Tiny magnetic robot armies could treat bleeds in the brain and'open new frontiers in medicine', experts have found. Researchers have created nanoscale robots – each about a twentieth of the size of a red blood cell – that can be remotely guided as a swarm. It is hoped they could enable precise, low-risk treatment of brain aneurisms, which cause around half a million deaths a year globally. The condition – a blood-filled bulge on a brain artery that can rupture and cause fatal bleeds – can also lead to stroke and disability. The team, co-led by the University of Edinburgh's School of Engineering, carried out lab tests using models of aneurisms and rabbits.

Over the last year, significant advancements have been made in the realms of large language models (LLMs) and multi-modal large language models (MLLMs), particularly in their application to autonomous driving. These models have showcased remarkable abilities in processing and interacting with complex information. In autonomous driving, LLMs and MLLMs are extensively used, requiring access to sensitive vehicle data such as precise locations, images, and road conditions. These data are transmitted to an LLM-based inference cloud for advanced analysis. However, concerns arise regarding data security, as the protection against data and privacy breaches primarily depends on the LLM's inherent security measures, without additional scrutiny or evaluation of the LLM's inference outputs. Despite its importance, the security aspect of LLMs in autonomous driving remains underexplored. Addressing this gap, our research introduces a novel security framework for autonomous vehicles, utilizing a multi-agent LLM approach. This framework is designed to safeguard sensitive information associated with autonomous vehicles from potential leaks, while also ensuring that LLM outputs adhere to driving regulations and align with human values. It includes mechanisms to filter out irrelevant queries and verify the safety and reliability of LLM outputs. Utilizing this framework, we evaluated the security, privacy, and cost aspects of eleven large language model-driven autonomous driving cues. Additionally, we performed QA tests on these driving prompts, which successfully demonstrated the framework's efficacy.

The idea of finding lost items in this chaotic world has been a constant problem over the years. How frustrating is it for a busy commuter to sift items to find one small object they misplaced? Researchers at MIT unveil a robot that can resolve this issue and prove helpful even in manufacturing and warehouse environments. With RFusion, a robotic arm with a camera and radio frequency(RF) antenna attached to its gripper, one can easily fetch their lost items. The prototype developed by researchers relies entirely on RFID tags.

By dividing the world into 3m x 3m squares, each with a unique 3-word address, what3words enables the most precise reference to any location around the globe. In contrast to GPS coordinates, 3-word addresses are easy to remember and, more importantly, they are designed for explicit and error-free voice input in more than 20 languages. Currently used by various businesses, governments, and individuals, this technology is now integrating with the automotive navigation system included in the all-new Mercedes-Benz User Experience (MBUX), featured in the new A-Class. At our Nuance Auto Forums in Detroit and Europe, attendees had the opportunity to experience the 3-word address system and to hear the what3word story and vision presented by Gigi Etienne, what3words Partnerships Manager, and Ashley Cashion, Head of Automotive and Mobility at what3words, personally. For those who could not attend, I had the good fortune to talk to these innovators about what3words and their vision on voice input in vehicles. What was the motivation for developing the 3-words address system?

All too often, we search for an app and end up finding what looks to be the best fit for our needs. But that is until one sees the long list of permissions the application thinks it needs to function. Some developers tend to call for permissions for functionality that their app clearly does not need, like an expense tracker needing the RECORD_AUDIO permission, indicating a high possibility of a nefarious motive. Google does realize that many such applications plague the Google Play Store. While the technologically adept users may keep a close eye on the permissions they grant to any app, the normal user usually just presses on "Accept" till they reach their end result.

Last March I wrote about an early experiment using Google's Cloud Vision API to perform deep learning-powered geocoding of 20 million global news images. In that experiment I compiled 20 million photographs that had appeared in online news articles worldwide as monitored by the open data GDELT Project over a period of two months and ran them through Google's Vision API service, which applies state-of-the-art deep learning algorithms to visually analyze an image much as a human would. The API returns a wealth of data about each image, including a list of objects and activities it depicts, recognizable logos, OCR text recognition in almost 80 languages, levels of violence, an estimate of how "happy" or "sad" people in the photograph appear to be and even the precise location on earth the image appears to depict. It is that last category that is so fascinating when it comes to trying to understand the visual geography of the world's news media. One year later the GDELT Project has now processed more than a quarter billion news photographs from news outlets in almost every corner of the world through Google's API – what can we learn through this deep learning powered "visual geocoding" of the world's news imagery?

A new study from MIT reveals that a brain region dedicated to reading has connections for that skill even before children learn to read. By scanning the brains of children before and after they learned to read, the researchers found that they could predict the precise location where each child's visual word form area (VWFA) would develop, based on the connections of that region to other parts of the brain. Neuroscientists have long wondered why the brain has a region exclusively dedicated to reading -- a skill that is unique to humans and only developed about 5,400 years ago, which is not enough time for evolution to have reshaped the brain for that specific task. The new study suggests that the VWFA, located in an area that receives visual input, has pre-existing connections to brain regions associated with language processing, making it ideally suited to become devoted to reading. "Long-range connections that allow this region to talk to other areas of the brain seem to drive function," says Zeynep Saygin, a postdoc at MIT's McGovern Institute for Brain Research.

Maybe the reason why Google isn't giving a formal name to its personal assistant software is that it's more than just one thing. Introduced at I/O this week is a new Android Awareness API that bundles all the sensor data from your smartphone or other Android device and presents it to apps, which can then act on that input to automatically assist you. "You can use this information to build more assistive and aware applications," says Google's Bhavik Singh, product manager of the Awareness API. He offers a number of scenarios where smart assistive apps could help: projecting the day's weather forecast on the nearest Chromecast TV, beaming out traffic alerts to your Google Home speaker to avoid being late for a meeting, or tagging photos with weather and activity data as well as location. In order to be so savvy, however, apps will need access to seven different parameters: the time and place (both type of place and precise location), your physical activity, any nearby wireless beacons, whether or not you have headphones connected, and the weather.