About The Forum
Artificial Intelligence
Spotlight Research Projects
|
The Red Sea Robotics Research Exploratorium was created in April 2012 through a generous research award from the King Abdullah University of Science and Technology (KAUST)
|
|||
|---|---|---|---|
The Red Sea Robotics Exploratorium|
About
The Red Sea Robotics Research Exploratorium was created in April 2012 through a generous research award from the King Abdullah University of Science and Technology (KAUST). As a part of the KAUST Global Collaborative Research Program, Stanford University is part of a team of universities working to build a major science and technology university along a marshy peninsula on Saudi Arabia’s western coast. Meka Robotics joined the collaboration and provides the hardware for the development of dexterous underwater robot arms.
|
|||
|---|---|---|---|
The Red Sea Robotics Exploratorium|
This video production documents the life and career of Ed Feigenbaum, "Father of Expert Systems," through archival photographs, a Computer History Museum oral history, and the recollections of his collaborators and students. These recollections were videotaped at the Feigenbaum 70th Birthday Symposium, held on
March 25-26, 2006 and co-sponsored by the Stanford Computer Forum.
|
|||
|---|---|---|---|
Ed Feigenbaum's Search for A.I.|
ImageNet is an image dataset organized according to the WordNet hierarchy. Each meaningful concept in WordNet, possibly described by multiple words or word phrases, is called a "synonym set" or "synset". There are more than 100,000 synsets in WordNet, majority of them are nouns (80,000+). In ImageNet, we aim to provide on average 1000 images to illustrate each synset. Images of each concept are quality-controlled and human-annotated. In its completion, we hope ImageNet will offer tens of millions of cleanly sorted images for most of the concepts in the WordNet hierarchy.
|
|||
|---|---|---|---|
'ImageNet: A Large-Scale Hierarchical Image Database'|
Given an image, we propose a hierarchical generative model that classifies the overall scene, recognizes and segments each object component, as well as annotates the image with a list of tags. To our knowledge, this is the first model that performs all three tasks in one coherent framework. For instance, a scene of a ‘polo game’ consists of several visual objects such as ‘human’, ‘horse’, ‘grass’, etc. In addition, it can be further annotated with a list of more abstract (e.g. ‘dusk’) or visually less salient (e.g. ‘saddle’) tags. Our generative model jointly explains images through a visual model and a textual model. Visually relevant objects are represented by regions and patches, while visually irrelevant textual annotations are influenced directly by the overall scene class. We propose a fully automatic learning framework that is able to learn robust scene models from noisy web data such as images and user tags from Flickr.com. We demonstrate the effectiveness of our framework by automatically classifying, annotating and segmenting images from eight classes depicting sport scenes. In all three tasks, our model significantly outperforms state-of- the-art algorithms.
|
|||
|---|---|---|---|
Towards Total Scene Understanding:Classification, Annotation and Segmentation in an Automatic Framework|
What kind of tools would you need to make a functional interactive prototype of a media player in 30 minutes?
d.tools is a hardware and software system that enables designers to rapidly prototype the bits (the form) and the atoms (the interaction model) of physical user interfaces in concert. d.tools was built to support design thinking rather than implementation tinkering. With d.tools, designers place physical controllers (e.g., buttons, sliders), sensors (e.g., accelerometers), and output devices (e.g., LEDs, LCD screens) directly onto form prototypes, and author their behavior visually in our software workbench.
|
|||
|---|---|---|---|
d.tools: Enabling rapid prototyping for physical interaction design|
Who will be writing software in the future and how will they be doing it? As computing becomes increasingly important in people's work andhobbies, a much broader range of people are engaging in programming. Understanding and building tools for professional software developers has a long history, but there has been relatively little research on
how to support amateur, opportunistic programmers. Professor Scott R. Klemmer's NSF-funded research group at Stanford University is currently studying this problem.
So far, they have done fieldwork with exhibit designers at San
Francisco's Exploratorium Museum, and conducted several empirical
studies on how these programmers use information resources while
building software. Most notably, the Web has revolutionized the way
these individuals write software. They build entire applications by
iteratively searching for, understanding, and integrating pieces of functionality embodied in 15-line chunks of code!
Right now, Professor Klemmer's research group is building a number of tools to support amateur programmers that embody and support this reliance on Web resources. The broad goal of this work is to make software development faster, easier, and less error-prone for a much larger population.
|
|||
|---|---|---|---|
Opportunistic Programming|
Social software – email, blogs, wikis, forums, social networks – has revolutionized how people share expertise and collaborate on the web. However, in rural developing regions, many do not have direct access to Internet-connected PCs or the literacy skills to interact with textual content. How might we design a communications platform for these communities? In our research, we are designing voice-based applications for communities in rural India to access agricultural advice and share expertise, using the mobile phone. The key challenges are contending with limited capability speech recognition for regional languages, designing for illiterate users, and methods for search and filtering of user-generated audio content. We have deployed one pilot system for farmers in Gujarat, India, to record agricultural questions and get responses from experts and other farmers. Based on the enthusiastic response, the application will be launched later this year to serve over 500,000 farmers across the state.
|
|||
|---|---|---|---|
Voice-Based Social Media for Developing Regions|
An artist might spend weeks fretting over questions of depth, scale and perspective in a landscape painting, but once it is done, what's left is a two-dimensional image with a fixed point of view. But the Make3d algorithm, developed by Stanford computer scientists, can take any two-dimensional image and create a three-dimensional "fly around" model of its content, giving viewers access to the scene's depth and a range of points of view.
|
|||
|---|---|---|---|
Make3D|
Dryad is a design tool for virtual trees. The user creates a tree by visually exploring the space of all trees. Dryads around the world communicate to share which trees were picked in the past. A collaborative mapping of the tree space emerges, which every Dryad utilizes to guide users to higher-quality parts of the space.
|
|||
|---|---|---|---|
Dryad|
Wireless networks are convenient, increase productivity, and enable new services and applications. However, they are difficult to secure, configure, and manage. The falling prices of IEEE 802.11 hardware enables the use of higher number of access points, increasing overall network capacity and allowing services to scale to much higher number of users. In practice, however, such high-density deployments are challenged in many ways: (i) higher management costs, due to access point configuration and maintenance, (ii) increased security risks, due to misconfigured access points or improperly managed user credentials (passwords, private keys), (iii) potentially higher interference levels, which can considerably degrade performance.
The objective of the KIWI project is to build self-managed wireless LANs that scale to high number of access points and clients. We explore the capabilities of wireless LANs managed by a wireless appliance (WA), a device that controls all access points (APs). The WA gathers information about all clients in the network, being able amongst other things to establish their location accurately, restrict network connectivity to a targeted geographical area, and minimize the effects of misbehaving devices.
|
|||
|---|---|---|---|
KIWI Project - Towards Self-Managed Wireless LANs|
This work will help biodiversity researchers Acquire digital materials in the field, manage these online holdings (Curate), and Transfer the knowledge (or disseminate) to other researchers, museums, and the public. We will create three sets of tools:
|
|||
|---|---|---|---|
BioACT!|
Stanley and the Stanford Racing Team were awarded 2 million dollars for being the first team to complete the 132 mile DARPA Grand Challenge course. Stanley finished in just under 6 hours 54 minutes and averaged over 19 miles per hours on the course.
The Grand Challenge, first held in March 2004, is an off-road robot competition devised by DARPA to promote research in the area of autonomous vehicles. The challenge, in general terms, is to build a robot capable of navigating without human intervention over 130 miles of rough terrain in less than 10 hours.
Calling this task a Grand Challenge is not an exaggeration! The 2004 course started in Barstow, California (1.5 hours outside of Los Angeles) and ended in Primm, Nevada. The most successful competitor navigated only 7.5 miles out of the 142 mile course.
We invite you to visit www.stanfordracing.org to check on our progress!
|
|||
|---|---|---|---|
DARPA Grand Challenge|
Since 1996, research on light fields has followed a number of lines. On the theoretical side, researchers have developed spatial and frequency domain analyses of light field sampling and have proposed several new parameterizations of the light field, including surface light fields and unstructured Lumigraphs. On the practical side, researchers have experimented with literally dozens of ways to capture light fields, ranging from camera arrays to kaleidoscopes, as well as several ways to display them, such as an array of video projectors aimed at a lenticular sheet.
Researchers have also explored the relationship between light fields and other sampled representations of light transport, such as incident light fields and reflectance fields. At Stanford, we have focused on the boundary between light fields, photography, and high-performance imaging, an area we sometimes call computational photography. However, our research also touches on other aspects of light fields, such as interactive animation of light fields and computing shape from light fields.
|
|||
|---|---|---|---|
Light Fields & Computational Photography|
The goal of this research is to enable a multi-limbed robot to climb vertical rock using techniques similar to those developed by human climbers. The robot consists of a small number of articulated limbs. Only the limb end-points can make contact with the environment�a vertical surface with small, arbitrarily distributed features called holds. A path through this environment is a sequence of one-step climbing moves in which the robot brings a limb end-point to a new hold. The robot maintains balance during each move by pushing and/or pulling at other holds, exploiting contact and friction at these holds while adjusting internal degrees of freedom to avoid sliding. The fixed set of robot-hold contacts during a one-step move is called a stance. Our planner combines a multi-step and a one-step planner. The multi-step planner searches a graph representing the adjacency relation between stances to compute a sequence of steps from the initial to the goal stance. The one-step planner (a PRM planner) searches the robot�s configuration space at a given stance for a feasible path between the previous and the next stance in the sequence computed by the multi-step planner. The one-step planner makes use of an efficient test efficient of the quasi-static equilibrium of thee robot. The multi-step planner makes use of lazy search techniques to speed up the exploration of the stance graph. It also makes use of a trained classifier to quickly recognize infeasible transitions between stances. The planner was tested in simulation and on a real four-limbed robot�LEMUR IIB�created by NASA/JPL. A new climbing robot with an increased number of degrees of freedom is currently being developed. In another develop, we adapt and extend the planner to the ATHLETE robot, a six-legged vehicle designed by NASA/JPL to climb steep, irregular, and possibly non-rigid lunar terrain. In such terrain ATHLETE�s wheels are frozen and are used as �feet�. This project is a joint effort with the group of Prof. Steve Rock in the Stanford Aerospace Robotics Lab.
|
|||
|---|---|---|---|
Climbing Robots|
The Common Password Problem. Users tend to use a single password at many different web sites. By now there are several reported cases where attackers breaks into a low security site to retrieve thousands of username/password pairs and directly try them one by one at a high security e-commerce site such as eBay. As expected, this attack is remarkably effective.
A Simple Solution. PwdHash is a browser extension that transparently converts a user's password into a domain-specific password. The user can activate this hashing by choosing passwords that start with a special prefix (@@) or by pressing a special password key (F2). PwdHash automatically replaces the contents of these password fields with a one-way hash of the pair (password, domain-name). As a result, the site only sees a domain-specific hash of the password, as opposed to the password itself. A break-in at a low security site exposes password hashes rather than an actual password. We emphasize that the hash function we use is public and can be computed on any machine which enables users to login to their web accounts from any machine in the world. Hashing is done using a Pseudo Random Function (PRF).
Phishing protection. A major benefit of PwdHash is that it provides a defense against password phishing scams. In a phishing scam, users are directed to a spoof web site where they are asked to enter their username and password. SpoofGuard is a browser extension that alerts the user when a phishing page is encountered. PwdHash complements SpoofGuard in defending users from phishng scams: using PwdHash the phisher only sees a hash of the password specific to the domain hosting the spoof page. This hash is useless at the site that the phisher intended to spoof.
|
|||
|---|---|---|---|
PwdHash - Web Password Hashing|
Volumetric models of facial musculature enable highly realistic computer simulations of visual speech and expressive face motion. This research addresses the performance, control and analysis challenges that arise from the simulation of such high detail models and caters to both computer graphics and medical applications.
Realistic simulation and analysis of facial expressions is an open research problem with a wide range of potential applications such as special effects, communications, visual speech synthesis and maxillofacial medicine. The goals of visual realism and biophysical accuracy suggest a clear advantage for simulation methods that respect the anatomy of the face by modeling the tissue composition of the flesh and employ the underlying musculature as the driving force behind the formation of expressions.
This project involves the development of novel algorithms for increasing the computational performance of high resolution face models by several orders of magnitude, enhancing the quality of all visual elements involved in the simulation and providing convenient, intuitive and efficient ways of controlling the motion of the face, for example using easy to obtain motion capture data.
|
|||
|---|---|---|---|
Simulation & Analysis of Muscle Actuated 3D Face Models|
Trio is a new database system that manages not only data, but also the accuracy and lineage of the data. Approximate (uncertain, probabilistic, ncomplete, fuzzy, and imprecise!) databases have been proposed in the past, and the lineage problem also has been studied. The goals of the Trio project are:
|
|||
|---|---|---|---|
TRIO: A System for Integrated Management of Data, Accuracy, and Lineage|
Increasing use of computers and networks in business, government, recreation, and almost all aspects of daily life has led to a proliferation of online sensitive data, i.e., data that, if used improperly, can harm the data subjects. As a result, concern about the ownership, control, privacy, and accuracy of these data has become a top priority. This project focuses on both the technical challenges of handling sensitive data and the policy and legal issues facing data subjects, data owners, and data users.
|
|||
|---|---|---|---|
PORTIA: Managing sensitive information in a wired world