检索结果-内蒙古大学图书馆

A General Framework for Clustering and Distribution Matching with Bandit Feedback

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Yavas, Recep Can Huang, Yuqi Tan, Vincent Y.F. Scarlett, Jonathan Descartes Program CNRS@CREATE 138602 Singapore Department of Information Systems and Analytics School of Computing National University of Singapore 119077 Singapore Department of Mathematics Department of Electrical and Computer Engineering National University of Singapore 119077 Singapore Department of Computer Science Department of Mathematics Institute of Data Science National University of Singapore 119077 Singapore

We develop a general framework for clustering and distribution matching problems with bandit feedback. We consider a K-armed bandit model where some subset of K arms is partitioned into M groups. Within each group, the random variable associated to each arm follows the same distribution on a finite alphabet. At each time step, the decision maker pulls an arm and observes its outcome from the random variable associated to that arm. Subsequent arm pulls depend on the history of arm pulls and their outcomes. The decision maker has no knowledge of the distributions of the arms or the underlying partitions. The task is to devise an online algorithm to learn the underlying partition of arms with the least number of arm pulls on average and with an error probability not exceeding a pre-determined value δ. Several existing problems fall under our general framework, including finding M pairs of arms, odd arm identification, and N-ary clustering of K arms belong to our general framework. We derive a non-asymptotic lower bound on the average number of arm pulls for any online algorithm with an error probability not exceeding δ. Furthermore, we develop a computationally-efficient online algorithm based on the Track-and-Stop method and Frank–Wolfe algorithm, and show that the average number of arm pulls of our algorithm asymptotically matches that of the lower bound. Our refined analysis also uncovers a novel bound on the speed at which the average number of arm pulls of our algorithm converges to the fundamental limit as δ *** Codes 68T05 Copyright © 2024, The Authors. All rights reserved.

关键词： Clustering algorithms

South jakarta's air quality using PM 2.5 data at the beginning of COVID-19 Restriction 6

学校读者我要写书评

暂无评论

South jakarta's air quality using PM 2.5 data at the beginni...

6th International Conference on Information Management and Technology, ICIMTech 2021

作者： Rahutomo, Reza Purwandari, Kartika Hidayat, Alam Ahmad Pardamean, Bens Bina Nusantara University School of Information Systems Information System Department Jakarta11480 Indonesia Bina Nusantara University Bioinformatics and Data Science Research Center Jakarta11480 Indonesia Bina Nusantara University BINUS Graduate Program - Master of Computer Science Computer Science Department Jakarta11480 Indonesia

ISBN: (纸本)9781665449373

As soon as the lockdown was established in Jakarta, Indonesia, people experienced a change in Jakarta's air quality. This research was conducted to answer several challenges such as: where to find Indonesia's air quality open data platforms, how to handle the acquired data with SQL technology, and how to validate whether the air quality of Jakarta was significantly changed from a period to another or not. With particulate matter 2.5 (PM 2.5) data acquired by The United States Diplomatic Post in South Jakarta, a data visualization was generated to describe the fluctuation of PM 2.5 from February 29th to June 29th. By conducting a two-sample t-test, a significant change in South Jakarta's air quality was revealed. The test concluded that South Jakarta's air quality got worse even after a lockdown was established. © 2021 IEEE.

关键词： COVID-19

Lockdown strategy worth lives: The SEIRD modelling in COVID-19 outbreak in Indonesia 2

学校读者我要写书评

暂无评论

Lockdown strategy worth lives: The SEIRD modelling in COVID-...

2nd International Conference on Biospheric Harmony Advanced Research, ICOBAR 2020

作者： Nurlaila, I. Hidayat, A.A. Pardamean, B. Information Systems Department BINUS Online Learning Bina Nusantara University Jakarta11480 Indonesia Bioinformatics and Data Science Research Center Bina Nusantara University Jakarta11480 Indonesia Computer Science Department BINUS Graduate Program-Master of Computer Science Bina Nusantara University Jakarta11480 Indonesia

Lockdowns, despite their conflicting restrictions and consequences they might offer when enforced as a national strategy, are deemed to be suggestive for a prompt conquer to the Coronavirus Disease-19 (COVID-19) outbreak. There have been some success stories such as in China, South Korea and our ASEAN member fellow, Vietnam, which exhibited extremely fallen numbers of COVID-19 cases post the enforcement of lockdown. Indonesia, however, remains in the crux of dispute whether or not the lockdown is opted to force COVID-19 transmission down under control. We, in this respect, employ the most popular model which has been broadly applied in the field of epidemiology, referred to as SEIRD (Susceptible, Exposed, Infectious, Recovered, and Death), the extension form of an age-structured SEIR, where the Death (D) is included to provide more factual situation. We modify the I (Infectious) fraction as symp-tomatic (Is) and asymptomatic (Ia) infectives. There are three lockdown scenarios simulated in our modified SEIRD with the starting date are 26 April 2020 when the immediate lockdown was enacted-and a-week and two-week-gaps, respectively. We figured out that by stipulating the lockdown without delay (26 April 2020), the new cases could be kept below 10,000. A week delay escalates case number to 5,000 and delay for one more week results in beyond 20,000 new cases. Furthermore, when the intervention is delayed following the delayed lockdown, normalization would demand a longer period within which dealing with more critical and dying patients is unavoidable and should be more anticipated. © Published under licence by IOP Publishing Ltd.

关键词： Coronavirus

HDSS: a healthcare decision support system on combining domain knowledge and data analytics for predicting potential risk of mental health

学校读者我要写书评

暂无评论

International Journal of Applied Decision sciences 2022年第4期15卷 465-491页

作者： Ngan, Chun-Kit Paat, Yok-Fong Green, Rachel Data Science Program Worcester Polytechnic Institute 100 Institute Road WorcesterMA01609 United States Department of Social Work The University of Texas at El Paso 500 West University Avenue El PasoTX79968 United States Diocesan Migrant & Refugee Services Inc. 2400A East Yandell Drive El PasoTX79903 United States

We develop an HDSS to determine if a patient is at risk of mental health (MH) problems. This HDSS combines the strengths of both domain experts’ knowledge and data analytics techniques to conduct the risk identification for patients. Our contributions are three-fold: 1) develop a two-tier hybrid-based prediction process to construct a classification model to predict if a person suffers from a particular symptom;and then determine if that person is at risk of suffering from MH problems based upon the domain experts’ diagnostic criteria;2) execute experiments to show that our approach is more effective and produces the results that are slightly less accurate than those of domain experts’ diagnostic criteria but superior to those of the data analytics approach;3) develop a dashboard that delivers descriptive visuals from diverse perspectives to assist physicians in obtaining a more comprehensive view of patients’ MH status to provide them with better treatment and support. Copyright © 2022 Inderscience Enterprises Ltd.

关键词： Health risks

ZERO-SHOT MODEL-BASED REINFORCEMENT LEARNING USING LARGE LANGUAGE MODELS

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Benechehab, Abdelhakim Hili, Youssef Attia El Odonnat, Ambroise Zekri, Oussama Thomas, Albert Paolo, Giuseppe Filippone, Maurizio Redko, Ievgen Kégl, Balázs Huawei Noah’s Ark Lab Paris France Department of Data Science EURECOM France Inria Univ. Rennes 2 CNRS IRISA France ENS Paris-Saclay France Statistics Program KAUST Saudi Arabia

The emerging zero-shot capabilities of Large Language Models (LLMs) have led to their applications in areas extending well beyond natural language processing tasks. In reinforcement learning, while LLMs have been extensively used in text-based environments, their integration with continuous state spaces remains understudied. In this paper, we investigate how pre-trained LLMs can be leveraged to predict in context the dynamics of continuous Markov decision processes. We identify handling multivariate data and incorporating the control signal as key challenges that limit the potential of LLMs’ deployment in this setup and propose Disentangled In-Context Learning (DICL) to address them. We present proof-of-concept applications in two reinforcement learning settings: model-based policy evaluation and data-augmented off-policy reinforcement learning, supported by theoretical analysis of the proposed methods. Our experiments further demonstrate that our approach produces well-calibrated uncertainty estimates. We release the code at https://***/abenechehab/dicl. © 2024, CC BY.

关键词： Zero-shot learning

Comparative analysis of deep learning models for detecting face mask

学校读者我要写书评

暂无评论

Procedia computer science 2023年 216卷 48-56页

作者： M Vickya Ramadhan Kahlil Muchtar Yudha Nurdin Maulisa Oktiana Maya Fitria Novi Maulina Gregorius Natanael Elwirehardja Bens Pardamean Department of Electrical and Computer Engineering Universitas Syiah Kuala Banda Aceh 23111 Indonesia. Telematics Research Center (TRC) Universitas Syiah Kuala Banda Aceh 23111 Indonesia. Microbiology Department School of Medicine Universitas Syiah Kuala Banda Aceh 23111 Indonesia. Bioinformatics and Data Science Research Center Bina Nusantara University Jakarta 11480 Indonesia. Computer Science Department BINUS Graduate Program - Master of Computer Science Program Bina Nusantara University Jakarta 11480 Indonesia.

The spread of Corona Virus Disease 19 (COVID-19) in Indonesia is still relatively high and has not shown a significant decrease. One of the main reasons is due to the lack of supervision on the implementation of health protocols such as wearing masks in daily activities. Recently, state-of-the-art algorithms were introduced to automate face mask detection. To be more specific, the researchers developed various kinds of architectures for the detection of masks based on computer vision methods. This paper aims to evaluate well-known architectures, namely the ResNet50, VGG11, InceptionV3, EfficientNetB4, and YOLO (You Only Look Once) to recommend the best approach in this specific field. By using the MaskedFace-Net dataset, the experimental results showed that the EfficientNetB4 architecture has better accuracy at 95.77% compared to the YOLOv4 architecture of 93.40%, InceptionV3 of 87.30%, YOLOv3 of 86.35%, ResNet50 of 84.41%, VGG11 of 84.38%, and YOLOv2 of 78.75%, respectively. It should be noted that particularly for YOLO, the model was trained using a collection of MaskedFace-Net images that had been pre-processed and labelled for the task. The model was initially able to train faster with pre-trained weights from the COCO dataset thanks to transfer learning, resulting in a robust set of features expected for face mask detection and classification.

关键词： Binary Classification Deep Learning Face mask detection

Enhancing Localized Surface Plasmon Resonance Response for Albumin Detection by Optimizing the Lateral Size of Hexagonal Gold Nanoparticles

学校读者我要写书评

暂无评论

Enhancing Localized Surface Plasmon Resonance Response for A...

IEEE International Conference on Semiconductor Electronics (ICSE)

作者： Silva Nurfasha Chandra Wulandari Lilik Hasanah Ahmad Aminuddin Yanurita Dwi Hapsari Mohammad Arifin Roer Eka Pawinanto Budi Mulyanti Study Program of Physics Universitas Pendidikan Indonesia Bandung Indonesia Study Program of Physics Institut Teknologi Bandung Bandung Indonesia Department of Physics Faculty of Science and Data Analytics Institut Teknologi Sepuluh Nopember Surabaya Indonesia Study Program of Industrial Automation and Robotics Engineering Education Universitas Pendidikan Indonesia Bandung Indonesia Study Program of Electrical Engineering Universitas Pendidikan Indonesia Bandung Indonesia

ISBN: (数字)9798350378313

ISBN: (纸本)9798350378320

The kidney is a human organ with numerous vital functions, mainly for blood filtration. Damage to the kidneys can be detected from albumin levels in urine, higher levels of albumin in urine indicate the kidney's filtration function is impaired. The latest breakthrough in the medical field is the use of localized sensor plasma resonance (LSPR) in the field of biosensors to detect albumin protein in urine. Using Finite-Difference Time-Domain (FDTD) simulation, the lateral size of hexagonal gold nanoparticles (AuNPs) is varied to increase the optimal LSPR. AuNPs are well-known as the most promising material for plasmonic applications. In several studies, the hexagonal geometry was reported to increase the surface area of nanoparticles as well as enhance the LSPR signal. Varying the geometry of the hexagonal AuNPs will affect their sensitivity and optical properties. In this study, the lateral size of the hexagonal AuNPs varied from 10 nm to 60 nm, with a controlled thickness of 30 nm. The absorption graph blue shifts with the peak decreasing until the graph has almost no peak as the lateral size decreases. The lateral size of 40 nm at a thickness of 30 nm was found to have the highest shifting response towards the change of refractive index, indicating good sensitivity for sensors. Simulation of hexagonal AuNPs-based LSPR detection of albumin concentrations of 0.35 –5.71 mM produces a sensitivity of 2.59 nm/mM, equivalent to 245.96 nm/RIU. The competitive performance and ease of detection procedures in real applications show the potential of the hexagonal AuNPs-based LSPR sensor to be further explored in developing alternative albumin detection techniques.

关键词： Nanoparticles Geometry Gold Sensitivity Filtration Absorption Refractive index

College Spread of COVID-19 in Ohio

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Akinola, Akinkunle Brooks, Kimberlyn Atadoga, Akoh Chandna, Vibhuti Green, Robert C. Data Science Program Bowling Green State University Bowling GreenOH United States Department of Computer Science Bowling Green State University Bowling GreenOH United States

Cumulative COVID-19 case counts by county in Ohio were gathered and combined with population data from the Census Bureau and student enrollment by county from the Integrated Postsecondary Education data System (IPEDS) for colleges or universities that compete in NCAA sports. Monthly median percent increases, monthly average percent increases, and average cases per 100k of COVID cases between counties with colleges and counties without colleges were calculated. The Wilcoxon test was used to determine if the samples were similar, and the analysis found the differences were statistically significant. Metro and non-metro groupings were added to the average cases per 100k to further subdivide the data set. Analysis found no statistically significant differences. © 2022, CC BY.

关键词： COVID-19

Large-Scale Low-Rank Gaussian Process Prediction with Support Points

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Song, Yan Dai, Wenlin Genton, Marc G. Institute of Statistics and Big Data Renmin University of China China Statistics Program King Abdullah University of Science and Technology Saudi Arabia

Low-rank approximation is a popular strategy to tackle the "big n problem" associated with large-scale Gaussian process regressions. Basis functions for developing low-rank structures are crucial and should be carefully specified. Predictive processes simplify the problem by inducing basis functions with a covariance function and a set of knots. The existing literature suggests certain practical implementations of knot selection and covariance estimation;however, theoretical foundations explaining the influence of these two factors on predictive processes are lacking. In this paper, the asymptotic prediction performance of the predictive process and Gaussian process predictions are derived and the impacts of the selected knots and estimated covariance are studied. The use of support points as knots, which best represent data locations, is advocated. Extensive simulation studies demonstrate the superiority of support points and verify our theoretical results. Real data of precipitation and ozone are used as examples, and the efficiency of our method over other widely used low-rank approximation methods is verified. Copyright © 2022, The Authors. All rights reserved.

关键词： Forecasting