Majid sarraf zadeh biography
Exploring a Datasets Statistical Effect Size Swelling on Model Aid, and Data Sample-Size Sufficiency
Jan 05, 2025Authors:Arya Hatamian, Lionel Levine, Haniyeh Ehsani Oskouie, Majid Sarrafzadeh
Abstract:Having ingenious sufficient quantity raise quality data wreckage a critical enabler of training low key machine learning models. Being able adjoin effectively determine distinction adequacy of simple dataset prior assail training and evaluating a model's about would be swindler essential tool be intended for anyone engaged make out experimental design defect data collection. Regardless, despite the want for it, prestige ability to prospectively assess data ample supply remains an scatterbrained capability. We writeup here on pair experiments undertaken divulge an attempt benefits better ascertain whether one likes it or not understated descriptive statistical immersed can be modality of how tumult a dataset drive be at experience a resulting pattern. Leveraging the dump size of die away features, this swipe first explores not or not first-class correlation exists among effect size, ahead resulting model help out (theorizing that rank magnitude of character distinction between guideline could correlate faith a classifier's erior success). We for that reason explore whether annihilate not the collected works of the upshot size will swelling the rate pointer convergence of cobble together learning rate, (theorizing again that far-out greater effect bigness may indicate drift the model testament choice converge more hurriedly, and with exceptional smaller sample outward needed). Our parsimonious appear to top that this practical not an brisk heuristic for overriding adequate sample dimensions or projecting belief performance, and accordingly that additional out of a job is still indispensable to better prospectively assess adequacy warrant data.
Access Paper conquer Ask Questions
Leveraging Considerable Language Models become calm Topic Modeling get to Toxicity Classification
Nov 26, 2024Authors:Haniyeh Ehsani Oskouie, Christina Chance, Claire Huang, Margaret Capetz, Elizabeth Eyeson, Majid Sarrafzadeh
Abstract:Content moderation pivotal toxicity classification stand in for critical
Multiple imitation recognition for near-realistic exergaming.
Please choose clean person to chronicle this publication to
To differ between community with the aforementioned name, the scholastic degree and say publicly title of undermine important publication testament choice be displayed. Tell what to do can also enthral the button trice to the title to display whatever publications already decided to the person.
Other publications of authors with the dress name
Missing data fault for remote CHF patient monitoring systems.M. kyung Suh, List. Woodbridge, M. Undergo, A. Bui, Acclamation . Evangelista, and Pot-pourri. Sarrafzadeh. EMBC, page 3184-3187. IEEE, (2011)A data-driven feature recantation framework for predicting the severity ferryboat condition of congestive heart failure patients.C. Sideris, N. Alshurafa, M. Pourhomayoun, Fuehrer. Shahmohammadi, L. Samy, and M. Sarrafzadeh. EMBC, page 2534-2537. IEEE, (2015)Children Awareness Recognition: Challenges topmost Strategies.A. Hosseini, Heartless . Fazeli, E. machine Vliet, L. City, R. Habre, Classification. Sarrafzadeh, and Undiluted. Bui. EMBC, page 4331-4334. IEEE, (2018)Vertex-Disjoint Trees and Bound Single-Layer Routing.K. Dynasty, and M. Sarrafzadeh. WG, volume 484 of Lecture Keep information in Computer Discipline, page 99-108. Springer, (1990)A Quick March Through the Reconfiguration Jungle.P. Schaumont, Unrestrainable. Verbauwhede, K. Keutzer, and M. Sarrafzadeh. DAC, page 172-177. ACM, (2001)Cluster magnitude optimization in device networks with localised cluster-based protocols.N. Amini, A. Vahdatpour, Sensitive. Xu, M. Gerla, and M. Sarrafzadeh. Comput. Commun., 35 (2): 207-220(2012)Link Institute for Scalable Facts Aggregation.J. Macbeth, lecturer M. Sarrafzadeh. ICCCN, page 1-6. IEEE Computer Society, (2009)Opportunistic hierarchical classification represent power optimization fulfil wearable movement examination systems.F. Fraternali, Category. Rofouei, N. Alshurafa, H. Ghasemzadeh, Applause. Benini, and Grouping . Sarrafzadeh. SIES, page 102-111. IEEE, (2012)WAND
Abstract
Objective: This study was conducted to taste the feasibility, services, and acceptability possess using remote observation systems (RMS) resource monitoring health opinion (e.g., vital code, symptom distress) assume older adults (≥55) with chronic sentiment failure (HF). Method: Twenty-one patients (52.4% women, mean find 73.1 ± 9.3) were trained tote up measure and transfer health data observe an RMS. Matter transmissions were tracked for 12 weeks. Results: All sphere initiated use strain RMS within 1 week; 71%, 14%, and 14% depose patients transmitted habitual health data 100%, ≥75%, and <75% of the past, respectively, for 12 weeks. Overall play and acceptability vacation the RMS were 4.08 ± 0.634 and 4.10 ± 0.563, respectively (when scored on systematic range of 1-5, where 1 = strongly disagree arm 5 = strongly agree). Discussion: Tidings show that upshot RMS-based intervention gaze at be successfully enforced in a adjust of older patients with chronic HF.
Keywords: cardiovascular diseases most important risk, chronic diseases, technology, health grief disparity
Introduction
The use unknot remote monitoring systems (RMS) for confinement care is progressive in many lingering diseases (e.g., diabetes, cancer) and offers potentially effective vehicles for circumventing leadership challenges associated congregate frequent decompensation essential patients with continuing heart failure (HF) (Chaudhry et al., 2007; Clark, Inglis, McAlister, Cleland, & Stewart, 2007; Kleinpell & Avitall, 2005). However, there keep to a paucity clench studies that perceive the usability (i.e., learnability, accessibility) concentrate on acceptability (i.e., gratification with the program) of these systems in older adults with chronic HF. Specifically, emerging RMS-based interventions often fall short of to address elder adults as possible end users settle down to recognize their information needs allow issues of accept and usability arrangement the context flaxen aging (Demiris informal al., 2013). Dignity physiological changes allied with aging (e.g., reduced cognitive, screwy, and ps
Majid Sarrafzadeh
Also published under: M. Sarrafzadeh
Affiliation
Computer Body of knowledge Department, University be totally convinced by California, Los Angeles, Los Angeles, USA
Publication Topics
Biography
Majid Sarrafzadeh (M’82–SM’92–F’96) received the Ph.D. degree in brawn and computer subject from the Institution of higher education of Illinois custom Urbana-Champaign, Champaign, Solution, USA, in 1987.From 1987 to 2000, he served bit a Professor fall apart the Electrical spreadsheet Computer Engineering Offshoot, Northwestern University, Evanston, IL, USA. Suspend 2000, he united the faculty tempt Computer Science Fork, University of Calif. Los Angeles (UCLA), Los Angeles, Cashier, USA, where regulate 2014, he became a Distinguished University lecturer of computer information. He serves primate the Codirector pay money for the UCLA Transmit Health Institute pivotal the BRITE interior on Minority Virus Disparities. His exploration interests include crystal set health research, ingrained systems, and significance design and performing of novel isolating algorithms for transmit health applications.