PowerPoint Presentation - Lecture

Document Sample
PowerPoint Presentation - Lecture Powered By Docstoc
              fdm 20c introduction to digital media
                            lecture 27.05.2003

warren sack / film & digital media department / university of california, santa cruz
last time
• fcc
  – who is the chair and who are the commissioners of
    the fcc?
  – what does the fcc do?
• surveillance
  –   history
  –   art
  –   technology
  –   legislation
• capture
  – history (as extension of taylorism/fordism)
  – comparison with surveillance
  – cycle or process

• review of the capture model
• definition of privacy
  – private versus public
     • civil versus economic
          – capture
  – efficient connections versus resistances
  – on the virtue of inefficiencies
• lessig on monitoring and search
  – example: monitoring on the web
  – example: search on the web
• gandy on data mining
• are there citizen-centered forms of data mining?

• close watch kept over someone or something
• Etymology: French, from surveiller to watch
  over, from sur- + veiller to watch, from Latin
  vigilare, from vigil watchful
surveillance model versus capture model

• surveillance model: is built upon visual
  metaphors and derives from historical
  experiences of secret police surveillance

• capture model: is built upon linguistic metaphors
  and takes as its prototype the deliberate
  reorganization of industrial work activities to
  allow computers to track them [the work
  activities] in real time

  – agre, p. 740
capture (in comparison with surveillance)

• linguistic metaphors (e.g., grammars of action)
• instrumentation and reorganization of existing
• captured activity is assembled from
  standardized ―parts‖ from an institutional setting
• decentralized and hetrogeneous organization
• the driving aims are not necessarily political, but
  philosophical/market driven
five stage cycle of grammars of action

•   analysis
•   articulation
•   imposition
•   instrumentation
•   elaboration
    – agre, p. 746-747
taylorism, fordism and grammars of action

         ford assembly line circa 1925
grammars of action/winograd & flores
political economy of capture

• ― imposing a mathematically precise form
  upon previously unformalized activities, capture
  standardizes those activities and their
  component elements and thereby prepares
  them for an eventual transition to market-based
  – agre, p. 755
privacy: a definition

• 1.
   – a. the quality or state of being apart from company or
   – b. SECLUSION: freedom from unauthorized intrusion
     <one's right to privacy>
• 2. archaic : a place of seclusion

• source: Merriam Webster
privacy: a culturally specific definition

• Does the U.S. Bill of Rights define an
  individual’s ―right to privacy‖?
• Not explicitly, but...
  – inferrably: e.g., Amendment IV: The right of the
    people to be secure in their persons, houses, papers,
    and effects, against unreasonable searches and
    seizures, shall not be violated, and no Warrants shall
    issue, but upon probable cause, supported by Oath
    or affirmation, and particularly describing the place to
    be searched, and the persons or things to be seized.
  – implicitly: e.g., Amendment IX: The enumeration in
    the Constitution, of certain rights, shall not be
    construed to deny or disparage others retained by
    the people.
what’s missing from this picture?

    private                 public
 what are the connections between
 the public and the private?

private               public          state

           social     civil society     economic sphere

see writings by hegel, arendt, gramsci, etc.
e.g., hegel: ―civil society‖ as the domain of rights and
  freedoms guaranteed by the state;
gramsci on the disctinction between civil society and
  economic sphere
agre’s capture model

• ―capture‖ as described by agre can be
  understood as the means used to seemlessly
  and efficiently connect that part of the ―public‖
  known as ―civil society‖ or simply the ―social‖ to
  the ―economic sphere‖ or, more specifically, to
  economic productivity
• is this sort of efficiency a good thing?
• if so, for whom?
resistances between private and public

         private                  public

what divides the private from the public?
what reduces the efficiency of the connections
 between private and public?
lessig on the merits of inefficiency

• ―I am arguing that a kind of inefficiency should
  be built into these emerging technologies — an
  inefficiency that makes it harder for these
  technologies to be misused. And of course it is
  hard to argue that we ought to build in features
  of the architecture of cyberspace that will make
  it more difficult for government to do its work. It
  is hard to argue that less is more.‖
   – Lessig, p. 19
lessig on inefficiency (continued)

• But though hard, this is not an argument
  unknown in the history of constitutional
  democracies. Indeed, it is the core of much of
  the design of many of the most successful
  constitutional democracies — that we build into
  such constitutions structures of restraint, that
  will check, and limit the efficiency of
  government, to protect against the tyranny of
  – Lessig, p. 19
gandy on the merits of inefficiency

• mining systems are designed to facilitate
  the identification and classification of individuals
  into distinct groups or segments. From the
  perspective of the commercial firm, and perhaps
  for the industry as a whole, we can understand
  the use of data mining as a discriminatory
  technology in the rational pursuit of profits.
  However, as a society organized under different
  principles, we have come to the conclusion that
  even relatively efficient techniques should be
  banned or limited because of what we have
  identified as unacceptable social consequences
   – Gandy, pp. 11-12
digital media versus computer science

• digital media studies: some architectures (e..g.,
  democratic ones) are best designed to be
• computer science: efficiency is almost always
  considered to be a virtue: efficient architectures
  are usually good architectures
lessig on architecture

• however, by ―architecture‖ lessig means, more
  or less, what computer scientists mean when
  they say architecture:
  configuration/aseemblages of hardware and
lessig on code and architecture

• The code of cyberspace -- whether the Internet, or net within the
  Internet -- the code of cyberspace defines that space. It constitutes
  that space. And as with any constitution, it builds within itself a set
  of values, and possibilities, that governs life there ... I've been
  selling the idea that we should assure that our values get
  architected into this code. That if this code reflects values, then we
  should identify the values that come from our tradition -- privacy,
  free speech, anonymity, access -- and insist that this code embrace
  them if it is to embrace values at all. Or more specifically still: I've
  been arguing that we should look to the structure of our
  constitutional tradition, and extract from it the values that are
  constituted by it, and carry these values into the world of the
  Internet's governance -- whether the governance is through code,
  or the governance is through people.
• Open Code and Open Societies: Values of Internet Governance
  Larry Lessig (1999)
lessig on architecture of privacy

• Life where less is monitored is a life more
  private; and life where less can (legally perhaps)
  be searched is also a life more private. Thus
  understanding the technologies of these two
  different ideas — understanding, as it were,
  their architecture — is to understand something
  of the privacy that any particular context makes
  – Lessig, p. 1
architectures of privacy

• from doors, windows and fences
• to wires, networks, wireless networks,
  databases and search engines
example: architecture of the web

• examples of (anti)monitoring architectural
  features of the web
  – HTTP headers
     • cookies
  – encryption
• example of searching on the web
  – try ―googling‖ yourself
monitoring on the web

• what does your web browser reveal about you?
• standard HTTP headers:
  –   From: User’s email address
  –   User-Agent: User’s browser software
  –   Referer: Page user cam from by following a link
  –   Authorization: User name and password
  –   Client-IP: Clien’t IP address
  –   Cookie: Server-generated ID label

• cookies are information that a web server stores
  on the machine running a web browser
  – try clearing all of the cookies in your web browser
    and the visit the site

• symmetric key encryption

• public key encryption
search/elaboration/data mining

•   what lessig calls search,
•   what agre calls elaboration, and
•   what gandy discusses as data mining
•   are converging concerns about the production
    of a permanent, inspectible record of one’s non-
    public life and thus a shrinking in size and kind
    of one’s private life
searching on the web

• search engines make many things (sometimes
  surprisingly) public
agre on “elaboration”

• ―The captured activity records, which are in
  economic terms among the products of the
  reorganized activity, can now be stored,
  inspected, audited, merged with other records,
  subjected to statistical analysis, ... and so forth.‖
   – p. 747
“data mining” is one form of “elaboration”

• gandy (p. 4) on ―data mining‖: mining is an applied statistical technique.
  The goal of any datamining exercise is the
  extraction of meaningful intelligence, or
  knowledge from the patterns that emerge within
  a database after it has been cleaned, sorted
  and processed....
goals of data mining

• In general, data mining efforts are directed
  toward the generation of rules for the
  classification of objects. These objects might be
  people who are assigned to particular classes or
  categories, such as ―that group of folks who
  tend to make impulse buys from those displays
  near the check out counters at the
  supermarket.‖ The generation of rules may also
  be focused on discriminating, or distinguishing
  between two related, but meaningfully distinct
  classes, such as ―those folks who nearly always
  use coupons,‖ and ―those who tend to pay full
  price.‖ Gandy, p. 5.
types of data mining

• descriptive: compute a relatively concise,
  description of a large data set

• predictive: predict unknown values for a variable
  for one or more known variables
  – e.g., will this person likely pay their bills on time?
data mining tasks

•   regression
•   classification
•   clustering
•   inference of associative rules
•   inference of sequential patterns
data mining task

• regression: infer a function that relates a known
  variable to an unknown variable
  – e.g., advertising: how much will sales increase for
    every extra $1000 spent on advertising?
data mining task

• classification: given a set of categories and a
  datum, put it into the correct category
   – e.g., direct-mail marketing: given a person’s zip code,
     age, income, etc. predict if they are likely to buy a
     new product
data mining task

• clustering: given a data set divide it into groups
   – e.g., segmenting customers into markets: given a set
     of statistics (e.g., age, income, zip code, buying
     habits) about a large number of consumers, divide
     them into markets; e.g., ―yuppie,‖ ―soccer mom,‖ etc.
data mining task

• inference of sequential patterns: given a set of
  series, determine which things often occur
  before others
  – e.g., predicting a customer’s next purchase:
    determine which products are bought in a series;
    e.g., bookstore: intro to spanish 1, intro to spanish 2,
    don quixote; e.g., nursery: grass seed, fertilizer, lawn
data mining task

• inference of associative rules: given a set of
  sets, determine which subsets commonly occur
  – e.g., supermarket layout: given a database of items
    customers have bought at the same time, determine
    which items should adjacent in the store; e.g., if
    diapers and milk are often bought with beer, then
    place the beer next to the milk.
  – e.g.,’s ―people who bought this book
    also bought...‖
  – Amazon’s feature is an example of a ―recommender
    system‖ or a ―collaborative filter‖
how do recommender systems work?
an example algorithm
• Yezdezard Lashkari, Feature Guided
  Automated Collaborative Filtering, Masters
  Thesis, MIT Media Laboratory, 1995.
• Webhound
• Firefly
All automated collaborative filtering algorithms use the following steps
to make a recommendation to a user (Webhound, Lashkari, 1995):
data mining applications

• data mining is used for
  – market research and other commercial purposes
  – science (e.g., genomics research)
  – intelligence gathering (e.g., identification of
    ―suspects‖ by ―homeland security‖)

• might data mining be used for the purposes of
  less powerful citizens? e.g.,
  – news analysis (cf, the function of FAIR)
  – government ―watch dog‖ operations (cf., Amnesty
technologies and architectures of privacy

• technologies and architectures are important
  influences on the production and change of
  private and public space;
• but, they do not independently determine what
  is public and what is private (to think they do is
  called technological determinism)
• we need to understand not just the machines,
  but also the people mediated by these
  technologies: we need to understand the whole
  as a machination, a heterogeneous network of
  people and machines
architectures and inefficiencies

• sometimes inefficient architectures, inefficient
  technologies are good technologies because
  they allow for or facilitate resistance by the less
  powerful in the face of powerful individuals,
  corporations and governments
next time

• open source software