Mellon University Research Articles

MMDA: Multi‐person marginal distribution awareness for monocular 3D pose estimation

AbstractMost existing 3D pose representations cannot completely decouple the overlapping two or more human joints of the same type. In this paper, the authors propose a novel 2.5 D representation of the human pose by projecting human joints in 3D space onto the three orthogonal planes. The authors apply for the first time the permutation module to a multi‐person 3D human pose estimation task and use Geometric Constraints Loss (GCL) to guide the learning of the model. The authors overcome the negative effects of the inductive bias of convolutional neural networks (CNNs) by aligning the intermediate feature space with the output feature space. The effectiveness of the authors’ approach is validated on the carnegie mellon university (CMU) panoptic dataset and MuPoTS‐3D dataset. The authors’ proposed representations can effectively decouple the human joints in their selected data from overlapping human joints.

Presidential Address: A Coevolutionary History of COVID-19; Culture, Biology, and Mental Health

Previous articleNext article No AccessPresidential Address: A Coevolutionary History of COVID-19; Culture, Biology, and Mental HealthEdmund RussellEdmund RussellCarnegie Mellon University Search for more articles by this author PDFPDF PLUSFull Text Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinkedInRedditEmail SectionsMoreDetailsFiguresReferencesCited by Volume 27, Number 1January 2022 Published for the American Society for Environmental History and the Forest History Society Views: 345Total views on this site Article DOIhttps://doi.org/10.1086/717600 Views: 345Total views on this site HistoryPublished online December 22, 2021 © 2021 American Society for Environmental History and the Forest History Society. All rights reserved.PDF download Crossref reports no articles citing this article.

U.S. frontline workers and COVID-19 inequities

We overcome a lack of frontline worker status information in most COVID-19 data repositories to document the extent to which occupation has contributed to COVID-19 disparities in the United States. Using national data from over a million U.S. respondents to a Facebook-Carnegie Mellon University survey administered from September 2020 to March 2021, we estimated the likelihoods of frontline workers, compared to non-frontline workers, 1) to ever test positive for SARs-Cov-2 and 2) to test positive for SARs-Cov-2 within the past two weeks. Net of other covariates including education level, county-level political environment, and rural residence, both healthcare and non-healthcare frontline workers had higher odds of having ever tested positive for SARs-Cov-2 across the study time period. Similarly, non-healthcare frontline workers were more likely to test positive in the previous 14 days. Conversely, healthcare frontline workers were less likely to have recently tested positive. Our findings suggest that occupational exposure has played an independent role in the uneven spread of the virus. In particular, non-healthcare frontline workers have experienced sustained higher risk of testing positive for SARs-Cov-2 compared to non-frontline workers. Alongside more worker protections, future COVID-19 and other highly infectious disease response strategies must be augmented by a more robust recognition of the role that structural factors, such as the highly stratified U.S. occupational landscape, have played in the uneven toll of the COVID-19 pandemic.

Towards inclusive scholarly publishing: developments in the university press community

This article provides an overview of the ways in which the members of the Association of University Presses are working towards more inclusive practices in scholarly publishing. The authors consider the Mellon University Press Diversity Fellowship Program (now in its fourth year), the work of the Association’s Diversity and Inclusion Task Force, the Gender, Equity and Cultures of Respect Task Force and the new Equity, Justice and Inclusion Committee. They also look at press-based working groups and several ‘Toolkits for Equity’ that are currently in development. The volunteers engaged in these and other efforts are working to document how bias has shaped universities and university presses, to propose actions to disrupt this powerful force and to share what they have learned with their colleagues as well as with the larger scholarly publishing and academic communities.

News from the Field

IU launches rare newspaper, magazine collection2019 Scholarly Communications Research Grant Recipients announcedGPO digitizes Congressional hearingsCarnegie Mellon University, Elsevier announce access agreementCairn.info joins OASIS Publisher-Direct ProgramNominations sought for ACRL Board of Directors

772: PREDICTING DELAYED CEREBRAL ISCHEMIA USING SEQUENTIAL PATTERNS IN PLETHYSMOGRAPHY DATA

1Carnegie Mellon University Auton Lab, Pittsburgh, PA 2University of Pittsburgh School of Medicine, Pittsburgh, PA 3University of Pittsburgh, Pittsburgh, PA

Achieving Personalized $k$-Anonymity-Based Content Privacy for Autonomous Vehicles in CPS

Enabled by the industrial Internet, intelligent transportation has made remarkable achievements such as autonomous vehicles by carnegie mellon university (CMU) Navlab, Google Cars, Tesla, etc. Autonomous vehicles benefit, in various aspects, from the cooperation of the industrial Internet and cyber-physical systems. In this process, users in autonomous vehicles submit query contents, such as service interests or user locations, to service providers. However, privacy concerns arise since the query contents are exposed when the users are enjoying the services queried. Existing works on privacy preservation of query contents rely on location perturbation or $k$ -anonymity, and they suffer from insufficient protection of privacy or low query utility incurred by processing multiple queries for a single query content. To achieve sufficient privacy preservation and satisfactory query utility for autonomous vehicles querying services in cyber-physical systems, this article proposes a novel privacy notion of client-based personalized $k$ -anonymity (CP $k$ A). To measure the performance of CP $k$ A, we present a privacy metric and a utility metric, based on which, we formulate two problems to achieve the optimal CP $k$ A in term of privacy and utility. An approach, including two modules, to establish mechanisms which achieve the optimal CP $k$ A is presented. The first module is to build in-group mechanisms for achieving the optimal privacy within each content group. The second module includes linear programming-based methods to compute the optimal grouping strategies. The in-group mechanisms and the grouping strategies are combined to establish optimal CP $k$ A mechanisms, which achieve the optimal privacy or the optimal utility. We employ real-life datasets and synthetic prior distributions to evaluate the CP $k$ A mechanisms established by our approach. The evaluation results illustrate the effectiveness and efficiency of the established mechanisms.

Spatio-temporal summarization of dance choreographies

An important issue in performing dance analysis is the automatic extraction of its choreographic patterns, since these elements provide an abstract representation of the semantics of the dance and encode the overall dance storytelling. However, application of conventional video summarization algorithms on dance sequences cannot appropriately retrieve their choreographic patterns, since a dance is composed of an ordered set of sequential elements which are often repeated in time. Additionally, 3D geometry is lost using color information. For this reason, in this paper we propose a new dance summarization scheme of 3D motion captured data (in the form of skeleton joints coordinates) recorded using the Vicon motion capture system. The proposed key frame extraction method implements a hierarchical scheme that exploits spatio-temporal variations of dance features. Initially, global holistic descriptors are extracted to localize the key choreographic steps of a dance (coarse representation). Then, each segment is further decomposed into finer sub-segments to improve dance representativity (fine representation). The abstraction scheme exploits the concepts of a Sparse Modeling Representative Selection (SMRS) appropriately modified to enable spatio-temporal modelling of the dance sequences through a hierarchical decomposition algorithm. Our approach is evaluated on thirty folkloric dance sequences recorded at the Aristotle University of Thessaloniki under the framework of Terpsichore project representing five different choreographies and on publicly available datasets from Carnegie–Mellon University, which depict performances on theatrical kinesiology. Comparisons with other traditional video summarization methods indicate a clear superiority of the proposed hierarchical spatio-temporal decomposition scheme.

Terahertz Pioneer: Michael Bass “The THz Light at the End of the Tunnel”

As a young man growing up in the Bronx, NY, in the 1950s, Michael Bass was able to take full advantage of the museums and other unique multi-cultural offerings of a major metropolis. The extensive New York City subway system cost a nickel a ride, and the trip from the Bronx to the American Museum of Natural History on 83rd street in Manhattan, took less than an hour. Michael found himself visiting the museum frequently and developing an early interest in science. While attending the prestigious Stuyvesant High School, a public secondary school then on 15th street in Manhattan—which boasts four Nobel laureates as alumni—Michael remembers being very impressed by an exhibit at the United Nations building in NYC in the spring of 1955. The exhibit was part of the U.S. “Atoms for Peace” initiative. There he saw a functional cloud chamber, which focused his science interests on physics. In 1956 he enrolled in Carnegie Institute of Technology (now Carnegie Mellon University) in Pittsburgh, PA, to formalize a career in science.

Project Planningmeasures in CMMI

Computer information project planning is one of the most important activities in the modern software development process. Without an objective and realistic plan of software project, the development of software process cannot be managed effectively.This research will identify general measures for the specific goals and its specific practices of Project PlanningProcess Area in Capability Maturity Model Integration (CMMI). CMMI is developedin USA by Software Engineering Institute (SEI) inCarnegie Mellon University. CMMI is a framework for assessment and improvement of computer information systems. The procedure we used to determine the measures is to apply the Goal Questions Metrics (GQM) approach to the three specific goals and its fourteen specific pra ctices of Project PlanningProcess Area in CMMI.

NAE Elects New Members In 2013

The National Academy of Engineering elected 69 members and 11 foreign associates last month. The academy now totals 2,250 U.S. members and 211 foreign associates. Individuals are elected to NAE on the basis of outstanding contributions to engineering research, practice, or education. Of the new members and foreign associates, 20 are scientists who work in chemistry-related areas or are members of the American Chemical Society. The chemistry-related new members are James M. Anderson, Case Western Reserve University, for researching interactions between tissue and biomaterials and developing medical devices; Peter L. Andresen, GE Global Research Center, for predicting and preventing stress corrosion cracking in nuclear materials; Lorenz T. Biegler, Car­ne­gie Mellon University, for contributing to large-scale nonlinear optimization theory and its application; Donna G. Blackmond, Scripps Research Institute, La Jolla, Calif., for studying catalytic organic reactions and chiral amplification; Dawn A. Bonnell, ...

Palliative Care and Pain Medicine Come Together to Optimally Treat Cancer Pain: What We Can Learn from the British Pain Society

“We don't beat the Reaper by living longer. We beat the Reaper by living well.” Professor Randy Pausch declared this statement during his famous last lecture at Carnagie Mellon University prior to his death from pancreatic cancer in 2008 [1]. As opposed to the widely practiced WHO approach of utilizing opioids in a stepladder fashion, many now advocate a mechanism-based approach to cancer pain treatment. Old dogma treated cancer pain as a problem mainly toward the end of life situation in the context of metastatic, progressive disease, whereas new data show pain to be problematic throughout the cancer care cycle [2]. The effective management of these acutely painful surgeries and related treatments may limit the development of chronic pain states in long-term survivors [3]. Effective treatment strategies include multidisciplinary, multimodal care utilizing 1) combinations of long-acting opioids for constant pain with short-acting opioids for incidental pain; 2) “adjuvant” co-analgesics including nonsteroidal anti-inflammatories, anticonvulsants, antidepressants, and topical agents to optimize analgesia and minimize opioid doses, thereby reducing concomitant opioid-related side effects; 3) prophylactic treatment of constipation, nausea, and other common troublesome symptoms; 4) interventional …

In Memoriam: Cheng-Ching Yu, 1956−2009

ADVERTISEMENT RETURN TO ISSUEEditorialNEXTIn Memoriam: Cheng-Ching Yu, 1956−2009Babatunde Ogunnaike and Larry BieglerView Author Information Department of Chemical Engineering, University of Delaware, Newark, Delaware 19716 Department of Chemical Engineering, Carnegie−Mellon University, Pittsburgh, Pennsylvania 15213Cite this: Ind. Eng. Chem. Res. 2009, 48, 16, 7455–7456Publication Date (Web):July 15, 2009Publication History Received29 June 2009Published online15 July 2009Published inissue 19 August 2009https://doi.org/10.1021/ie901045jCopyright © 2009 American Chemical SocietyRIGHTS & PERMISSIONSArticle Views976Altmetric-Citations1LEARN ABOUT THESE METRICSArticle Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated. Share Add toView InAdd Full Text with ReferenceAdd Description ExportRISCitationCitation and abstractCitation and referencesMore Options Share onFacebookTwitterWechatLinked InReddit PDF (614 KB) Get e-AlertsSUBJECTS:Chemical engineering and industrial chemistry,Distillation,Process systems Get e-Alerts

Incremental Linear Discriminant Analysis for Face Recognition

Dimensionality reduction methods have been successfully employed for face recognition. Among the various dimensionality reduction algorithms, linear (Fisher) discriminant analysis (LDA) is one of the popular supervised dimensionality reduction methods, and many LDA-based face recognition algorithms/systems have been reported in the last decade. However, the LDA-based face recognition systems suffer from the scalability problem. To overcome this limitation, an incremental approach is a natural solution. The main difficulty in developing the incremental LDA (ILDA) is to handle the inverse of the within-class scatter matrix. In this paper, based on the generalized singular value decomposition LDA (LDA/GSVD), we develop a new ILDA algorithm called GSVD-ILDA. Different from the existing techniques in which the new projection matrix is found in a restricted subspace, the proposed GSVD-ILDA determines the projection matrix in full space. Extensive experiments are performed to compare the proposed GSVD-ILDA with the LDA/GSVD as well as the existing ILDA methods using the face recognition technology face database and the Carneggie Mellon University Pose, Illumination, and Expression face database. Experimental results show that the proposed GSVD-ILDA algorithm gives the same performance as the LDA/GSVD with much smaller computational complexity. The experimental results also show that the proposed GSVD-ILDA gives better classification performance than the other recently proposed ILDA algorithms.

Comment

Comment

Authenticity: Modernity, Stardom, and Rock & Roll

David Shumway is Professor of English and Literary and Cultural Studies, and Director of the humanities Center at Carnegie mellon university. he is the author of Michel Foucault, Creating American Civilization, and Modern Love: Romance, Intimacy, and the Marriage Crisis. he has co-edited three volumes, including Making and Selling Culture, and has published articles in Cinema Journal, The Film Genre Reader, PMLA, South Atlantic Quarterly, and Genre, among others. he has recently comleted a book about rock stars as cultural icons for Nyu Press and is at work on John Sayles for the university of Illinois Press Contemporary Directors Series. Authenticity: Modernity, Stardom, and Rock & Roll

Indeed I Was Pleased with the World (review)

Reviewed by: Indeed I Was Pleased with the World Jason Koo Indeed I Was Pleased with the World By Mary RuefleCarnegie Mellon University Press, 2007, 78 pp., $14.95 By far my favorite book of the year is Mary Ruefle's tenth collection of poetry, Indeed I Was Pleased with the World. Ruefle's books appear without fanfare: slim, unadorned, they dare to let the poems inside speak for themselves. Indeed comes packaged in a simple white cover with only the author's name, the title of the book, and a picture of a wilted little orange something on the front, either an orange or ball or pincushion. There's a randomness, a what-the-hell quality to the chosen image and title that mark Ruefle's work as a poet and make her lovable. Ruefle shares Keats's aversion to the "egotistical sublime." Her poems don't announce themselves; they sidle up to you, whispering their little charms in your ear. There are plenty of "good" books of poetry, but only a few that make you so fall in love with them that you want to tuck them in a shoebox and feed them spoonfuls of jam. Indeed I Was Pleased with the World is just such a book. The speaker of the poems shares the same cosmic charm as the Little Prince, the Fool in King Lear or Twain's Adam in The Diaries of Adam and Eve. The book opens with the mirthful, odd "The Day the Earth Stood Still": I remember the day the official letter came announcing that on such and such a day. And on such and such a day I was ready. I toweled off and sat by the window. I love that "I toweled off"; it is the kind of charming touch that gives Ruefle's poetry its magic. Indeed is full of such touches. Take the closing [End Page 174] passage of "Hope": "Bonnard. He's dead./Maybe because he's dead/you pronounce the final d/or maybe because he's dead/you don't." Or the beginning of "Something of a Tractate": It was gone now, the boring summer, and those who survived it were stuffed with famous haiku. Now they were asked to eat bold and dangerous amounts of macaroni. They were taught to say "he died" instead of "he is dead" and given a couple of sweaters. What is there to say in the face of such poetry? How does Ruefle move so swiftly from summer to haiku to macaroni to sweaters? Where does the reflection about the relation between death and the final "d" in "Bonnard" come from? Poetry of this kind of weirdness is always in danger of being dismissed as a "curiosity" by curmudgeonly canonmeisters. But these readers need to give Ruefle her due: her weirdness is world-class, like that of a Dali or Pessoa. Sometimes Ruefle's weirdness is mainly a source of amusement, as in "My Timid Eternity," in which the speaker imagines a heaven occupied solely by her and the first president of the United States: I am thinking how lonesome it will be in Heaven with only George Washington and me there. I suppose we will recite the Beatitudes and wonder when they are coming— the meek, the merciful, the peacemakers, those who are pure in heart. Roasting marshmallows in the evening I will broach the subject of lies. He will hand me a wig and some leeches, which I will decline, still thinking about the others— if they went to the Babies Camp by mistake we could maybe get a letter out. (I don't know what the weirdest detail here is, the marshmallows, the lies, the leeches or the Babies Camp.) But in the best poems of the book—"Peccadillo," "After a Rain" and "The Refrigerator"—Ruefle's weirdness takes her [End Page 175] into a zone of consciousness occupied solely by her and the greatest of poets. In "After a Rain," Ruefle appropriately makes the power of her own weird noticing the subject of her poem: I noticed it was silly of me to notice so much but I noticed there is no stationery in heaven...

The impact on tropospheric ozone formation on the implementation of a program for mobile emissions control: a case study in São Paulo, Brazil

The main sources of reactive hydrocarbons (RHC) and nitrogen oxides (NOx), ozone precursors, in the Metropolitan Area of Sao Paulo (MASP) in the southeast of Brazil are emissions from vehicles fleets. Ambient surface ozone and particulate matter concentrations are air quality problem in the MASP. This study examined the impact that implementing a control program for mobile emissions has on ozone concentrations, An episode of high surface ozone concentrations occurring in the MASP during the March 13–15, 2000 period was used as a case study that was modeled for photochemical oxidants using the California Institute of Technology/Carnegie Mellon University three-dimensional photochemical model. Different scenarios were analyzed in relationship to the implementation of the Programa Nacional de Controle de Poluicao por Veiculos Automotores (PROCONVE, National Program to Control Motor Vehicle Pollution). Scenario 1 assumed that all vehicles were operating within PROCONVE guidelines. Scenarios 2 and 3 considered hypothetical situations in which the PROCONVE was not implemented. Scenario 2 set the premise that vehicles were using pre-1989 technology, whereas scenario 3 allowed for technological advances. A base case scenario, in which the official emission inventory for the year 2000 was employed, was also analyzed. The CIT model results show agreement with most measurements collected during 13–15 March 2000 modeling episode. Mean normalized bias for ozone, CO, RHC and NO x are approximately 9.0, 6.0, −8.3, 13.0%, respectively. Tropospheric ozone concentrations predicted for scenario 2 were higher than those predicted for scenarios 1, 3 and base case. This study makes a significant contribution to the evaluation of air quality improvement and provides data for use in evaluating the economic costs of implementing a program of motor vehicle pollution control aimed at protecting human health.

The History of the Gulag: From Collectivization to the Great Terror . By Oleg V. Khlevniuk. Translated by Vadim A. Staklo. With editorial assistance and commentary by David J. Nordlander. Annals of Communism. Edited by Jonathan Brent. New Haven, CT: Yale University Press, 2004. Pp. xix+418. $39.95.

Previous articleNext article No AccessBook ReviewThe History of the Gulag: From Collectivization to the Great Terror . By Oleg V. Khlevniuk. Translated by Vadim A. Staklo. With editorial assistance and commentary by David J. Nordlander. Annals of Communism. Edited by Jonathan Brent. New Haven, CT: Yale University Press, 2004. Pp. xix+418. $39.95.Wendy Z. GoldmanWendy Z. GoldmanCarnegie Mellon University Search for more articles by this author Carnegie Mellon UniversityPDFPDF PLUSFull Text Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinkedInRedditEmail SectionsMoreDetailsFiguresReferencesCited by The Journal of Modern History Volume 79, Number 1March 2007 Article DOIhttps://doi.org/10.1086/517584 Views: 72Total views on this site Permission to reprint a book review printed in this section may be obtained only from the author.PDF download Crossref reports no articles citing this article.

Assets 2005 doctoral consortium

The ASSETS 2005 Doctoral Consortium was a resounding success, involving eight doctoral candidates and three faculty advisors, plus one transitional advisor. The doctoral candidates included:•Miguel Alonso Florida International University •Rita Hubert Pace University •Kanav Kahol Arizona State University •Peter Parente UNC-Chapel Hill •Kathy Price UMBC •Andrea Tartaro Northwestern University •Ted Wattenberg NOVA Southeastern University •Jacob Wobbrock Carnegie Mellon University