School of
Information
Previously School of Library & Information Studies
Friday Afternoon Seminar: Summaries.
296a-1 Seminar: Information Access, Fall 2019.
Fridays 3-5. 107 South Hall.
Schedule. Weekly
mailing list.
Details will be added as they become available.
Aug 30 No Seminar meeting
Sep 6: Clifford LYNCH: Machine Learning, Archives and Special
Collections.
Introduction to the Seminar.
In this session, I'll summarize what I expect will be the most
significant near term (next ten years) applications for machine
learning in the context of archives and special collections, building
on a brief essay I wrote over the summer for the International Council
on Archives; in this discussion, I'll focus in particular on the
architecture of facial recognition systems and how these may interact
with collections, including privacy and ethical issues.
Sept 13: Clifford LYNCH: Public Policy Initiatives That Might Improve
Digital stewardship.
I. Brief Updates from Seminar Participants.
II. Public Policy Initiatives That Might Improve
Digital Stewardship.
Traditionally most discussions of public policy efforts
to enable stewardship of the cultural record have focused on copyright
law -- in particular, first sale, fair
use, preservation exemptions for memory institutions, copyright deposit
and duration and registration (and orphan works). I'll briefly
summarize some of these discussions, and underscore how poorly they are
matched to some classes of cultural materials in the
current world. I'll suggest some very tentative speculations about
alternative approaches and formulations
that might be worth considering for examination and discussion.
Sept 20: Sean BROOKS: Scaling Cybersecurity Support for Civil Society.
The internet's promise as an open platform for free
expression and assembly has come under increasing threat as governments,
hate groups, terrorists, and multinational corporations have deployed
sophisticated attacks against the operations and legitimacy of civil
society organizations, journalists, researchers, and human rights
defenders. Some responses have emerged to combat this threat - improved
security product offerings for nonprofits, emergency response funds for
organizations under attack - but if the scale of the problem is the
whole of online civil society, assistance models must find methods to
address a monumental volume of need. Citizen Clinic - the world's
first public interest cybersecurity clinic - is researching the barriers
to scaling technical assistance projects and using its experience
supporting politically targeted organizations to examine how efforts
to protect online freedom can meet the growing need. More at cltc.berkeley.edu/about-us/citizen-clinic/.
Sean Brooks is the Director of the Citizen Clinic
and a Research Fellow at the Center for Long-Term Cybersecurity at UC
Berkeley. His work in cybersecurity and privacy span a decade of
experience in the government, private sector, civil society and academia,
including helping build the Privacy Engineering program at the National
Institute of Standards and Technology and working as an advocate at
the Center for Democracy and Technology. More at https://www.linkedin.com/in/seanbrooks.
Sept 27: *2 pm* The School's 101st Birthday Celebration.
No Seminar meeting. Attend instead the School's
101st birthday celebration in South Hall 202, 2 - 4 pm.
Also no-host luncheon at the Faculty Club.
Advance registration requested. See https://www.ischool.berkeley.edu/events/2019/101st-birthday-celebration.
Oct 4: Michael BUCKLAND: This School's Social Mission: The Early
Days.
The School originated in 1918 to mitigate an acute shortage
of qualified librarians in California. In addition to the technical
skills taught, this mission was founded on
a strong social agenda to achieve economic efficiency, to foster social
harmony, to promote individual personal development, to protect civil
liberties, and to promote liberal democracy. I will review this social
agenda with special attention to the role of public libraries, the development of bibliography, the rise
of authoritarian political ideologies in the 1930s, and the interpretation
of that social agenda in today’s
circumstances.
Oct 11: CAMPUS CLOSED. Rescheduled to March
6, 2020.
Howard BESSER: Archiving the Non-Organizational Born-Digital:
The Challenges Posed by Material from Individuals, Communities, Social
Movements, and Events.
Oct 18: Aditya PARAMESWARAN and Mugdha BHUSARY.
Mugdha BHUSARY: Managing Privacy in an Era of Self Driving
Cars.
Self driving or assisted driving cars are equipped with
360° cameras. They are constantly recording people and license plates
around them. However recording someone's location at different times
would be a privacy concern. This project looks at the potential of
adversarial machine learning to tackle this problem. It look at the
possibility of adding a perturbation (measured noise) while ensuring
model accuracy is not affected. Brief progress report.
Aditya PARAMESWARAN: Simplifying and Accelerating Data
Exploration.
Interactive data exploration tools are extremely popular;
as examples, the spreadsheet tool Microsoft Excel is used by nearly
10% of the world's population, while the visual analytics tool
Tableau was valued at 16B prior to its acquisition by Salesforce.
Despite their popularity across a spectrum of domains, it is still
challenging to use these tools to derive insights, especially on
large datasets that are increasingly the norm, leading to frustration,
missed opportunities and errors, and tedious manual effort. Drawing
on examples from spreadsheets and visual analytics, we will describe
our work on simplifying and accelerating data exploration, as well as
some takeaways from our experience.
Aditya Parameswaran is an Assistant Professor in the
School of Information (I School) and in Electrical Engineering and Computer
Sciences (EECS) at the University of California, Berkeley. Until June 2019,
Aditya was an Assistant Professor in Computer Science at the University
of Illinois, Urbana-Champaign. He spent a year as a PostDoc at MIT
CSAIL following his PhD at Stanford University. He develops systems
and algorithms for "human-in-the-loop" data analytics, synthesizing
techniques from database systems, data mining, and human-computer
interaction. His website is at http://adityagp.net.
Oct 25: Niloufar SALEHI and Ankit BANSAL.
Ankit BANSAL: Protecting World Leaders Against Deep
Fakes.
Brief progress report. Advances in deep learning i
have made it very easy to create sophisticated fake videos called
deep fakes. These pose a significant threat to democracy, national
security, and society. Researchers have developed forensic
techniques to typify an individual’s speaking pattern. However,
these methods need to scale, and I am working on different
techniques to create deepfakes, using which these models can
be improved.
Niloufar SALEHI: Intimate and Accountable
Reconfiguration of Social Platforms.
As online social spaces have evolved so have their
problems. I will discuss how online spaces may be reconfigured to
address those challenges. I will focus on two problems: First, social
pressure to perform perfection and second, online harassment and abuse.
Our research shows that one way that young people deal with mounting
social pressures is by reconfiguring online platforms and changing
their purposes, norms, expectations, and currencies. For instance, a
finsta (fake+Instagram account) provides an outlet to share emotional,
low-quality, or indecorous content. Carving out smaller spaces
accessible only to close friends allows finsta users the opportunity
for a more unguarded, vulnerable, and unserious performance. Drawing
on feminist theory, we have termed this process: intimate
reconfiguration. Through this reconfiguration finsta users repurpose
an existing and widely-used social platform to create opportunities
for more meaningful and reciprocal forms of social support.
Second, I will discuss how we might reconfigure online social spaces
to deal more effectively with online harm by relying on theories
of restorative and transformative justice. I will present a fictional
platform, The Internet's Good Place (TIGP), that we have developed through
participatory design with online moderators and transformative and
restorative justice practitioners. TIGP questions dominant approaches
to moderation online that focus on what "content" is not allowed and
presents paths to alternative approaches that rely on community
accountability.
Niloufar Salehi is an Assistant Professor at the
School of Information at UC, Berkeley. Her research interests are
in social computing, technologically mediated collective action,
digital labor, and computer supported cooperative work. Her work
has been published and received awards in premier venues in
human-computer interaction including CHI and CSCW. Through building
computational social systems in collaboration with existing
communities, controlled experiments, and ethnographic fieldwork,
her research contributes to the design of alternative social configurations
online. More at http://niloufar.org/.
Nov 1: Michael BUCKLAND: ABCs of Information: Content, Context,
Relevance, and other Challenges.
Useful information services can be provided without great
intellectual rigor or technical precision, but even a little critical
scrutiny reveals some basic, useful terms (such as content, context
and relevance) to be problematic conceptually and/or in
practice. Meanwhile technological changes continually generate
additional options. How could we evolve a more satisfactory tool-kit of
concepts and terms? Continuing past Seminar discussions, I will review
these and other less-known concepts and terms that could enable a
stronger and more coherent basis for information studies, including
affordances, autodocuments, documentarity, and particulars.
Join us for discussion.
Nov 8: Michael BUCKLAND & Clifford LYNCH.
Change of program. John Chuang's talk concelled.
Michael BUCKLAND: Incorporating Sense-making Into
Descriptions of Information Systems.
Both the theorizing of information and
representations of information systems and services
give prominent attention to information technology and logical flows.
But human sensing, sense-making, and becoming informed,
although obviously important, get less attention. I will propose how
this weakness could be remedied
with a more hermeneutic view of information systems.
Clifford LYNCH: Scenarios and Issues in "Send
Computation to the Data" Repository Models.
In recent years, there has been a resurgence of
interest in data repository models that propose to send the computation
to the repository rather than transferring data from a repository to
a computational environment. This is much more complex than it appears,
depending on the assumptions that are being made. I'll examine some
of the justifications and motivations for doing this, and outline the
issues involved at a high level. Topic to be announced.
Nov 15: Morgan AMES and Michele PRIDMORE-BROWN: Conversation on Morgan
Ames' new book The
Charisma Machine: The Life, Death, and Legacy of One Laptop per Child.
Morgan Ames's new book The Charisma Machine
chronicles the life and legacy of the One Laptop per Child project and
explains why--despite its failures--the same utopian visions that
inspired OLPC still motivate other projects trying to use technology
to “disrupt” education and development.
In this conversation between the author and LA Review of Books editor Michele Pridmore-Brown, we will explore why the project remained charismatic to many, even with (and even because of) its fundamentally flawed vision of who the computer was made for and what role technology should play in learning.
Drawing on an ethnography of a model OLPC project in
Paraguay, The Charisma Machine reveals that OLPC's laptops were
not only frustrating to use, easy to break, and hard to repair, they
were designed for “technically precocious boys”--idealized younger
versions of the developers themselves--rather than the diverse range
of children who might actually use them. Reaching fifty years into
the past and across the globe, this book offers a cautionary tale
about the allure of technology hype and the problems that result
when utopian dreams drive technology development.
For The Charisma Machine see mitpress.mit.edu/books/charisma-machine.
Morgan Ames is an assistant adjunct professor
in the School of Information and interim associate director of research
for the Center for Science, Technology, Medicine and Society at the
University of California, Berkeley, where she teaches in Data Science
and administers the Designated Emphasis in Science and Technology
Studies. She researches the ideological origins of inequality in the
technology world, with a focus on utopianism, childhood, and learning.
More at morganya.org.
Michele Pridmore-Brown is the science editor of
the Los Angeles Review of Books. She holds a Ph.D. in Comparative
Literature with an emphasis on the history of science. She has
written and taught on subjects like Virginia Woolf and radio
technology, gender and science, designer babies and the history and future of reproductive technologies. Her work appears in scholarly and popular venues. She is a research fellow with CSTMS.
More at cstms.berkeley.edu/people/michele-pridmore-brown-2/.
Nov 22: Ankit BANSAL & Mugdha BHUSARY: Final progress reports.
Ankit BANSAL: Protecting World Leaders Against Deep
Fakes.
The creation of sophisticated fake videos has been
largely relegated to Hollywood studios or state actors. However,
advances in deep learning have made it very easy to create sophisticated
fake videos called deep fakes. These pose a significant threat to
democracy, national security, and society. Researchers have developed
forensic techniques to typify an individual’s speaking pattern.
However, these methods need to scale. So I am working on a technique
to automate the creation of deepfakes in order to provide a better
test environment for forensic technique developers.
I will also be addressing the current challenges in the creation
of these videos.
Mugdha BHUSARY: Managing Pedestrian Privacy in an
Age of Self Driving Cars.
Self driving or assisted driving cars are equipped
with 360° cameras. They are constantly recording people and license
plates around them. However using the protected information of
people's faces and license plates for a task is not acceptable.
Through the time of this course, a project has been developed that
would look at the potential of differential privacy and adversarial
machine learning to tackle this problem. It would evaluate the
possibility of adding a perturbation (measured noise) while
ensuring model accuracy is not affected. Further, once we realise
the potential of building such a model, the project would tackle
the question of where and how the notification should be implemented.
Nov 29: Thanksgiving. No Seminar meeting.
Dec 6: Clifford LYNCH: Developments in Networked Information:
Review and Prospects.
This is the now-traditional extended preview of parts
of the plenary talk that I'll give at the CNI fall member meeting
in the week of December 9. It will cover recent developments that I'm
tracking, and discuss near-term prospects and opportunities.
Since we have more time than I will at the CNI plenary and the
scope of this talk is more constrained, I'll be able to go into
more depth on some of the topics.
Dec 13: Niall KELEHER: Demographics, Social Networks, and Mobile
Phone Usage: Introduction and Take up of Community Cellular Networks
in the Philippines.
In the context of remote locations of the Philippines,
I investigate the determinants of cellular network adoption. I
leverage unique circumstances where all households in seven localities
were interviewed before the installation of cellular network towers.
I link rich socio-economic as well as social network data with call
detail records from the initial months of cellular network service to
shed light on the drivers of network usage. The results contribute
to the literature on cellular networks by revealing household-level
characteristics that should be taken into consideration as “last-mile”
ICT interventions are considered.
Niall Keleher is a Postdoctoral Researcher at
the Stanford Graduate School of Business. His background is in
development economics and research methods. His current research
focuses on the application of machine learning to improve the targeting
of public and private services. Niall completed his Ph.D. at the UC
Berkeley School of Information. More at http://www.nkeleher.com.
Spring 2019 schedule and summaries.
Spring 2020 schedule and summaries.