Love and Authentication One of the most commonly neglected security vulnerabilities associated with typical online service providers lies in the password reset process. By being based on a small number of questions whose answers often can be derived using data-mining techniques, or even guessed, many sites are open to attack. To exacerbate the problem, many sites pose the very same questions to users wishing to reset their forgotten passwords, creating a common "meta password'' between sites: the password reset questions. At the same time, as the number of accounts per user increases, so does the risk for the user to forget her password. Unfortunately, the cost of a customer-service mediated password reset is far beyond possible for most service providers. In this talk, an alternative technique will be presented. It is fast and efficient, is compatible with input-constrained devices (such as handheld devices), and has low error rates. More about PARC's security work...
Heuristic Search for Target-Value Path Problem We define a class of combinatorial search problems in which the objective is to find a set of paths in a graph whose elements’ value is as close as possible to some target value. Unlike the usual shortest path problem, the goal is not necessarily to find paths with minimum length. We show that in most cases it is possible to decompose the problem into components where heuristic search can be used. We demonstrate the benefits of this approach on a synthetic domain and illustrate an instantiation of the approach for a problem in model-based diagnosis. More about PARC's intelligent control & autonomous systems work...
Certainly, software engineers require abstractions of objects in the world and the processes that animate them when designing and developing computer systems. Indeed, a design is an abstraction, a formal representation or model that serves as a kind of plan for the builders. But does this mean that when the objects of interest are human beings and their behavior – the social order nicely described by SEI as the ‘ecosystem’ for any and all computing – that researchers who study humans, social scientists in particular, ought to then have similar kinds of abstractions and formal models as their own goal, as a basic requirement of their science? Most social and behavioral scientists would answer in the affirmative, arguing without an analytically stipulated conceptual scheme or model from which to proceed (and to which to return), any orderliness in the vast plentitude of human beings’ lived experience cannot be found (seen), let alone explained. The social science tradition that informs my work, ethnomethodology, provides a different answer, however, one based on a rather different perspective on the organization of human behavior. This alternate perspective also directs our attention to humans’ own understandings of their actions: How do they make sense of their world, display this understanding to others, and produce the mutually shared social order in which they live – and in which the computer systems we design and build for them must operate? In addressing these questions, I very briefly describe the ethnomethodological approach, discuss the problem of representation and abstraction in the analysis of behavior, and argue about why this matters for system design.More about PARC ethnography...
IEEE LEOS 2008 Summer Topicals
21-23 July 2008, Acapulco, Mexico Peter Kiesel, Markus Beck, Michael Bassler, Noble Johnson
Micro-fluidic-based optical detection platform for characterizing fluorescing objects with integrated wavelength detection
We describe a compact, low-cost analyte detection platform that combines a fluidic channel, large area fluorescence excitation and on chip wavelength detection. The unit is optimized to record native fluorescence spectra from moving analytes. More about PARC's optical detector systems work...
High Throughput Membrane-less Water Purification This paper describes a highly scalable fluidic technology that presents a transformative approach to the practice of conventional water treatment. Features include: a high throughput, purely fluidic, continuous flow, membrane-less, size selective method for particulate extraction; and accelerated agglomeration kinetics from mixing and transporting chemicals and raw water in confined channels... More about PARC's water technology work...
An integrated approach to qualitative model-based diagnosis This work extends model-based diagnosis to systems which convert, move, and process material. Examples of such systems are printers, refineries and food processing plants. Such plants present two challenges to model-based diagnosis: (1) the plant may process 100s-1000s of items per minute so retaining full details of behavior of all past objects is impractical, and (2) complex multi-way interactions can occur among components operating on the same object. We address the first challenge by synopsizing past behavior in a data structure of fixed size. We address the second challenge by introducing the notion of interaction fault which represents the situation where a set of components operating on the same object damage the object even though each component alone produces no noticeable damage. Introducing interaction faults is much simpler than introducing fine-grained models of component object interactions. We demonstrate the approach on a highly redundant printer.
More about PARC's intelligent control & autonomous systems work...
Type-checking in Formally Non-typed Systemsguage Understanding: Question answering and textual 'entailment' Type checking defines and constrains system output and intermediate representations. We report on the advantages of introducing multiple levels of type checking in deep parsing systems, even with untyped formalisms. More about PARC's natural language work...
Low-Cost Membrane-less Water Purification
Conventional municipal water treatment is a multi-stage process, requiring large amounts of land and/or expensive filtration processes. PARC has developed a compact, low-cost purification system that reduces maintenance down-time, increases throughput, decreases power consumption, and shrinks real estate requirements. Alternate applications exist for the core IP. Membrane filtration (MF) is the state of the art for removing particles from water, but it suffers from clogging and frequent down-time to back flush. Cyclone separators require sedimentation which results in long processing times and cannot address particles smaller than 70mm depending on transit time. Our technology will separate particles down to the 1-5um size range, does not need membranes, can address neutrally buoyant particles, and can operate continuously. Water utilities, membrane suppliers, and water solution providers will be potential clients. Advantages include: minimal clogging as our approach is self-cleaning; low cost of manufacturing and direct integration with existing pumping and plumbing infrastructure; lighter TSS loading on membranes elsewhere in the system; replacement for intermediate filtration steps; small footprint with elimination of flocculation and sedimentation – particularly valuable for growing capacity of existing facilities in urban areas; reduced coagulant dosage and elimination of most flocculant cost; and faster processing or reduced time between raw and finished water leading to reduced holding capacity. More about PARC's water technology work...
Membrane-less Filtration for Conventional Potable Utility Applications
This talk describes an innovative highly scalable fluidic technology that presents a transformative approach to the practice of conventional water treatment. Features include: a high throughput, purely fluidic, continuous flow, membrane-less, size selective method for separation of neutrally buoyant suspensions; and accelerated agglomeration kinetics from enhanced mixing and high fluidic shear resulting in reduced coagulant chemical dosage by 30-50%. Together, the combined effects allow for... More about PARC's water technology work...
Crimeware: What lies ahead?
Malware these days is pushed by organized crime, aided by phishing-like deceit tactics, and spread via advertisements, social networks and shrink-wrapped electronics. It captures keystrokes of individuals, spies on corporations and politicians, and threatens our national security by means of server takeovers, information leakage, and a potential deterioration of trust in the infrastructure. We no longer call it malware - we call it crimeware. Crimeware can only survive if attacks can be monetized, and is suppressed by the risk of traceability and legal action. The traditional form of monetization relies on stealing credentials... In this talk, we present and explain yet another, not yet publicized, attack type that would let absolutely untraceable attackers to cash out huge amounts, at the cost of tremendous damages to victim corporations. Discussion of this new kind of attack will allow us to look into the future and consider what new countermeasures are needed. More about PARC's security & privacy work...
Reaching for Seamless Interaction in Information Environments The overarching goal of PARC's research in ubiquitous computing today is to achieve an environment of "seamless interaction" -- where people control information environments as a seamless whole as opposed to managing collections of networked components. We take a three-front strategy toward this objective: Ubiquity, Natural Interaction, and Proactivity. Hardware trends continue to drive the ubiquity of computation as miniaturization and the digitalization of media creates technologies that are increasingly portable as well as embedded in buildings, appliances, furniture, clothing and more. New technologies create ways for people to use more natural interaction styles that match the embodied ways we interact with the physical world. Sensor systems have made it possible for systems to become proactive in filtering, sorting, and presenting information, even sometimes taking action that is appropriate to the situation. This presentation describes several projects in these research thrusts and reflect on some of the lessons learned for user-centered methods for the design of pervasive, behavior-adapting applications. More about PARC's ubiquitous computing work...
Advances in Biodefense Technology
7-8 May 2008, Barcelona, Spain
Peter Kiesel, Markus Beck, Michael Bassler, Noble Johnson, & Oliver Schmidt
On-the-flow pathogen characterization based on native fluorescence detection
Native fluorescence spectroscopy is promising for pathogen detection since it is sensitive and requires neither specific binding nor tagging. Our compact detection platform combines fluidic channels with chip-size spectrometers and records fluorescence from analytes as they traverse the channel. More about PARC's optical detector systems work...
Enabling Extensibility of Sensing Systems through Automatic Composition over Physical Location
Networked sensing systems are increasingly adopted in many applications, but today's systems are generally single purpose and hard to extend. This paper addresses the problem of enabling developers to develop extensible networked sensing systems. We propose a design methodology, which centers on a novel automatic composition service where the sensor processing software modules are parameterized by a physical location region. The automatic composer automatically configures the processing and communication occurring in a networked sensing system based on up-to-date sensing needs and sensor device availability. Our approach also enables adaptability and robustness against sensor failures. Related PARC work...
“Lifting the Veil: Improving accountability and social transparency in Wikipedia with WikiDashboard”
“Augmented Information Assimilation: Social and algorithmic web aids for the information long tail”
“Keyhole Tagging: Selective sharing in close collaboration”
“Escape: A target selection technique using visually-cued gestures”
“Crowdsourcing user studies with Mechanical Turk”
“Activity-Based Serendipitous Recommendations with the Magitti Mobile Leisure Guide”
“Responsive Mirror: Fitting information for fitting rooms” -- at workshop on Surrounded by Persuasive Ambient Intelligence
"Social Information Foraging and Sensemaking"; "We digital sensemakers"; "Tracing the microstructure of sensemaking" -- at Sensemaking workshop
"Cognitive engineering and the psychology of human-computer interaction" -- at Invited Season: Celebrating the 25th anniversary of The Psychology of Human Computer Interaction (by Stuart K. Card, Thomas P. Moran, Allen Newell)
Compact, microfluidic-based detection platform for on-the-flow analyte characterization PARC has developed various key technologies which are essential for on-chip optical detection system. This talk will give a brief overview of these technologies and will then focus on our work in on-the-flow pathogen detection based on native fluorescence spectroscopy. This is a very promising approach that does not require specific binding or tagging of the analyte. However, the variety of cells is large compared to the number of basic molecular building blocks. Therefore, the fluorescence spectra of different species are often very similar, and sophisticated detection methods are required to reveal differences. The specificity of this approach can be further improved by implementing high spectral resolution and using multiple excitation wavelengths. We have developed a compact platform that combines... More about PARC's optical detector systems work...
FTC roundtable discussion on phishing education We have been invited by the U.S. Federal Trade Commission to share insights about phishing, the online identity theft technique that uses deceptive spam to trick consumers into divulging sensitive information. The FTC’s goal is to identify ways to raise public awareness of phishing to help consumers protect themselves, giving them a "playbook" for recognizing and reacting to this malicious practice. More about PARC's security & privacy work...
Spectrum Fusion: Using Multiple Mass Spectra for De Novo Peptide Sequencing We report on a new algorithm for combining the information from several mass spectra of the same peptide. The algorithm automatically learns peptide fragmentation patterns, so that it can handle spectra from any instrument and fragmentation technique. We demonstrate the utility of the algorithm, and the power of multiple spectra, by showing that combining pairs of spectra (one CID and one ETD) greatly improves de novo sequencing success rates. More about PARC's bioinformatics work in this area...
We report on work on and with the English ParGram grammar (error mining, construction of semantic representations by means of term-rewriting rules, improved probabilistic disambiguation). More about PARC and ParGram...
Customized software for peptide identification by tandem mass spectrometry There are a number of different programs for identifying peptides by tandem mass spectrometry (MS/MS), including Mascot, SEQUEST, X!Tandem, Spectrum
Mill, Phenyx, and Paragon. A couple years ago, I made the beginner's mistake
of writing another such program, ByOnic. But now that I've got the program
and can quickly customize it for new types of data, I am finding that this
enables some interesting collaborations. I will talk about three such projects: data-independent MS/MS, alcohol-induced carbonylation, and oxidative surface mapping using laser-induced oxidation. More about PARC's bioinformatics work in this area...
Computing Linguistically-based Textual Inferences
This talk provides an overview and a demo of PARC's Bridge system. The particular task that we focus on is entailment and contradiction detection (ECD), a more refined variant of the PASCAL RTE (Recognizing Textual Entailment) challenge. Given a passage of text and a query, does the query sentence follow from the text in the passage, is it contradicted by it, or neither? More about PARC's natural language work...
Control of Large-Scale Reconfigurable Systems
The rise in embedded computing, sensing and actuation is leading to the development of larger scale distributed systems in a variety of domains. These systems can have many appealing qualities, such as modularity - which confers benefits during both design time and run time - and reconfigurability - the ability to change the system structure for customization purposes, for repairs or in response to environmental conditions or component failure. To reap these benefits, however, the software designed for such complex systems must address a number of challenges in dynamically coordinating and controlling the distributed components. This presentation covers some of the challenges, solutions and lessons learned during the design and implementation of a prototype highly modular, reconfigurable printing system at PARC. More about PARC's intelligent control & autonomous systems work...
Class identification of pathogens based on native fluorescence spectroscopy on-a-chip
Native fluorescence spectroscopy is a promising approach for pathogen detection since it is sensitive and requires neither specific binding nor tagging of the analyte. Specificity can be met by a combination of multi-color excitation, collection of detailed spectral information and optimized evaluation techniques. Our approach achieves these requirements with a compact solution that includes on-chip native fluorescence spectroscopy. More...
Status and a Technology Roadmap for Document Recognition Intelligence agencies collect vast quantities of data from diverse sources including video, still images, audio recordings, and documents. To extract useful content to "connect the dots," they need to employ computer vision, speech recognition, and document recognition technology. This talk takes stock of the document image analysis portion of the challenge. In this talk, we: (1) survey the state of the art, exploring what works, and what doesn't; (2) consider the drivers and leading edge of both the commercial document recognition industry and current academic research; (3) touch on major trends and a dozen hot topics; (4) consider how and where progress gets made, and how to move it from the laboratory to applications; and (5) discuss major fundamental challenges to the field, and recommend process steps that would maximally leverage R&D investment. More about PARC's intelligent image analysis work...
Towards Language Understanding: Question answering and textual 'entailment' After a brief introduction about the role of questing answering and textual inferencing in language understanding, we discuss the PASCAL RTE challenge. We describe an experiment that was intended to evaluate how close the RTE challenge comes to representing the judgments of the 'man in the street'. We then decompose the textual inferencing task in real entailments and in more pragmatic notions and illustrate how the PARC XLE system handles both kinds of inferences. More about PARC's natural language work...
Computing linguistically-based textual inferences
A long-standing goal of computational linguistics is to build a system for answering natural language questions. A successful QA system has to recognize semantic relations between sentences. If the user would like to know the answer a question such as Did Shackleton reach the South Pole?, the system should recognize that the sentence Shackleton failed to reach the South Pole contains the answer. None of the current search engines is capable of delivering a simple NO answer in such cases. The system described in this talk does make the correct inference. It is the Bridge system (a bridge from language to logic) developed at PARC. More about PARC's natural language work...
A Content-Driven Access Control System
Protecting identity in the Internet age requires the ability to go beyond the identification of explicitly identifying information like social security numbers, to also find the broadlyheld attributes that, when taken together, are identifying. We present a system that can work in conjunction with natural language processing algorithms or user-generated tags, to protect identifying attributes in text. The system uses a new attribute-based encryption protocol to control access to such identifying attributes and thus protects identity. The system supports the definition of user access rights based on role or identity. We extend the existing model of attributebased encryption to support threshold access rights and provide a heuristic instantiation of revocation. More about PARC's security & privacy work...
Regional Institutions and Their Impact on the Innovation Environment in Silicon Valley The history of innovation in Silicon Valley and the greater San Francisco Bay Area is frequently explained through the history of the people and companies who made the major technological advances and their personal and organizational triumphs and failures. What is often ignored or relegated to the background is the role that regional institutions played in nurturing the environment and culture in which these people and companies operated and how that role has evolved over time... PARC Vice President of Business Development John Knights examines the evolving roles of these institutions in light of the concurrent globalization of the high technology industry. More about PARC innovation milestones...
The Magitti* Activity-Aware Leisure Guide: Opportunity Discovery, Innovation and New Technology Platform Development at PARC We describe an example of PARC's innovation services for corporate clients. The project was undertaken for Dai Nippon Printing Co. Ltd. (DNP), to assist them in developing a new opportunity beyond their traditional printing business. The Magitti solution*, developed at PARC in close collaboration with DNP personnel, was designed to be synergistic with DNP's existing strengths in the publishing industry whilst incorporating the latest in context- and activity-aware computing techniques to recommend published content. We explain our market and opportunity discovery fieldwork and innovation exercises, as well as the system components and user experience and an early field evaluation. We also discuss ways in which new techniques and technologies were successfully transferred to DNP. More about PARC client services and opportunity discovery through ethography...
*Magitti is an electronic mobile leisure guide that presents options for things to do, filtered by how well they match current interests. Users don't have to tell Magitti what they are doing; it uses an inference engine to figure this out for itself. Interests are inferred from time, location, past behavior, and predicted activity type. Taste profiles and preferences can be dynamically adjusted to further improve recommendations. Over time, Magitti learns from behavior to make its recommendations more personally targeted.
"What’s Next?" panel moderated by Wall Street Journal columnist Lee Gomes
In recent years, VCs have thrust themselves into the worlds of China, India, Web 2.0, and cleantech looking for the next great deal. But what is next on the horizon? Is it perhaps Eastern Europe, Web 3.0, or space? Looking into their crystal balls to give their thoughts: PARC Director and President Mark Bernstein; Prith Banerjee, SVP Research & Director HP Labs; Geoffrey Moore, Venture Partner, Mohr Davidow Ventures; and Vincent Pluvinage, General Manager, Strategic Alliance and Private Equity Partnerships, Intellectual Ventures. More about PARC client services...
Human Computer Interaction Consortium[members only] 30 January-3 February 2008, Frasier, Colorado
Peter Pirolli, Ed Chi
Beyond Information Foraging to Ecologies of Sense Making The LATEST (Learning Actively about Topics in Emerging Science and Technology) project combines psychological research with technology research in human-information interaction. LATEST involves a three-prong approach to understanding how expertise can be transferred to active learners. First, we are developing a model of expert information foragers. Second, based on this model, we are developing specific tools and interaction techniques to support more expert-like performance in active learners. A large part of the technical research will focus on Web-based social computing. Third, we developing a framework for the evaluation of social sensemaking tools to evaluate these technologies. More about PARC's web-based social computing work...
Computing linguistically-based textual inferences In this talk we give an overview and a demo of PARC's Bridge system. The particular task that we focus on is entailment and contradiction detection, a more refined variant of the PASCAL RTE (Recognizing Textual Entailment) challenge. Given a passage of text and a query, does the query sentence follow from the text in the passage, is it contradicted by it, or neither? More about PARC's natural language work...
Growth beyond the core: fusing inside and outside knowledge to reach new markets presentation available [.pdf] As Open Innovation has been matured, the concept itself has expanded. Today, the most progressive companies have recognized the advantage of looking outside not only for technology, but also for complementary expertise to help them create or enter new markets. PARC, known for its role in creating much of modern computing, has been deep in the trenches of expertise-based relationships in the last six years. This talk will discuss the nuts-and-bolts of PARC's recent work with both multinational corporations and new ventures. The cases will include a summary of the important lessons for making expertise-based engagements successful, particularly in developing new market opportunities. It will also touch on PARC's own transition from corporate research lab to cross-industry business catalyst. More about PARC client services...
Computing linguistically-based textual inferences In this talk we give an overview and a demo of PARC's Bridge system. The particular task that we focus on is entailment and contradiction detection, a more refined variant of the PASCAL RTE (Recognizing Textual Entailment) challenge. Given a passage of text and a query, does the query sentence follow from the text in the passage, is it contradicted by it, or neither? More about PARC's natural language work...
An Intelligent Fitting Room Using Multi-Camera Perception
We describe the architecture of the vision system for the Responsive Mirror, a novel system for retail fitting rooms that enables online social fashion comparisons in physical stores based on multi-camera perception. This vision system provides implicitly controlled real-time interaction for “self” and “social” clothing comparisons by automatically tracking user’s motion as she tries on clothes. We describe the key components of the motion-tracking and clothes-recognition systems and evaluate their effectiveness against images collected during a previous user study and a dataset of images representing content from a social fashion network. More about PARC's ubiquitous computing work...
Mobile Recommendations for Leisure Activities
We demonstrate a context-aware mobile system for recommending information about leisure activities (Shopping, Eating, Doing, Seeing, and Reading), codenamed Magitti, which infers the user’s leisure activity from context and patterns of behavior. Magitti filters a database of city-guide-style leisure information to find the most relevant items based on the user’s profile, history, context, and predicted activity. Users can also customize the profile or dynamically adjust the current preferences if they wish to improve the recommendations further. More about PARC's context-aware and ubiquitous computing work...
All Additive Printing of Flexible Backplanes for Large Area Displays Methods used to deposit and integrate solution-processed materials to fabricate thin-film transistor (TFT) backplanes by ink-jet printing are presented. The materials studied allow the development of an all-additive process in which materials are deposited only where their functionality is required. We demonstrate successful integration of a complete additive process with the fabrication of simple prototype TFT backplanes on glass and on flexible plastic substrates, and we discuss the factors that make the process possible. More about PARC's large-area electronics research...
Planar Compound Microlenses for Particle Detection in Biomimetic Flicker Spectroscopy Water is a vital part of human life and a resource to be protected for national defense and security. The National Research Council has recommended that public water supplies be tested for dangerous contaminants, and that sensor systems be created to continuously monitor the water supply to maintain public safety and confidence. We propose a new biomimetic sensor technology for continuous monitoring of water systems for agent particulates using micro-optical-electro-mechanical (MOEMS) techniques. More about PARC's particle manipulation work...
Toward Seamless Interaction in Information Environments The early vision of Ubiquitous Computing is no longer a dream but an all too ubiquitous problem. Yesterday's vision of widespread computation, information, media, and communication has today become a reality where managing and controlling these many types of information services, devices, and applications is becoming impossible. Today's technological reality also creates opportunities for new kinds of applications and services that integrate information across the "virtual" and "real" aspects of one’s life. The overarching goal of PARC's research in ubiquitous computing today is to achieve an environment of "seamless interaction" – where people control information environments as a seamless whole, as opposed to managing collections of networked components. We take a three-front strategy toward this objective: 1) lightweight "at-hand" interaction techniques and systems that react to implicit user actions; 2) infrastructures and platforms that reduce technical boundaries; and 3) user behavior modeling to anticipate user goals and adapt systems to user needs. More about PARC's ubiquitous computing work...
KMWorld & Intranets
6-8 November 2007, San Jose, California
Ed Chi & Lawrence Lee
Web 2.0 in the Enterprise While the benefits of Web 2.0 social software for greater collaboration, innovation, and knowledge sharing are often discussed, the coordination and interaction costs that occur in social systems are often overlooked. Based on extensive studies of social systems such as del.icio.us and Wikipedia, we have identified a number of factors that need to be managed to realize the full benefits of these systems within the enterprise. More about PARC's augmented social cognition work...
