Why PrivacySOS.org ? | Privacy SOS

http://privacysos.org

  • Facebook-DeepFace: Closing the Gap to Human-Level Performance in Face Verification

    Abstract
    In modern face recognition, the conventional pipeline consists of four stages: detect => align => represent => classify. We revisit both the alignment step and the representation step by employing explicit 3D face modeling in order to apply a piecewise affine transformation, and derive a face representation from a nine-layer deep neural network.
    https://www.facebook.com/publications/546316888800776

  • Missing the point: PERF and the surveillance industrial complex | Privacy SOS
    http://privacysos.org/node/376

    Intelligence-led vs. community-policing

    During the 1990s, a powerful movement for community control and oversight of police practices shook the law enforcement world. After decades of failed war on drugs policies that inflicted severe punishments and harsh police crackdowns on poor communities, law enforcement finally began to listen to organizers from those criminalized groups. Policing literature from the 1990s is full of references to ‘community policing’, a theory that prioritizes community-officer engagement, trust and mutual respect as the building blocks necessary to build and maintain safe cities. Indeed, when PERF opposes SB 1070 and like laws, it does so using the language of community policing and it is true to the ideology: local immigrant communities must trust their police departments, or else crimes won’t be reported, making cities more dangerous for everyone. That’s correct and it’s sound public policy.
     
    But then September 11, 2001 came and went, and with it went the federal government’s investment in the community-policing paradigm. All of a sudden, the community’s concerns and the importance of building trust among officers and locals became irrelevant; what mattered was getting all of the information possible, to attempt to pre-empt crime, to prevent it from happening.
     
    Everyone became suspect, so there was no one in the community left to trust. The new ideology was fully articulated by many in the Bush administration, but not least by ‘Total Information Awareness’ champion and convicted felon John Poindexter. Still, police didn’t have the tools they needed to execute the new mission; the missing key was the technology that would enable police to become like the futuristic crime stoppers in the film Minority Report. With the right tools, the story goes, police from the FBI on down to the local sheriff could prevent crime by predicting it. Intelligence-led, or predictive policing was born.
     
    Luckily for companies like Booz Allen and Lockheed Martin, intelligence-led policing requires vast expenditures in surveillance technologies and information sharing architectures. Over the past ten years, the federal government has doled out seemingly unlimited amounts of cash to local cops who want the latest spy and info sharing tech. […]
     
    Meanwhile, the Department of Justice hasn’t forgotten about community policing; indeed, it has an entire funding stream called “Community Oriented Policing Services” (COPS). Upon close inspection, however, it becomes clear that instead of investing in programs that actually work towards the goals of that methodology, DOJ has simply been giving its so-called “Community Policing” grants to local cops to buy – you guessed it – surveillance tools. 
     
    That’s right: DOJ is now simply calling intelligence-led policing community policing. Maybe it hopes no one will notice? But it’s hard not to notice that community-policing grants are funding major surveillance technology expenditures on the local level. One example among many: Brockton, MA police got nearly $500,000 from DOJ this year via the “Secure Schools” program as part of a community policing funding stream. The money will be largely spent on new surveillance cameras and lighting, and a “state of the art” door lock system capable of putting the school into full prison-style lockdown.