poor initialization. We explain why the symmetries cause the learning difficulty and propose a novel strategy that substantially improves the estimation. Overall, the proposed method solves PR in regimes not accessible by the previous methods, and our work synergizes deep learning and iterative methods to solve a difficult scientific inverse problem.

Global River Monitoring Using Semantic Fusion Networks

Abstract: Global river monitoring is an important mission within the remote sensing society. One of the main challenges faced by this mission is generating an accurate water mask from remote sensing images (RSI) of rivers (RSIR), especially on a global scale with various river features. Aiming at better water area classification using semantic information, this paper presents a segmentation method for global river monitoring based on semantic clustering and semantic fusion. Firstly, an encoder–decoder network (AEN)-based architecture is proposed to obtain the semantic features from RSIR. Secondly, a clustering-based semantic fusion method is proposed to divide semantic features of RSIR into groups and train convolutional neural networks (CNN) models corresponding to each group using data augmentation and semi-supervised learning. Thirdly, a semantic distance-based segmentation fusion method is proposed for fusing the CNN models result into final segmentation mask. We built a global river dataset that contains multiple river segments from each continent of the world based on Sentinel-2 satellite imagery. The result shows that the F1-score of the proposed segmentation method is 93.32%, which outperforms several state-of-the-art algorithms, and demonstrates that grouping semantic information helps better segment the RSIR in global scale.

Semi-supervised Classification using Attention-based Regularization on Coarse-resolution Data.

Abstract: Many real-world phenomena are observed at multiple resolutions. Predictive models designed to predict these phenomena typically consider different resolutions separately. This approach might be limiting in applications where predictions are desired at fine resolutions but available training data is scarce. In this paper, we propose classification algorithms that leverage supervision from coarser resolutions to help train models on finer resolutions. The different resolutions are modeled as different views of the data in a multi-view framework that exploits the complementarity of features across different views to improve models on both views. Unlike traditional multi-view learning problems, the key challenge in our case is that there is no one-to-one correspondence between instances across different views in our case, which requires explicit modeling of the correspondence of instances across resolutions. We propose to use the features of instances at different resolutions to learn the correspondence between instances across resolutions using an attention mechanism.Experiments on the real-world application of mapping urban areas using satellite observations and sentiment classification on text data show the effectiveness of the proposed methods.

Model-agnostic Methods for Text Classification with Inherent Noise

Abstract: Text classification is a fundamental problem, and recently, deep neural networks (DNN) have shown promising results in many natural language tasks. However, their human-level performance relies on high-quality annotations, which are time-consuming and expensive to collect. As we move towards large inexpensive datasets, the inherent label noise degrades the generalization of DNN. While most machine learning literature focuses on building complex networks to handle noise, in this work, we evaluate model-agnostic methods to handle inherent noise in large scale text classification that can be easily incorporated into existing machine learning workflows with minimal interruption. Specifically, we conduct a point-by-point comparative study between several noise-robust methods on three datasets encompassing three popular classification models. To our knowledge, this is the first time such a comprehensive study in text classification encircling popular models and model-agnostic loss methods has been conducted. In this study, we describe our learning and demonstrate the application of our approach, which outperformed baselines by up to 10 % in classification accuracy while requiring no network modifications.

BHi-Cect: a top-down algorithm for identifying the multi-scale hierarchical structure of chromosomes.

Abstract: High-throughput chromosome conformation capture (Hi-C) technology enables the investigation of genome-wide interactions among chromosome loci. Current algorithms focus on topologically associating domains (TADs), that are contiguous clusters along the genome coordinate, to describe the hierarchical structure of chromosomes. However, high resolution Hi-C displays a variety of interaction patterns beyond what current TAD detection methods can capture. Here, we present BHi-Cect, a novel top-down algorithm that finds clusters by considering every locus with no assumption of genomic contiguity using spectral clustering. Our results reveal that the hierarchical structure of chromosome is organized as 'enclaves', which are complex interwoven clusters at both local and global scales. We show that the nesting of local clusters within global clusters characterizing enclaves, is associated with the epigenomic activity found on the underlying DNA. Furthermore, we show that the hierarchical nesting that links different enclaves integrates their respective function. BHi-Cect provides means to uncover the general principles guiding chromatin architecture.

Learning with Small Data

Abstract: In the era of big data, data-driven methods have become increasingly popular in various applications, such as image recognition, traffic signal control, fake news detection. The superior performance of these data-driven approaches relies on large-scale labeled training data, which are probably inaccessible in real-world applications, i.e., "small (labeled) data" challenge. Examples include predicting emergent events in a city, detecting emerging fake news, and forecasting the progression of conditions for rare diseases. In most scenarios, people care about these small data cases most and thus improving the learning effectiveness of machine learning algorithms with small labeled data has been a popular research topic. In this tutorial, we will review the trending state-of-the-art machine learning techniques for learning with small (labeled) data. These techniques are organized from two aspects: (1) providing a comprehensive review of recent studies about knowledge generalization, transfer, and sharing, where transfer learning, multi-task learning, and meta-learning are discussed. Particularly, we will focus more on meta-learning, which improves the model generalization ability and has been proven to be an effective approach recently; (2) introducing the cutting-edge techniques which focus on incorporating domain knowledge into machine learning models. Different from model-based knowledge transfer techniques, in real-world applications, domain knowledge (e.g., physical laws) provides us with a new angle to deal with the small data challenge. Specifically, domain knowledge can be used to optimize learning strategies and/or guide the model design. In data mining field, we believe that learning with small data is a trending topic with important social impact, which will attract both researchers and practitioners from academia and industry.

End to End learning for Phase Retrieval

Abstract: We consider the end-to-end deep learning approach for phase retrieval, a central problem in scientific imaging. We highlight a fundamental difficulty for learning that previous work has neglected, likely due to the biased datasets they use for training and evaluation. We propose a simple yet different formulation for PR that seems to overcome the difficulty and return consistently better qualitative results.

Automated Plantation Mapping in Southeast Asia Using MODIS Data and Imperfect Visual Annotations

Abstract: Expansion of large-scale tree plantations for commodity crop and timber production is a leading cause of tropical deforestation. While automated detection of plantations across large spatial scales and with high temporal resolution is critical to inform policies to reduce deforestation, such mapping is technically challenging. Thus, most available plantation maps rely on visual inspection of imagery, and many of them are limited to small areas for specific years. Here, we present an automated approach, which we call Plantation Analysis by Learning from Multiple Classes (PALM), for mapping plantations on an annual basis using satellite remote sensing data. Due to the heterogeneity of land cover classes, PALM utilizes ensemble learning to simultaneously incorporate training samples from multiple land cover classes over different years. After the ensemble learning, we further improve the performance by post-processing using a Hidden Markov Model. We implement the proposed automated approach using MODIS data in Sumatra and Indonesian Borneo (Kalimantan). To validate the classification, we compare plantations detected using our approach with existing datasets developed through visual interpretation. Based on random sampling and comparison with high-resolution images, the user’s accuracy and producer’s accuracy of our generated map are around 85% and 80% in our study region.

Analysis of tri-cum biserial bulk queue model connected with a common server

Abstract: The present paper demonstrates the tri-cum biserial bulk queue model linked with a common server with fixed batch size. The development of the model has been done in the steady-state condition. The arrival and servicing patterns of the customers are postulated to follow the Poisson law. Various queuing model performances have been assessed by using the probability generating function technique and other statistical tools. The broad parametric examination has been documented to show the adequacy of the current arrangement procedure.

Opportunistic offlining for faulty devices in datacenters

Abstract: Embodiments relate to determining whether to take a resource distribution unit (RDU) of a datacenter offline when the RDU becomes faulty. RDUs in a cloud or datacenter supply a resource such as power, network connectivity, and the like to respective sets of hosts that provide computing resources to tenant units such as virtual machines (VMs). When an RDU becomes faulty some of the hosts that it supplies may continue to function and others may become unavailable for various reasons. This can make a decision of whether to take the RDU offline for repair difficult, since in some situations countervailing requirements of the datacenter may be at odds. To decide whether to take an RDU offline, the potential impact on availability of tenant VMs, unused capacity of the datacenter, a number or ratio of unavailable hosts on the RDU, and other factors may be considered to make a balanced decision.

Process Guided Deep Learning for Modeling Physical Systems: An Application in Lake Temperature Modeling

Abstract: In this paper, we introduce a new paradigm that combines scientific knowledge within process-based models and machine learning models to advance scientific discovery in many physical systems. We will describe how to incorporate physical knowledge in real-world dynamical systems as additional constraints for training machine learning models and how to leverage the hidden knowledge encoded by existing process-based models. We evaluate this approach on modeling lake water temperature and demonstrate its superior performance using limited training data and the improved generalizability to different scenarios.