3,613 research outputs found
Learning value functions from undirected state-only experience
No full textSubmission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2023-04-12 without embargo termsThe student, Arjun Gupta, accepted the attached license on 2022-12-01 at 20:24.The student, Arjun Gupta, submitted this Thesis for approval on 2022-12-01 at 20:31.This Thesis was approved for publication on 2022-12-06 at 08:49.DSpace SAF Submission Ingestion Package generated from Vireo submission #18729 on 2023-04-12 at 07:37:03This thesis tackles the problem of learning value functions from undirected state-only experience (state transitions without action labels i.e. (s,s′,r) tuples). We first theoretically characterize the applicability of Q-learning in this setting. We show that tabular Q-learning in discrete Markov decision processes (MDPs) learns the same value function under any arbitrary refinement of the action space. This theoretical result motivates the design of Latent Action Q-learning or LAQ, an offline RL method that can learn effective value functions from state-only experience. Latent Action Q-learning (LAQ) learns value functions using Q-learning on discrete latent actions obtained through a latent-variable future prediction model. We show that LAQ can recover value functions that have high correlation with value functions learned using ground truth actions. Value functions learned using LAQ lead to sample efficient acquisition of goal-directed behavior, can be used with domain-specific low-level controllers, and facilitate transfer across embodiments. Our experiments in 5 environments ranging from 2D grid world to 3D visual navigation in realistic environments demonstrate the benefits of LAQ over simpler alternatives, imitation learning oracles, and competing methods
Policy and guidance on the use of generative artificial intelligence in UK higher education
No full textThis paper describes the discussions, contentions, and arguments for and against the use of generative artificial intelligence (GenAI) in Higher Education (HE). It includes an overview of artificial intelligence, synthesises the guidelines and policies on the use of GenAI, and investigates the opportunities to optimise the application of GenAI tools in HE. We suggest that while there is uncertainty surrounding such tools it is important to look at their potential utility and efficacy
Generative AI in UK Higher Education: contentions and considerations for policy and practice
No full text
Deep learning for cardiologist-level myocardial infarction detection in electrocadiograms
No full textHeart disease is the leading cause of death worldwide. Among patients with cardiovascular
diseases, myocardial infarction is the main cause of death. In order to provide adequate healthcare
support to patients who may experience this clinical event, it is essential to gather supportive
evidence in a timely manner to help secure a correct diagnosis. In this thesis we design domain-inspired
neural network models, trained, tested and validated with the Physikalisch-Technische
Bundesanstalt (PTB) data set, to conduct a series of studies. First, acknowledging that the
identification of suggestive electrocardiographic (ECG) changes may help in the classification of
heart conditions, we adapt the ConvNetQuake neural network model---originally designed to identify
earthquakes---to train, validate and test neural network models that take in from one to several ECG
leads. This systematic analysis, first of its kind in the literature, indicates that out of 15 ECG leads,
data from the v6, vz, and ii leads are critical to correctly identify myocardial infarction. Second, we
show that using two independent train-validation-test data splits, namely, record-wise and patient-wise,
does not change the finding that the combination of the leads v6, vz, and ii provides the best
classification results for myocardial infarction, achieving 99.43% classification accuracy on a record-wise
split, and 97.83% classification accuracy on a patient-wise split. These two results represent
cardiologist-level performance for myocardial infarction detection after feeding only 10 seconds of raw ECG data into our multi-ECG-channel (v6-vz-ii) neural network model. Third, we show that our
multi-ECG-channel neural network model achieves cardiologist-level performance without the need
of any kind of manual feature extraction or data pre-processing.U of I Only Restriction set for Item 114826 on 2020-06-11T17:34:59Z with date by [email protected] by Janice Progen ([email protected]) on 2020-06-11T17:57:31Z
No. of bitstreams: 1
SP20-ECE499-Thesis-Gupta, Arjun.pdf: 1439863 bytes, checksum: 072f6462435f5e93fac6b7c32ea31783 (MD5)U of I Only Restriction set for Item 114826 on 2020-06-11T18:55:42Z with date by [email protected] for entry into archive by James Hutchinson ([email protected]) on 2020-06-11T18:56:21Z (GMT) No. of bitstreams: 1
SP20-ECE499-Thesis-Gupta, Arjun.pdf: 1439863 bytes, checksum: 072f6462435f5e93fac6b7c32ea31783 (MD5)Made available in DSpace on 2020-06-11T18:56:21Z (GMT). No. of bitstreams: 1
SP20-ECE499-Thesis-Gupta, Arjun.pdf: 1439863 bytes, checksum: 072f6462435f5e93fac6b7c32ea31783 (MD5)
Previous issue date: 2020-05Embargo set by: James Hutchinson for item 114826
Lift date: 10000-01-01
Reason: Undergraduate senior thesis not recommended for open accessUndergraduate senior thesis not recommended for open accessU of I Onl
Rebellion in a ‘Little-Known’ district of the Empire: 1857 and the Hos of Singhbhum
No full textThis paper focuses on the nature of rebellion of 1857 in the Singhbhum district of Chotanagpur, then considered a relatively unimportant region of the british Empire in India, chiefly inhabited by the Ho tribe. In particular, it argues that niether the nationalist nor the subaltern opinions of various scholars writing on 1857 have taken into account the specific situation prevailing in Singhbhum were the rebellion was a multi-layered movement with the raja Arjun Singh of Porahat, the sepoys and the people coming together in an anti-British uprising
Data for Gupta et al., "Estimating the Meridional Extent of Adiabatic Mixing in the Stratosphere using Age-of-Air", JGR:Atmospheres,
No full textModel data and post-processed data supporting the creation of the manuscript "Estimating the Meridional Extent of Adiabatic Mixing in the Stratosphere using Age-of-Air" submitted to JGR:Atmospheres in August 2022.
1) The netCDF files created through post-processing of full model data in FORTRAN are shared in the /data/ directory. These file contains the zonal mean circulation statistics based on Gupta et al. (2020), age-of-air transport diagnostics based on Linz et al. (2021), and the novel \Gamma-\Theta circulation streamfunction introduced in this study. The /data/ directory also contains MATLAB .mat data files for the transport diagnostics obtained from WACCM. 150 days of actual GFDL-FV3 model data in the northern hemisphere, between 0.1 hPa-500 hPa pressure levels is also provided to support external computations and validation.
2) The Jupyter notebook used for final computation and figures production is provided in .ipynb, .html and .pdf formats in /code/. All the files referred to in the notebook are stored in the /data/ directory.
Corresponding author : Aman Gupta, [email protected], [email protected], [email protected]
Polyaniline/Fe-0 composite nanofibers: An excellent adsorbent for the removal of arsenic from aqueous solutions
No full textA novel nano-scale polyaniline/Fe-0 composite (termed as PANI/Fe-0) was synthesized via reductive deposition of nano-Fe-0 onto the PANI nanofibers matrix at room temperature. PANI/Fe-0 was characterized by SEM, TEM, XRD, FTIR and XPS. PANI/Fe-0 was tested for arsenic removal in batch experiments. The effects of initial pH value, adsorbent mass loading, As initial concentration, and co-solutes on As removal were explored. Results showed nano-Fe-0 particles with average diameter 6.4 nm embedded onto the PANI matrix. As removal was rapid for both As(III) and As(V) species. The equilibrium isotherm data fitted well with Langmuir isotherm model. The maximum adsorption capacities at pH 7.0 for As(III) and As(V) were 232.5 and 227.3 mg/g, respectively. These capacities are significantly higher than most of the reported adsorbents. The co-existing HCO3-, SiO32-, and SO42- ions had no significant effect on As removal. However, the presence of NO3- and PO43- significantly reduced the As removal efficiencies. Achieved results further show that PANI/Fe-0 is a promising adsorbent for safe drinking provision in low-income communities owing to its excellent performance, simple and low-cost synthesis process. (C) 2015 Elsevier B.V. All rights reserved
Review: Evidence-Based Management of Acute Heart Failure
No full textAcute heart failure (AHF) is a complex, heterogeneous clinical syndrome with high morbidity and mortality, incurring significant healthcare costs. Patients transition from home to the emergency department, the hospital and home again, and require decisions surrounding diagnosis, treatment and prognosis at each step of the way. The purpose of this review is to examine the epidemiology, etiologies and classifications of AHF, and specifically focus on practical information relevant to the clinician. We examine the mechanisms of decompensation relevant to clinical presentations, including precipitating factors, neuroendocrine interactions and inflammation, along with how consideration of these factors these may help select therapies for an individual patient. The prevalence and significance of end-organ manifestations like renal, gastrointestinal, respiratory and neurologic manifestations are discussed. We also highlight how the development of renal dysfunction relates to the choice of a variety of diuretics that may be useful in specific circumstances and review guideline-directed medical therapy. We discuss the practical use (and pitfalls) of a variety of evidence-based clinical scoring criteria available to risk stratify patients with AHF. Finally, evidence-based management of AHF is discussed, including both pharmacologic and nonpharmacologic therapies, including the lack of evidence for using old and new vasodilators and the recent evidence regarding initiation of newer therapies in hospital. Overall, we suggest that clinicians consider implementing the newer data in AHF and subject existing practice patterns and treatments to the same rigor as new therapies
Corrigendum: Capital Inflows and House Prices: Aggregate and Regional Evidence from China
No full textIn the paper ‘Capital Inflows and House Prices: Aggregate and Regional Evidence from China’ by H. An, et al., printed in the December 2016 issue, there was a missing acknowledgement section for funding resources.
On page 451, the acknowledgement section should appear after the corresponding information as:
“Correspondence: Rakesh Gupta, Department of Accounting, Finance and Economics, Griffith Business School, Griffith University, Nathan Campus QLD 4111. [email protected]
*This work was financially supported by the Humanities and Social Science Foundation of Ministry of Education of China (16YJA790001).”
The author apologises for this error and any confusion it may have caused.No Full Tex
- …
