Lifeboat Foundation

Artificial intelligence (AI) and machine learning techniques have proved to be very promising for completing numerous tasks, including those that involve processing and generating language. Language-related machine learning models have enabled the creation of systems that can interact and converse with humans, including chatbots, smart assistants, and smart speakers.

To tackle dialog-oriented tasks, language models should be able to learn high-quality dialog representations. These are representations that summarize the different ideas expressed by two parties who are conversing about specific topics and how these dialogs are structured.

Researchers at Northwestern University and AWS AI Labs have recently developed a self-supervised learning model that can learn effective dialog representations for different types of dialogs. This model, introduced in a paper pre-published on arXiv, could be used to develop more versatile and better performing dialog systems using a limited amount of training data.

An engineer from the Johns Hopkins Center for Language and Speech Processing has developed a machine learning model that can distinguish functions of speech in transcripts of dialogs outputted by language understanding, or LU, systems in an approach that could eventually help computers “understand” spoken or written text in much the same way that humans do.

Developed by CLSP Assistant Research Scientist Piotr Zelasko, the new model identifies the intent behind words and organizes them into categories such as “Statement,” “Question,” or “Interruption,” in the final transcript: a task called “dialog act recognition.” By providing other models with a more organized and segmented version of text to work with, Zelasko’s model could become a first step in making sense of a conversation, he said.

“This new method means that LU systems no longer have to deal with huge, unstructured chunks of text, which they struggle with when trying to classify things such as the topic, sentiment, or intent of the text. Instead, they can work with a series of expressions, which are saying very specific things, like a question or interruption. My model enables these systems to work where they might have otherwise failed,” said Zelasko, whose study appeared recently in Transactions of the Association for Computational Linguistics.

As robots are gradually introduced into various real-world environments, developers and roboticists will need to ensure that they can safely operate around humans. In recent years, they have introduced various approaches for estimating the positions and predicting the movements of robots in real-time.

Researchers at the Universidade Federal de Pernambuco in Brazil have recently created a new deep learning model to estimate the pose of robotic arms and predict their movements. This model, introduced in a paper pre-published on arXiv, is specifically designed to enhance the safety of robots while they are collaborating or interacting with humans.

“Motivated by the need to anticipate accidents during human-robot interaction (HRI), we explore a framework that improves the safety of people working in close proximity to robots,” Djamel H. Sadok, one of the researchers who carried out the study, told TechXplore. “Pose detection is seen as an important component of the overall solution. To this end, we propose a new architecture for Pose Detection based on Self-Calibrated Convolutions (SCConv) and Extreme Learning Machine (ELM).”

An estimated one-quarter of adults in the U.S. have nonalcoholic fatty liver disease (NAFLD), an excess of fat in liver cells that can cause chronic inflammation and liver damage, increasing the risk of liver cancer. Now, UT Southwestern researchers have developed a simple blood test to predict which NAFLD patients are most likely to develop liver cancer.

“This test lets us noninvasively identify who should be followed most closely with regular ultrasounds to screen for liver cancer,” said Yujin Hoshida, M.D. Ph.D., Associate Professor of Internal Medicine in the Division of Digestive and Liver Diseases at UTSW, a member of the Harold C. Simmons Comprehensive Cancer Center, and senior author of the paper published in Science Translational Medicine.

NAFLD is rapidly emerging as a major cause of chronic liver disease in the United States. With rising rates of obesity and diabetes, its incidence is expected to keep growing. Studies have found that people with NAFLD have up to a seventeenfold increased risk of liver cancer. For NAFLD patients believed to be most at risk of cancer, doctors recommend a demanding screening program involving a liver ultrasound every six months. But pinpointing which patients are in this group is challenging and has typically involved invasive biopsies.

An international team of researchers has demonstrated a technique that allows them to align gold nanorods using magnetic fields, while preserving the underlying optical properties of the gold nanorods.

“Gold nanorods are of interest because they can absorb and scatter specific wavelengths of light, making them attractive for use in applications such as biomedical imaging, sensors, and other technologies,” says Joe Tracy, corresponding author of a paper on the work and a professor of materials science and engineering at North Carolina State University.

It is possible to tune the wavelengths of light absorbed and scattered by engineering the dimensions of the gold nanorods. Magnetically controlling their orientation makes it possible to further control and modulate which wavelengths the nanorods respond to.

The invention of the transistor in 1947 by Shockley, Bardeen and Brattain at Bell Laboratories ushered in the age of microelectronics and revolutionized our lives. First, so-called bipolar transistors were invented, in which negative and positive charge carriers contribute to the current transport; unipolar field effect transistors were only added later. The increasing performance due to the scaling of silicon electronics in the nanometer range has immensely accelerated the processing of data. However, this very rigid technology is less suitable for new types of flexible electronic components, such as rollable TV displays or medical applications.

For such applications, transistors made of organic material, or carbon-based semiconductors, have come into focus in recent years. Organic field effect transistors were introduced as early as 1986, but their performance still lags far behind silicon components.

A research group led by Prof. Karl Leo and Dr. Hans Kleemann at the TU Dresden has now succeeded for the first time in demonstrating an organic, highly efficient bipolar transistor. Crucial to this was the use of highly ordered thin organic layers. This new technology is many times faster than previous organic transistors, and for the first time the components have reached operating frequencies in the gigahertz range (i.e., more than a billion switching operations per second).

Biometric authentication like fingerprint and iris scans are a staple of any spy movie, and trying to circumvent those security measures is often a core plot point. But these days the technology is not limited to spies, as fingerprint verification and facial recognition are now common features on many of our phones.

Now, researchers have developed a new potential odorous option for the biometric security toolkit: your breath. In a report published in Chemical Communications, researchers from Kyushu University’s Institute for Materials Chemistry and Engineering, in collaboration with the University of Tokyo, have developed an olfactory sensor capable of identifying individuals by analyzing the compounds in their breath.

Combined with machine learning, this “artificial nose,” built with a 16-channel sensor array, was able to authenticate up to 20 individuals with an average accuracy of more than 97%.

An international team of researchers has developed a scanning tool to make websites less vulnerable to hacking and cyberattacks.

The black box security assessment prototype, tested by engineers in Australia, Pakistan and the UAE, is more effective than existing web scanners which collectively fail to detect the top 10 weaknesses in web applications.

UniSA mechanical and systems engineer Dr. Yousef Amer is one of the co-authors of a new international paper that describes the development of the tool in the wake of escalating global cyberattacks.

An autonomous vehicle is able to navigate city streets and other less-busy environments by recognizing pedestrians, other vehicles and potential obstacles through artificial intelligence. This is achieved with the help of artificial neural networks, which are trained to “see” the car’s surroundings, mimicking the human visual perception system.

But unlike humans, cars using artificial neural networks have no memory of the past and are in a constant state of seeing the world for the first time—no matter how many times they’ve driven down a particular road before. This is particularly problematic in adverse weather conditions, when the car cannot safely rely on its sensors.

Continue reading “Technology helps self-driving cars learn from their own memories” »