Skip to Content

From the Labs: Information Technology

New publications, experiments and breakthroughs in information technology–and what they mean.

Automated Towel Folder
A robot learns to reliably manipulate flexible objects

Fold, repeat: This PR2 robot from Willow Garage in Menlo Park, CA, has been programmed to use twin high-resolution cameras to examine a towel before folding it. It started by pulling this towel from a disordered pile.

Source: “Cloth Grasp Point Detection Based on Multiple-View Geometric Cues with Application to Robotic Towel Folding”
Jeremy Maitin-Shepard et al.
2010 IEEE International Conference on Robotics and Automation (ICRA2010), May 3-8, 2010, Anchorage, AK

Results: For the first time, a robot can reliably sort, fold, and stack different-sized towels that it encounters in a disordered pile. Previous robots have had only limited success at folding towels, even when the towels were laid out in a particular orientation and the robots were programmed with their dimensions.

Why it matters: Even robots that are adept at manipulating rigid objects, such as cups or tools, have not been able to dependably handle unfamiliar objects made of cloth, which changes shape. The ability to do so could eventually allow robots to help with laundry or other tasks that involve malleable objects.

Methods: After the robot picks up a towel from a pile, it rotates the towel with a two-fingered gripper and uses high-resolution cameras to view it from multiple angles. An algorithm developed by researchers at the University of California, Berkeley, enables the robot to distinguish the towel’s edges from folds in the cloth by estimating the curvature of the cloth (an edge is slightly sharper than a fold). By noting where two edges meet, it can identify a corner. After grasping the corner, it identifies and grasps an adjacent corner; then it untwists the towel and checks its configuration before folding it. To keep it from getting stuck, the researchers designed a self-correcting process. If it detects an error, it attempts to correct it or, if necessary, drops the towel and starts over. The robot successfully folded towels on its own 50 out of 50 times.

Next steps: The robot currently takes about 20 minutes to fold a single towel; the researchers are adapting the algorithm to speed up the process. They also plan to program robots to fold other objects, such as shirts, and to learn how to fold new items after observing a human demonstration.

Uncovering Search Histories
Personalized ­services on the Internet need high levels of security

Source: “Private Information Disclosure from Web Searches (The Case of Google Web History)”
Claude Castelluccia et al.
Proceedings of the 10th Privacy Enhancing Technologies Symposium, July 21-23, 2010, Berlin, Germany

Results: Researchers successfully reconstructed the Web search histories of specific Google users by stealing the users’ credentials and impersonating them. They were able to identify about 65 percent of what the users had been searching for, and they could tell whether a user had searched for a particular term.

Why it matters: Personalized Web services can help make searches and other tasks faster, but the new research suggests that they could also be used to collect information about search histories that people might prefer to keep private. A single search on a public Wi-Fi network would be enough to expose a person’s search history to a potential attacker. Although Google has made changes to prevent search histories from being discovered, the researchers say that other search engines are likely to have similar vulnerabilities. They recommend that Web applications encrypt all searches and credentials.

Methods: Google encrypts sensitive information such as passwords, but it doesn’t encrypt the authentication credentials that it uses to identify particular users of its search service. By intercepting these credentials, the researchers were able to impersonate a given user. Then they performed automated test searches in the user’s name and pieced together the Web search history from the personalized recommendations that Google provided.

Next steps: The researchers plan to analyze other search engines for similar leaks. They also continue to track the progress at fixing the problems they found.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.