DOCSLIB.ORG

Incorporating a Physics Engine 4.1 the Galileo Model

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Incorporating a Physics Engine 4.1 the Galileo Model Computational Perception of Physical Object Properties by Jiajun Wu B.Eng., B.Ec., Tsinghua University (2014) Submitted to the Department of Electrical Engineering and Computer Science in partial fulfillment of the requirements for the degree of Master of Science at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY February 2016 @ Massachusetts Institute of Technology 2016. All rights reserved. Signature redacted Author ................. Department of E)rical Engineering and Computer Science January 29, 2016 Certified by ................... Signature redacted......... William T. Freeman Thomas and Gerd Perkins Professor of Electrical Engineering and Computer Science Thesis Supervisor Signature redacted C ertified by ....... .................... Joshua B. Tenenbaum Professor of Computational Cognitive Science Thesis Supervisor Signature redacted A ccepted by ................. ...... U---- - .. .... .. .... .... .. Leslie A. Kolodziejski on Graduate Students MASSACHUSMS INSTTUTE Chair, Depar ent Committee OF TECHNOLOGY APR 15 2016 LIBRARIES ARCHIVES 2 Computational Perception of Physical Object Properties by Jiajun Wu Submitted to the Department of Electrical Engineering and Computer Science on January 29, 2016, in partial fulfillment of the requirements for the degree of Master of Science Abstract We study the problem of learning physical object properties from visual data. In- spired by findings in cognitive science that even infants are able to perceive a physical world full of dynamic content at a early age, we aim to build models to characterize object properties from synthetic and real-world scenes. We build a novel dataset con- taining over 17, 000 videos with 101 objects in a set of visually simple but physically rich scenarios. We further propose two novel models for learning physical object prop- erties by incorporating physics simulators, either a symbolic interpreter or a mature physics engine, with deep neural nets. Our extensive evaluations demonstrate that these models can learn physical object properties well and, with a physic engine, the responses of the model positively correlate with human responses. Future research directions include incorporating the knowledge of physical object properties into the understanding of interactions among objects, scenes, and agents. Thesis Supervisor: William T. Freeman Title: Thomas and Gerd Perkins Professor of Electrical Engineering and Computer Science Thesis Supervisor: Joshua B. Tenenbaum Title: Professor of Computational Cognitive Science 3 4 Acknowledgments I would like to express sincere gratitude to my advisors, Professor William Freeman and Professor Joshua Tenenbaum. Bill and Josh are always inspiring and encouraging, and have led me through my research with profound insights. Not only have they taught me how to aim for top-quality research, but they have been sharing with me invaluable lessons about life. I deeply appreciate the guidance and support from my undergraduate advisor, Professor Zhuowen Tu, who introduced me into the world of AI and vision, and has long been my mentor and friend since then. I also thank Professor Andrew Chi- Chih Yao and Professor Jian Li for advising me during my undergraduate study, Dr. Yuandong Tian for mentoring me at Facebook AI Research, and Dr. Kai Yu and Dr. Yinan Yu for mentoring me at Baidu Research. The thesis would not have been possible without the inspiration and support from my colleagues in the MIT Vision Group and Computational Cognitive Science (Co- CoSci) Group. I would like to deliver my appreciation to my collaborators, Dr. Joseph Lim, Tianfan Xue, and Dr. Ilker Yildirim. I am also thankful to other encouraging and helpful group members, especially Andrew Owens, Donglai Wei, Dr. Tomer Ull- man, Katie Bouman, Kelsey Allen, Tejas Kulkarni, Dr. Dilip Krishnan, Dr. Hossein Mohabi, Dr. Tali Dekel, Dr. Daniel Zoran, Pedro Tsividis, and Hongyi Zhang. I would like to extend my appreciation to my dear friends for their backing in my academic and daily life. I received the Edwin S. Webster Fellowship during my first year, and have been partially funded by NSF-6926677 (Reconstructive Recognition). I appreciate the sup- port from all funding agencies. Finally, I thank my parents, for their lifelong encouragement and love. 5 6 Contents 1 Introduction 13 2 Modeling the Physical World 17 2.1 Scenarios ......... .................................. 18 2.2 The Physics 101 Dataset ......................... 20 3 Physical Object Model: Learning with a Symbolic Interpreter 23 3.1 Visual Property Discoverer ... .... ..... .... .... .... 24 3.2 Physics Interpreter ... .... .... ..... .... .... .... 25 3.3 Physical World Simulator . .... .... .... .... .... ... 26 3.4 Experim ents ..... .... .... .... .... .... .... ... 27 3.4.1 Learning Physical Properties ... .... .... .... ... 28 3.4.2 Detecting Objects with Unusual Properties . .... .... .. 30 3.4.3 Predicting Outcomes . ..... ..... .... ..... ... 31 4 Physical Object Model: Incorporating a Physics Engine 35 4.1 The Galileo M odel . .... .... ... .... ... .... ... .. 35 4.1.1 Tracking as Recognition .... ..... ...... ...... 38 4.1.2 Inference .. .... .... .... .... .... .... ... 38 4.2 Sim ulations ... .... .... .... ..... .... .... .... 38 4.3 Bootstrapping as Efficient Perception in Static Scenes ... ..... 39 4.4 Experim ents .. ...... ..... .... ..... ..... ..... 41 4.4.1 Outcome Prediction . ..... ..... .... ..... ... 41 7 4.4.2 M ass Prediction ..... ...... ...... ..... .... 42 4.4.3 "Will it move" Prediction ... ...... ...... ...... 43 5 Beyond Understanding Physics 45 6 Conclusion 47 8 List of Figures 2-1 Abstraction of the physical world, and a snapshot of our dataset. 18 2-2 Illustrations of the scenarios in our Physics 101 dataset. ... .... 19 2-3 Physics 101: this is the set of objects we used in our experiments. We vary object material, color, shape, and size, together with external conditions such as the slope of a surface or the stiffness of a string. Videos recording the motions of these objects interacting with target objects will be used to train our algorithm. .. .... .... .... 20 3-1 Our first model exploits the advancement of machine learning algo- rithm (convolutional neural network) - we supervise all levels by a physics interpreter. This interpreter provides the physical constraints on what each layer can take values. During the training and testing, our model has no label of physical properties, in contrast to the stan- dard approaches. .... .... .... .... .... .... ... 24 3-2 Charts for the estimations of rings. The physical properties, especially density, of the first ring is different from those of the other rings. The difference is hard to perceive by merely visual appearances; however, by observing videos with object interactions, our algorithm is able to learn the properties and find the outlier. All figures are on a log-normalized scale. .... ... .... .... .... .... .... .... .... 30 3-3 Heat maps of user predictions, model outputs (in orange), and ground truths (in white). Objects from top to bottom, left to right: dough, metal coin, metal pole, plastic block, plastic doll, and porcelain. .. 31 9 4-1 Our second model formalizes a hypothesis space of physical object rep- resentations, where each object is defined by its mass, friction coef- ficient, 3D shape, and a positional offset w.r.t. an origin. To model videos, we draw objects from that hypothesis space into the physics engine. The simulations from the physics engine are compared to ob- servations in the velocity space. ... ........ ....... ... 36 4-2 Simulation results. Each row represents one video in the data: (a) the first frame of the video, (b) the last frame of the video, (c) the first frame of the simulated scene generated by Bullet, (d) the last frame of the simulated scene, (e) the estimated object with larger mass, (f) the estuimateu oUjecUtwih larig iiction cIeUfIcIent. .... ...... .. 39 4-3 Mean squared errors of oracle estimation, our estimation, and uniform estimations of mass on a log-normalized scale, and the correlations between estimations and ground truths . ..... ...... ..... 41 4-4 The log-likelihood traces of several chains with and without recognition- model (LeNet) based initializations. .... .... .... .... ... 41 4-5 Mean errors in numbers of pixels of human predictions, Galileo out- puts, and a uniform estimate calculated by averaging ground truth ending points over all test cases. As the error patterns are similar for both target objects (foam and cardboard), the errors here are averaged across target objects for each material. .. .... ..... ..... 43 4-6 Heat maps of user predictions, Galileo outputs (orange crosses), and ground truths (white crosses). ... .... .... ... .... .... 43 4-7 Average accuracy of human predictions and Galileo outputs on the tasks of mass prediction and "will it move" prediction. Error bars indicate standard deviations of human accuracies. .. .... ... .. 44 10 List of Tables 3.1 Accuracies (%, for oracle) or clustering purities (%, for joint training) on material estimation. In the joint training case, as there is no su- pervision on the material layer, it is not necessary for the network to specifically map the responses in that layer to material labels, and we do not expect the numbers to be comparable with the oracle case. Our analysis is just to show even in this case the network implicitly grasps some knowledge of object materials. ..... ...... ...... .. 29 3.2 Correlation coefficients of our estimations and ground truth for mass, density, and
Load more

Recommended publications
  • Photographing Black Holes with the Event Horizon Telescope - Past, Present and Future
    Photographing Black Holes with the Event Horizon Telescope - Past, Present and Future - Kazu Akiyama MIT Haystack Observatory The Shadow of a Black Hole BH ~5.2 Rs Rs = 2GMBH/c2 (Hilbert 1916) Credit: Hung-Yi Pu Kazu Akiyama, NEROC Symposium 2020, Online, 2020/11/17 (Tue) Event Horizon Telescope Kazu Akiyama, NEROC Symposium 2020, Online, 2020/11/17 (Tue) Event Horizon Telescope Sgr A* EHT 50μas Credit: Hotaka Shiokawa M87 EHT 40μas Credit: Monika Moscibrodzka Kazu Akiyama, NEROC Symposium 2020, Online, 2020/11/17 (Tue) Units of the Angular Size Protractor：1 ticks = 1 degree x 1/60 = 1 arcmin x 1/60 = 1 arcsec x 1/1000 = 1 mas x 1/1000 = 1 μas 0.5 deg 30 arcmin 40 - 50 μas Kazu Akiyama, NEROC Symposium 2020, Online, 2020/11/17 (Tue) Event Horizon Telescope Collaboration >300 members, >59 institutes, >18 countries in North & South America, Europe, Asia, and Africa. Kazu Akiyama, NEROC Symposium 2020, Online, 2020/11/17 (Tue) Meet the Telescope SMT, Arizona LMT, Mexico IRAM 30m Spain APEX, Chile JCMT, Hawaii Photos: ALMA, Sven Dornbusch, Junhan Kim, Helge Rottmann, David Sanchez, Daniel Michalik, Jonathan Weintroub, William Montgomerie Meet the Telescope Photos: ALMA, Sven ALMA, Chile Dornbusch, Junhan Kim, Helge Rottmann, David Sanchez, Daniel Michalik, Jonathan Weintroub, William Montgomerie SPT, South Pole SMA, Hawaii From Observations to Images Kazu Akiyama, NEROC Symposium 2020, Online, 2020/11/17 (Tue) Credit: Lindy Blackburn EHT Hardware EHT Backend (R2DBE, Mark 6) Recording Rate: - VLBA, GMVA: 2-4 Gbps - EHT: 32 Gbps (2017), 64 Gbps (2018-) Kazu Akiyama, NEROC Symposium 2020, Online, 2020/11/17 (Tue) From Observations to Images MIT Haystack Observatory 8 TB x 8 HDD (x 92 modules) Credit: Bryce Vickmark Credit: Lindy Blackburn Kazu Akiyama, NEROC Symposium 2020, Online, 2020/11/17 (Tue) Data Calibration HOPS Pipeline (EHT-HOPS) CASA Pipeline (rPICARD) EHT AIPS Pipeline Blackburn et al.
    [Show full text]
  • "Seeing a Black Hole" the First Image of a Black Hole from the Event Horizon Telescope
    "Seeing a Black Hole" The First Image of a Black Hole from the Event Horizon Telescope Matthew Newby Temple University Department of Physics May 1, 2019 "Seeing a Black Hole" The First Image of a Black Hole from the Event Horizon Telescope ● Black Holes – a Background ● Techniques to observe M87* ● Implications Matthew Newby, Temple University, May 1, 2019 2 "Seeing a Black Hole" Matthew Newby, Temple University, May 1, 2019 3 What is a black hole? Classical escape velocity: Let vesc → c, and the escape velocity is greater than the speed of light → “Black” Matthew Newby, Temple University, May 1, 2019 4 In General Relativity Einstein (and Hilbert) Field Equation Metric tensor Stress-energy tensor Ricci curvature tensor Scalar curvature “Solution” (“source” term) Matthew Newby, Temple University, May 1, 2019 5 Schwarzschild Metric Spherically symmetric, isolated, vacuum solution: Constant rs Schwarzschild Radius: Looks like classical escape velocity! Matthew Newby, Temple University, May 1, 2019 6 Space-Time Interval ● τ is proper time ● t is time measured at infinity (τ∞) ● r, θ, φ, are Schwarzschild spherical coordinates (i.e., coordinates as viewed at infinity) ● ds is a path element in spacetime Matthew Newby, Temple University, May 1, 2019 7 General Relativistic Time Dilation Allow a photon emitted at r to travel to infinity; in that photon’s rest frame: Since the frequency of a photon is a proper time interval, this implies that the photon’s frequency (and energy) are lower as it travels away from a spherical mass. Matthew Newby, Temple University, May 1, 2019 8 The Event Horizon rs is the “Surface of infinite redshift” or event horizon The event horizon is the “surface” of a black hole.
    [Show full text]
  • An Unprecedented Global Communications Campaign for the Event Horizon Telescope First Black Hole Image
    An Unprecedented Global Communications Campaign Best for the Event Horizon Telescope First Black Hole Image Practice Lars Lindberg Christensen Colin Hunter Eduardo Ros European Southern Observatory Perimeter Institute for Theoretical Physics Max-Planck Institute für Radioastronomie [email protected] [email protected] [email protected] Mislav Baloković Katharina Königstein Oana Sandu Center for Astrophysics | Harvard & Radboud University European Southern Observatory Smithsonian [email protected] [email protected] [email protected] Sarah Leach Calum Turner Mei-Yin Chou European Southern Observatory European Southern Observatory Academia Sinica Institute of Astronomy and [email protected] [email protected] Astrophysics [email protected] Nicolás Lira Megan Watzke Joint ALMA Observatory Center for Astrophysics | Harvard & Suanna Crowley [email protected] Smithsonian HeadFort Consulting, LLC [email protected] [email protected] Mariya Lyubenova European Southern Observatory Karin Zacher Peter Edmonds [email protected] Institut de Radioastronomie de Millimétrique Center for Astrophysics | Harvard & [email protected] Smithsonian Satoki Matsushita [email protected] Academia Sinica Institute of Astronomy and Astrophysics Valeria Foncea [email protected] Joint ALMA Observatory [email protected] Harriet Parsons East Asian Observatory Masaaki Hiramatsu [email protected] Keywords National Astronomical Observatory of Japan Event Horizon Telescope, media relations, [email protected] black holes An unprecedented coordinated campaign for the promotion and dissemination of the first black hole image obtained by the Event Horizon Telescope (EHT) collaboration was prepared in a period spanning more than six months prior to the publication of this result on 10 April 2019.
    [Show full text]
  • Feature Extraction on Synthetic Black Hole Images
    Feature Extraction on Synthetic Black Hole Images Joshua Yao-Yu Lin 1 George N. Wong 1 Ben S. Prather 1 Charles F. Gammie 1 2 3 Abstract 80 80 The Event Horizon Telescope (EHT) recently re- 60 60 leased the first horizon-scale images of the black 40 40 hole in M87. Combined with other astronomical 20 20 data, these images constrain the mass and spin of as 0 0 the hole as well as the accretion rate and magnetic µ 20 20 flux trapped on the hole. An important question − − 40 40 for EHT is how well key parameters such as spin − − 60 60 and trapped magnetic flux can be extracted from − − 80 80 − 75 50 25 0 25 50 75 − 75 50 25 0 25 50 75 present and future EHT data alone. Here we ex- − − − − − − plore parameter extraction using a neural network µas µas trained on high resolution synthetic images drawn from state-of-the-art simulations. We find that the Figure 1. Synthetic image examples. Left panel: full resolution neural network is able to recover spin and flux simulated black hole image based on numerical model of black with high accuracy. We are particularly interested hole accretion flow with M87-like parameters. Right panel: the in interpreting the neural network output and un- same simulated image convolved with 20µas FWHM Gaussian derstanding which features are used to identify, beam meant to represent the resolving power of the EHT. e.g., black hole spin. Using feature maps, we find that the network keys on low surface brightness with data from other sources, the EHT images constrain features in particular.
    [Show full text]
  • Reasons for Lack of Diversity in Open Source: the Case Katie Bouman
    Reasons for lack of diversity in open source: The case Katie Bouman Lina Wang and Kristina Weinberger February 9, 2020 1 Introduction Open Source communities face similar problems with diversity as other areas of Computer Science. In this paper, we aim to take a look at the current situation and why diversity matters. Then, we discuss the relatively recent events around Katie Bouman and the first picture of a Black Hole as an example for some of the possible reasons for a lack of women. This is followed by an overview of other common issues that may contribute to imbalances in Open Source contributions. Finally, we spend some time looking at positive developments and initiatives aimed at bringing in more diversity. 1.1 Current Status of Diversity in Open Source It is well known that women and other minority groups tend to be underrepresented in computer science in general. This lack of diversity also extends into open source projects. According to a survey among more than 5000 randomly selected github1 users, more than 90% of all contributors to open source projects identify as male, while only around 3% identify as female [24]. The survey results show a lower percentage of women, non-binary people and racial minorities than present in the wider area of computer science [7], which is already well below the general population. While this survey only shows a narrow slice of the entire open source community and other communities around free and open technologies, it is reasonable to expect a relative lack of women and other minorities overall.
    [Show full text]
  • A Shot in the Dark: the First Image of a Black Hole
    A Shot in the Dark: The First etry (VLBI) (visit EHT’s home page for more information[14]). VLBI create a collection of radio Image of a Black Hole wave detections that when pieced together created Written by Priscilla Dohrwardt a ”symphony” of noise (i.e radio waves). The black hole in question is only a mere 55 As Johnny Cash once said, ”I fell into a burning million light years from earth in the Messier 87 ring of fire” and on April 10th, so did the world. galaxy. It is a supermassive black hole, weighing The first ever black hole image graces us with a in at 6.5 billion times the mass of our sun. To put black interior surrounded by what is referred to as that in perspective, the center of the black hole a ring of fire. is about the size of the Milky Way Galaxy. Not only are we all astonished by this mammoth, but even Professor Heino Falcke, the man behind the project, said, ”It is one of the heaviest black holes that we think exists. It is an absolute monster, the heavyweight champion of black holes in the Uni- verse” [15]. The team who captured the iconic black hole image happened to be in a collaborative project between NASA, MIT and various other partnered The first image of a black hole, located in countries within the National Science Foundation. the Messier 87 galaxy was captured by the They used 8 different radio telescopes around the Event Horizon Telescope collaboration. Credits: world to collect an array of information.
    [Show full text]
  • First Results from the Event Horizon Telescope
    First results from the Event Horizon Telescope Dom Pesce On behalf of the Event Horizon Telescope collaboration March 10, 2020 EDSU conference The goal (ca. 2017) The goal is to take a picture of a black hole The goal (ca. 2017) Subrahmanyan Chandrasekhar The goal is to take a picture of a black hole • General relativity makes a prediction for the apparent shape of a black hole “It is conceptually interesting, if not astrophysically very important, to calculate the precise apparent shape of the black hole… Unfortunately, there seems to be no hope of observing this effect.” - James Bardeen, 1973 James Bardeen Photon ring and black hole shadow The warped spacetime around a black hole causes photon trajectories to bend event horizon b photon orbit Photon ring and black hole shadow The warped spacetime around a black hole causes photon trajectories to bend The closer these trajectories get to a critical impact parameter, the more tightly wound they become Photon ring and black hole shadow The warped spacetime around a black hole causes photon trajectories to bend The closer these trajectories get to a critical impact parameter, the more tightly wound they become Photon ring and black hole shadow The warped spacetime around a black hole causes photon trajectories to bend The closer these trajectories get to a critical impact parameter, the more tightly wound they become The trajectories interior to this critical impact parameter intersect the horizon, and these collectively form the black hole “shadow” (Falcke, Melia, & Agol 2000) while the bright surrounding region constitutes the “photon ring” Adapted from: Asada et al.
    [Show full text]
  • Who Is Katie Bouman?
    Who Is Katie Bouman? In April 2019, the first image of a black hole was captured in a photograph. Dr Katie Bouman (aged 29) was a member of the team that made this happen. She has made headlines around the world because of this huge achievement. Previously, it was believed that black holes were invisible. Early Life Katie was born on 9th May 1989 in Indiana, USA. She showed a keen interest in science and first learnt about the Event Horizon Telescope (EHT) while she was at school. The EHT is a project that collects information from telescopes all over the world. It has helped scientists find out more about black holes. When she left school, she spent several years at university, studying computer science and electrical engineering, before becoming a member of the EHT team. How Did She Do It? In 2016, she gave a talk to explain how it might be possible to take a photo of a black hole, something scientists once thought was impossible. Katie led the way in developing an algorithm, a set of mathematical calculations, that allowed data from different telescopes to be gathered so the image could be created. Teamwork Although Katie made headlines around the world, she was keen to point out that the photo was a team effort, involving 200 scientists and taking over three years to achieve. The scientists were delighted when they finally captured the image on 10th April 2019. 'Black Hole Image Makes History' by NASA Goddard Photo and Video is licensed under CC BY 2.0 'Black Hole Image Makes History' by NASA Goddard Photo and Video is licensed under CC BY 2.0 Black Hole Facts The largest stars in space sometimes explode.
    [Show full text]
  • Written Testimony of Katherine L. Bouman, Phd Before the Committee on Science, Space, and Technology United States House of Repr
    Written Testimony of Katherine L. Bouman, PhD before the Committee on Science, Space, and Technology United States House of Representatives May 16, 2019 INTRODUCTION Chairwoman Johnson, Ranking Member Lucas, and Members of the Committee, it is an honor to be here today. I thank you for your interest in studying black holes through imaging, and your support for this incredible breakthrough. My name is Katherine (Katie) Bouman. I am currently a postdoctoral fellow at the Harvard- Smithsonian Center for Astrophysics, and in a few weeks will be starting as an Assistant Professor at the California Institute of Technology. However, like many Event Horizon Telescope (EHT) scientists, I began contributing to this project as a graduate student. My primary role in the project has been developing methods to reconstruct images from the EHT data, as well as designing procedures to validate these images. This morning I want to tell you about one piece of the full story that made this image possible: the diverse team and computational procedures used to make the first image of a black hole from data collected at seven telescopes around the globe. THE EHT’S COMPUTATIONAL TELESCOPE On April 10th we presented the first ever image of a black hole. This stunning image shows a ring of light surrounding the dark shadow of the supermassive black hole in the heart of the Messier 87 (M87) galaxy. Since M87 is 55 million light years away, this ring appears incredibly small on the sky: roughly 40 microarcseconds in size, comparable to the size of an orange on the surface of the Moon as viewed from our location on Earth.
    [Show full text]
  • Event Horizon Telescope: the Black Hole Seen Round the World
    EVENT HORIZON TELESCOPE: THE BLACK HOLE SEEN ROUND THE WORLD HEARING BEFORE THE COMMITTEE ON SCIENCE, SPACE, AND TECHNOLOGY HOUSE OF REPRESENTATIVES ONE HUNDRED SIXTEENTH CONGRESS FIRST SESSION MAY 16, 2019 Serial No. 116–19 Printed for the use of the Committee on Science, Space, and Technology ( Available via the World Wide Web: http://science.house.gov U.S. GOVERNMENT PUBLISHING OFFICE 36–301PDF WASHINGTON : 2019 COMMITTEE ON SCIENCE, SPACE, AND TECHNOLOGY HON. EDDIE BERNICE JOHNSON, Texas, Chairwoman ZOE LOFGREN, California FRANK D. LUCAS, Oklahoma, DANIEL LIPINSKI, Illinois Ranking Member SUZANNE BONAMICI, Oregon MO BROOKS, Alabama AMI BERA, California, BILL POSEY, Florida Vice Chair RANDY WEBER, Texas CONOR LAMB, Pennsylvania BRIAN BABIN, Texas LIZZIE FLETCHER, Texas ANDY BIGGS, Arizona HALEY STEVENS, Michigan ROGER MARSHALL, Kansas KENDRA HORN, Oklahoma RALPH NORMAN, South Carolina MIKIE SHERRILL, New Jersey MICHAEL CLOUD, Texas BRAD SHERMAN, California TROY BALDERSON, Ohio STEVE COHEN, Tennessee PETE OLSON, Texas JERRY MCNERNEY, California ANTHONY GONZALEZ, Ohio ED PERLMUTTER, Colorado MICHAEL WALTZ, Florida PAUL TONKO, New York JIM BAIRD, Indiana BILL FOSTER, Illinois JAIME HERRERA BEUTLER, Washington DON BEYER, Virginia JENNIFFER GONZA´ LEZ-COLO´ N, Puerto CHARLIE CRIST, Florida Rico SEAN CASTEN, Illinois VACANCY KATIE HILL, California BEN MCADAMS, Utah JENNIFER WEXTON, Virginia (II) CONTENTS May 16, 2019 Page Hearing Charter .....................................................................................................
    [Show full text]
  • Antonio Torralba Katie L
    DAILY Interview with: Presenting Work by: Editorial with: Antonio Torralba Katie L. Bouman Rita Cucchiara John McCormac Today’s Picks by: Workshop: Dèlia The Joint Video and Language... Fernandez A publication by Wednesday 2 Dèlia’s Picks For today, Wednesday 25 Dèlia Fernandez is an industrial PhD candidate at the startup company Vilynx and the group of Image and Video Processing (GPI) in the Polytechnic University of Catalonia (BarcelonaTech). “My research topic aims at introducing knowledge bases in the automatic understanding of video content. ICCV is the perfect place to learn about last trends and techniques for image and video analysis, get in touch with the community and hear about new interesting topics and challenges.” Dèlia presented a paper in the workshop on web-scale vision and social media. The paper is "VITS: Video Dèlia Fernandez Tagging System from Massive Web Multimedia Collections". Find it here! Dèlia’s picks of the day: • Morning P3-34: Temporal Dynamic Graph LSTM for Action-Driven Video Object Detection P3-42: Hierarchical Multimodal LSTM for Dense Visual-Semantic Embedding P3-50: Unsupervised Learning of Important Objects From First-Person Videos P3-52: Visual Relationship Detection With Internal and External Linguistic Knowledge Distillation P3-72: Common Action Discovery and Localization in Unconstrained Videos P4-69: Learning View-Invariant Features for Person Identification in Temporally Synchronized Videos Taken by Wearable Cameras P4-74: Chained Multi-Stream Networks Exploiting Pose, Motion, and Appearance for Action Classification and Detection “Venice is a magic place where you can feel Italian traditions alive! I was delighted to find out ICCV was taking place here and I am enjoying the chance to have nice Italian food and explore the city.” Wednesday Summary 3 Dèlia’s Picks Antonio Torralba Katie L.
    [Show full text]
  • How to Navigate an Information Media Environment Awash in Manipulation, Falsehood, Hysteria, Vitriol, Hyper-Partisan Deceit and Pernicious Algorithms
    How to Navigate an Information Media Environment Awash in Manipulation, Falsehood, Hysteria, Vitriol, Hyper-Partisan Deceit and Pernicious Algorithms A Guide for the Conscientious Citizen A Reflection Paper prepared for the Canadian Commission for UNESCO and the Canadian Committee for World Press Freedom By Christopher Dornan Ottawa, Canada, August 2019 To quote this article: DORNAN, Christopher. ‘‘How to Navigate an Information Media Environment Awash in Manipulation, Falsehood, Hysteria, Vitriol, Hyper-Partisan Deceit and Pernicious Algorithms’’, the Canadian Commission for UNESCO’s IdeaLab, July 2019. The views and opinions expressed in this article are those of the author and do not necessarily reflect the official policy or position of the Canadian Commission for UNESCO. About the author Christopher Dornan teaches at Carleton University, where he served for nine years as director of the School of Journalism and Communication and six years as director of the Arthur Kroeger College of Public Affairs. His academic work has appeared in venues from the Media Studies Journal to the Canadian Medical Association Journal. He is the co-editor (with Jon Pammett) of The Canadian Federal Election of 2015, along with five previous volumes in this series. In 2017, he wrote the Canadian Commission for UNESCO reflection paper “Dezinformatsiya: The Past, Present and Future of ‘Fake News’.” He is chair of the board of Reader’s Digest Magazines Canada, Inc. ii Table of Contents Introduction ................................................................................................................................
    [Show full text]