Filters








706 Hits in 1.7 sec

Imitation by Predicting Observations [article]

Andrew Jaegle, Yury Sulsky, Arun Ahuja, Jake Bruce, Rob Fergus, Greg Wayne
2021 arXiv   pre-print
Imitation learning enables agents to reuse and adapt the hard-won expertise of others, offering a solution to several key challenges in learning behavior. Although it is easy to observe behavior in the real-world, the underlying actions may not be accessible. We present a new method for imitation solely from observations that achieves comparable performance to experts on challenging continuous control tasks while also exhibiting robustness in the presence of observations unrelated to the task.
more » ... ur method, which we call FORM (for "Future Observation Reward Model") is derived from an inverse RL objective and imitates using a model of expert behavior learned by generative modelling of the expert's observations, without needing ground truth actions. We show that FORM performs comparably to a strong baseline IRL method (GAIL) on the DeepMind Control Suite benchmark, while outperforming GAIL in the presence of task-irrelevant features.
arXiv:2107.03851v1 fatcat:vfo2yknbuzatpcxrc4zus3ewiu

Semi-supervised Calibration of Risk with Noisy Event Times (SCORNET) Using Electronic Health Record Data [article]

Yuri Ahuja, Liang Liang, Sicong Huang, Tianxi Cai
2021 bioRxiv   pre-print
others, 2017; Liang and others, 2016; Zhao and others, 2014; Uno and others, 2018; Hassett and others, 2017; Chubak and others, 2012; Choi and others, 2015; Kaji and others, 2019; Ruan and others, 2019; Ahuja  ...  leverages the features used in MAP as well as 121 additional manually-selected EHR features including counts of relevant medications, ICD codes, and concept unique identifiers (CUIs) in clinical notes (Ahuja  ... 
doi:10.1101/2021.01.08.425976 fatcat:7j7vkvbbfzaerp6o5wmjeyo5ri

SAMGEP: A Novel Method for Prediction of Phenotype Event Times Using the Electronic Health Record [article]

Yuri Ahuja, Chuan Hong, Zongqi Xia, Tianxi Cai
2021 medRxiv   pre-print
Objective: While there exist numerous methods to predict binary phenotypes using electronic health record (EHR) data, few exist for prediction of phenotype event times, or equivalently phenotype state progression. Estimating such quantities could enable more powerful use of EHR data for temporal analyses such as survival and disease progression. We propose Semi-supervised Adaptive Markov Gaussian Embedding Process (SAMGEP), a semi-supervised machine learning algorithm to predict phenotype event
more » ... times using EHR data. Methods: SAMGEP broadly consists of four steps: (i) assemble time-evolving EHR features predictive of the target phenotype event, (ii) optimize weights for combining raw features and feature embeddings into dense patient-timepoint embeddings, (iii) fit supervised and semi-supervised Markov Gaussian Process models to this embedding progression to predict marginal phenotype probabilities at each timepoint, and (iv) take a weighted average of these supervised and semi-supervised predictions. SAMGEP models latent phenotype states as a binary Markov process, conditional on which patient- timepoint embeddings are assumed to follow a Gaussian Process. Results: SAMGEP achieves significantly improved AUCs and F1 scores relative to common machine learning approaches in both simulations and a real-world task using EHR data to predict multiple sclerosis relapse. It is particularly adept at predicting a patient's longitudinal phenotype course, which can be used to estimate population-level cumulative probability and count process estimators. Reassuringly, it is robust to a variety of generative model parameters. Discussion: SAMGEP's event time predictions can be used to estimate accurate phenotype progression curves for use in downstream temporal analyses, such as a survival study for comparative effectiveness research.
doi:10.1101/2021.03.07.21253096 fatcat:ahl3si5y65fajjtkyw6rqtqa5a

sureLDA: A Multi-Disease Automated Phenotyping Method for the Electronic Health Record [article]

Yuri Ahuja, Doudou Zhou, Zeling He, Jiehuan Sun, Victor M Castro, Vivian Gainer, Shawn N Murphy, Chuan Hong, Tianxi Cai
2020 bioRxiv   pre-print
Objective: A major bottleneck hindering utilization of electronic health record (EHR) data for translational research is the lack of precise phenotype labels. Chart review as well as rule-based and supervised phenotyping approaches require laborious expert input, hampering applicability to studies that require many phenotypes to be defined and labeled de novo. Though ICD codes are often used as surrogates for true labels in this setting, these sometimes suffer from poor specificity. We propose
more » ... fully automated topic modeling algorithm to simultaneously annotate multiple phenotypes. Methods: sureLDA is a label-free multidimensional phenotyping method. It first uses the PheNorm algorithm to initialize probabilities based on two surrogate features for each target phenotype, and then leverages these probabilities to constrain the Latent Dirichlet Allocation (LDA) topic model to generate phenotype-specific topics. Finally, it combines phenotype-feature counts with surrogates via clustering ensemble to yield final phenotype probabilities. Results: sureLDA achieves reliably high accuracy and precision across a range of simulated and real-world phenotypes. Its performance is robust to phenotype prevalence and relative informativeness of surogate versus non-surrogate features. It also exhibits powerful feature selection properties. Discussion: sureLDA combines attractive properties of PheNorm and LDA to achieve high accuracy and precision robust to diverse phenotype characteristics. It offers particular improvement for phenotypes insufficiently captured by a few surrogate features. Moreover, sureLDAs feature selection ability enables it to handle high feature dimensions and produce interpretable computational phenotypes. Conclusion: sureLDA is well suited toward large-scale EHR phenotyping for highly multi-phenotype applications such as PheWAS.
doi:10.1101/2020.04.13.038968 fatcat:bgk56aafpbfyjfyvgkoy54gn4u

Influence of Socioeconomic Status Trajectories on Innate Immune Responsiveness in Children

Meghan B. Azad, Yuri Lissitsyn, Gregory E. Miller, Allan B. Becker, Kent T. HayGlass, Anita L. Kozyrskyj, Sunil K. Ahuja
2012 PLoS ONE  
Objectives: Lower socioeconomic status (SES) is consistently associated with poor health, yet little is known about the biological mechanisms underlying this inequality. In children, we examined the impact of early-life SES trajectories on the intensity of global innate immune activation, recognizing that excessive activation can be a precursor to inflammation and chronic disease. Methods: Stimulated interleukin-6 production, a measure of immune responsiveness, was analyzed ex vivo for 267
more » ... ian schoolchildren from a 1995 birth cohort in Manitoba, Canada. Childhood SES trajectories were determined from parentreported housing data using a longitudinal latent-class modeling technique. Multivariate regression was conducted with adjustment for potential confounders. Results: SES was inversely associated with innate immune responsiveness (p = 0.003), with persistently low-SES children exhibiting responses more than twice as intense as their high-SES counterparts. Despite initially lower SES, responses from children experiencing increasing SES trajectories throughout childhood were indistinguishable from high-SES children. Low-SES effects were strongest among overweight children (p,0.01). Independent of SES trajectories, immune responsiveness was increased in First Nations children (p,0.05) and urban children with atopic asthma (p,0.01). Conclusions: These results implicate differential immune activation in the association between SES and clinical outcomes, and broadly imply that SES interventions during childhood could limit or reverse the damaging biological effects of exposure to poverty during the preschool years.
doi:10.1371/journal.pone.0038669 pmid:22685596 pmcid:PMC3369855 fatcat:35xisvz2b5ab7icdz3cnrrifxe

Diagnosis of feline acute intermittent porphyria presenting with erythrodontia requires molecular analyses

Sonia Clavero, Yuri Ahuja, David F. Bishop, Brittany Kwait, Mark E. Haskins, Urs Giger, Robert J. Desnick
2013 The Veterinary Journal  
Erythrodontia is the hallmark of human congenital erythropoietic porphyria (CEP), but is also a major phenotypic feature of acute intermittent porphyria (AIP) in cats. In this study, detailed biochemical and molecular analyses were performed on two unrelated cats with autosomal dominant AIP that presented with erythrodontia, yellow-brown urine and mild changes in erythrocytes. The cats had elevated concentrations of urinary 5-aminolevulinic acid and porphobilinogen, and half normal erythrocytic
more » ... hydroxymethylbilane synthase (HMBS) activity. Two novel HMBS mutations were detected; one cat had a deletion (c.107_110delACAG) and one cat had a splicing alteration (c.826-1G>A), both leading to premature stop codons and truncated proteins (p.D36Vfs*6 and p.L276Efs*6, respectively). These studies highlight the importance of appropriate biochemical and molecular genetic analyses for the accurate diagnoses of porphyrias in cats and extend the molecular genetic heterogeneity of feline AIP. Thus, although erythrodontia is a classic sign of congenital erythropoietic porphyria in human beings, cats with erythrodontia may have acute intermittent porphyria, a hepatic porphyria. The porphyrias are inborn errors of metabolism resulting from deficient activities of specific enzymes in the heme biosynthetic pathway. Acute hepatic porphyrias in human beings include autosomal dominant acute intermittent porphyria (AIP), hereditary coproporphyria (HCP), variegate porphyria (VP) and autosomal recessive 5-aminolevulinate dehydratase deficient porphyria (ADP). In these conditions, plasma and urinary concentrations of the porphyrin precursors 5-aminolevulinic acid (ALA) and porphobilinogen (PBG) are markedly increased during sudden, life-threatening neurological attacks. In contrast, human
doi:10.1016/j.tvjl.2013.10.008 pmid:24239138 pmcid:PMC3963809 fatcat:eoold74a6bc27ehjlvdiijftl4

Leveraging electronic health records data to predict multiple sclerosis disease activity

Yuri Ahuja, Nicole Kim, Liang Liang, Tianrun Cai, Kumar Dahal, Thany Seyok, Chen Lin, Sean Finan, Katherine Liao, Guergana Savovoa, Tanuja Chitnis, Tianxi Cai (+1 others)
2021 Annals of Clinical and Translational Neurology  
Ahuja et al.  ... 
doi:10.1002/acn3.51324 pmid:33626237 pmcid:PMC8045951 fatcat:x65ty5wepzb6nd3azihov5emr4

Adsorption separation of heavier isotope gases in subnanometer carbon pores

Sanjeev Kumar Ujjain, Abhishek Bagusetty, Yuki Matsuda, Hideki Tanaka, Preety Ahuja, Carla de Tomas, Motomu Sakai, Fernando Vallejos-Burgos, Ryusuke Futamura, Irene Suarez-Martinez, Masahiko Matsukata, Akio Kodama (+4 others)
2021 Nature Communications  
AbstractIsotopes of heavier gases including carbon (13C/14C), nitrogen (13N), and oxygen (18O) are highly important because they can be substituted for naturally occurring atoms without significantly perturbing the biochemical properties of the radiolabelled parent molecules. These labelled molecules are employed in clinical radiopharmaceuticals, in studies of brain disease and as imaging probes for advanced medical imaging techniques such as positron-emission tomography (PET). Established
more » ... llation-based isotope gas separation methods have a separation factor (S) below 1.05 and incur very high operating costs due to high energy consumption and long processing times, highlighting the need for new separation technologies. Here, we show a rapid and highly selective adsorption-based separation of 18O2 from 16O2 with S above 60 using nanoporous adsorbents operating near the boiling point of methane (112 K), which is accessible through cryogenic liquefied-natural-gas technology. A collective-nuclear-quantum effect difference between the ordered 18O2 and 16O2 molecular assemblies confined in subnanometer pores can explain the observed equilibrium separation and is applicable to other isotopic gases.
doi:10.1038/s41467-020-20744-6 pmid:33483513 fatcat:gletlyel4vaxjm7ame3mgg5tom

Inferring multimodal latent topics from electronic health records

Yue Li, Pratheeksha Nair, Xing Han Lu, Zhi Wen, Yuening Wang, Amir Ardalan Kalantari Dehaghi, Yan Miao, Weiqi Liu, Tamas Ordog, Joanna M. Biernacka, Euijung Ryu, Janet E. Olson (+7 others)
2020 Nature Communications  
Electronic health records (EHR) are rich heterogeneous collections of patient health information, whose broad adoption provides clinicians and researchers unprecedented opportunities for health informatics, disease-risk prediction, actionable clinical recommendations, and precision medicine. However, EHRs present several modeling challenges, including highly sparse data matrices, noisy irregular clinical notes, arbitrary biases in billing code assignment, diagnosis-driven lab tests, and
more » ... neous data types. To address these challenges, we present MixEHR, a multi-view Bayesian topic model. We demonstrate MixEHR on MIMIC-III, Mayo Clinic Bipolar Disorder, and Quebec Congenital Heart Disease EHR datasets. Qualitatively, MixEHR disease topics reveal meaningful combinations of clinical features across heterogeneous data types. Quantitatively, we observe superior prediction accuracy of diagnostic codes and lab test imputations compared to the state-of-art methods. We leverage the inferred patient topic mixtures to classify target diseases and predict mortality of patients in critical conditions. In all comparison, MixEHR confers competitive performance and reveals meaningful disease-related topics.
doi:10.1038/s41467-020-16378-3 pmid:32439869 fatcat:jhbngeavxnh5zjtm7h44jko4o4

Association of Interleukin 6 Receptor Variant With Cardiovascular Disease Effects of Interleukin 6 Receptor Blocking Therapy

Tianxi Cai, Yichi Zhang, Yuk-Lam Ho, Nicholas Link, Jiehuan Sun, Jie Huang, Tianrun A. Cai, Scott Damrauer, Yuri Ahuja, Jacqueline Honerlaw, Jie Huang, Lauren Costa (+10 others)
2018 JAMA cardiology  
Cai, Ahuja, Gaziano, Cho, O'Donnell, Liao); Brigham and Women's Hospital, Boston, Massachusetts Atherosclerosis of native arteries of the extremities with intermittent claudication log 10 P Value  ... 
doi:10.1001/jamacardio.2018.2287 pmid:30090940 pmcid:PMC6233652 fatcat:ykkx3xwg7beqhd7bog2n6cu2yq

Page 6695 of Mathematical Reviews Vol. , Issue 2004h [page]

2004 Mathematical Reviews  
E. .. 62047 Ahmed, Zafar 82040 Ahn, Inkyung 35054 Ahn, Min-Ho ...... 41002 Ahuja, O.  ...  PEDRAIMNVIOR Weds. 5 cv2 co Seaakencus See Abramovich, Yuri A. Abramovich, Yuri A. 47055 Abrams, A 60106 \bramsky, Samson 68076 Abrosimoy, M.  ... 

Page 1443 of Mathematical Reviews Vol. , Issue 2001B [page]

2001 Mathematical Reviews  
M. ... 47016,47056 Arlinskii, Yury See Arlinskii, Yu. M. Armanious, M.  ...  , Mangho 11002 Ahuja, Ravindra K 90072 Aichholzer, Oswin —e Aikawa, Hiroaki 31003-31004 Aiki, Toyohiko ................ 35305 REE, By «020020: 68093 Aizawa, N. .... 20085 Aizicovici, S. ........ ... 34110  ... 

Page 4551 of Mathematical Reviews Vol. , Issue 2000f [page]

2000 Mathematical Reviews  
32030 82002 81144 76084 68034 34182 53024 76118 See Abramovich, Yuri A . 47062 Abramovich, Yuri A Abramson, Michael A. Abrashin, V.N. .... Abrashkin, V. A.  ...  Ahuja, Ravindra K. Ainsworth, Mark ............. Ait-Ameur, Yamine Asyeme, Re. ........: Aizicovici, S. .... Akcoglu, Mustafa Akemann, Charles A. Akhmedoy, M.  ... 

Page 4605 of Mathematical Reviews Vol. , Issue 2002F [page]

2002 Mathematical Reviews  
32021 17044 03092 13008 93117 78002 39008 14071 See Abramovich, Yuri A Abramovich, Yuri A Abramson, David Abrashin, V.  ...  Ahn, Seung-Ho Ahuja, Om P. 41015 * 58001 See Aguado, Miguel 81115 46071 11070 37032 81003,81005 47055 11048 See Ahmad, Rais 47132,47137 $4029 See Ahmedov, Hadji 33022 65129 $2022 30012 AuthorIndex 2002f  ... 

Page 10587 of Mathematical Reviews Vol. , Issue 2004m [page]

2004 Mathematical Reviews  
G. 54006-54007 Ahuja, Ravindra K.  ...  See Abramovich, Yuri A. Abramovich, Yuri A. ................ 47079 Abramsky, Samson RUNNER: cc vs Ktnnnscoveeeaddewimeek Absil, P.-A.  ... 
« Previous Showing results 1 — 15 out of 706 results