Posts

Showing posts with the label language

Big language models can't plan, even if they write fancy essays

Image
This article is part of our coverage of the latest AI research. Large language models such as GPT-3 have grown to the point where it is difficult to measure the limits of their capabilities. When you have a very large neural network that can produce articles, write software code, and engage in conversations about feelings and life, you should expect it to be able to reason about tasks and plans like humans do, right? Wrong. A study by researchers at Arizona State University, Tempe, showed that when it comes to planning and methodical thinking, LLMs perform very poorly, and suffer from many of the same failures observed in today’s deep learning systems. Regards, humanoids Subscribe to our newsletter now for weekly recaps of our favorite AI stories in your inbox. Interestingly, this study found that, although very large LLMs such as GPT-3 and PaLM pass many tests intended to evaluate reasoning abilities and artificial intelligence systems, they do so because these benchmarks are t...

The study analyzes the spontaneous social interactions of children aged 2 and 4 years when interacting with peers

What do building a pyramid, going to the moon, pedaling a two-person canoe, or dancing the waltz have in common? All of these actions are the result of a common goal between many partners and lead to a sense of shared obligation, known as “mutual commitment”. This ability to cooperate is universal in humans and certain animal species, such as the great apes. However, humans seem to have a unique predisposition and strong desire for social interaction that may be one component of the emergence of language, according to the study authors. How do our social interactions differ from other species? And why? To answer this question, an international team analyzed the interactions of 31 children between the ages of 2 and 4 in four preschools in the United States (10 hours per child). There are only a few quantitative analyzes of the spontaneous social interactions of children aged 2 and 4 years when interacting with peers, even though this age is a critical age for the develop...

'Universal language network' identified in brain

Japanese, Italian, Ukrainian, Swahili, Tagalog, and dozens of other spoken languages ​​cause the same “universal language network” to fire in the brains of native speakers. This language processing center has been studied extensively in English speakers, but now neuroscientists have confirmed that the same network is activated in speakers of 45 different languages ​​representing 12 different language families. “This study is very basic, extending some of the findings from English to multiple languages,” senior author Evelina Fedorenko, a professor of neuroscience at MIT and a member of MIT’s McGovern Institute for Brain Research, said in a statement. statement (opens in a new tab) . “The hope is that now that we see that basic traits seem to be common across languages, we can ask about potential differences between languages ​​and language families in how they are implemented in languages. brain and we can study phenomena that don’t real...

Similar Activity in the Brain's Language Network, No Matter What Language You Speak - Neuroscience News

Image
Summary: In a study of speakers of 45 languages, researchers found similar patterns of brain activity and language selectivity. Source: MIT For decades, neuroscientists have created well-defined maps of the brain’s “language network,” or regions of the brain specialized for processing language. Found primarily in the left hemisphere, this tissue includes areas within Broca’s area, as well as in other parts of the frontal and temporal lobes. However, most of these mapping studies were conducted on English speakers while they were listening to or reading English texts. MIT neuroscientists have now conducted brain imaging studies of speakers of 45 different languages. The results show that the language network of speakers appears to be essentially the same as that of native English speakers. This finding, though not surprising, establishes that the location and key properties of language networks appear to be universal. This work also lays the groundwork for the...

Meet Wylah The Koorie Warrior, the new hero who connects children to Indigenous culture

Image
A bestselling children’s book stars a new kind of hero — she’s a girl, she’s an Indigenous Australian, and she’s a fighter. Guardians: Wylah the Koorie Warrior is an illustrated chapter book; a fantasy adventure set 40,000 years ago in the land of Peek Whurrong in southwest Victoria. Australian children instantly became obsessed with Wylah, making the book one of the best-selling children’s novels of the year so far, and sending it to the top of the charts at booksellers Booktopia and Readings. And the good news for kids who’ve read this book is that the author has mapped out a whole world of new characters and new adventures to ensure the Wylah series can run for years to come. Jordan Gould and Richard Pritchard created the Wylah book series. ( Provided: Allen & Unwin ) Authors inspired by their single mothers Warrnambool-based co-authors Richard Pritchard and Jordan Gould said the book began with a vision for Wylah—a strong First Nations girl...

AI model detects people's attitudes towards vaccines from their social media posts

People’s attitudes towards vaccines can now be detected from their social media posts with a smart AI model, developed by researchers at the University of Warwick. An AI-based model can analyze social media posts and determine the author’s attitude towards vaccines, by being ‘trained’ to recognize that attitude from a small number of sample tweets. As a simple example, if a post contains mention of distrust of health care institutions, fear of needles, or something related to a known conspiracy theory, the model can recognize that the person who wrote it may have negative feelings about vaccinations. The research, funded by UK Research and Innovation (UKRI), will be presented today (12 July) at the North American Association of Computational Linguistics Annual Conference 2022. It is led by Professor Yulan He from the University’s Department of Computer Science, supported by a 5-year Turing AI Fellowship funded by the EPSRC. Professor He and his colleagues a...

Can AI-chatbots promote healthy lifestyle changes?

Image
Artificial intelligence (AI) chatbots are capable of imitating human interactions with the help of spoken, written, or verbal communication with users. AI chatbots can provide critical health-related information and services, ultimately leading to promising technology-facilitated interventions. Study: Artificial Intelligence (AI)-based Chatbots in Promoting Health Behavior Change: A Systematic Overview. Image Credit: TippaPatt / Shutterstock.com AI chatbot in healthcare Today’s telehealth and digital therapeutic interventions are associated with several challenges including discontinuity, low adherence, and inflexibility. AI chatbots are able to overcome these challenges and provide personalized on-demand support, higher interactivity and higher sustainability. The AI ​​chatbot utilizes input data from various sources, followed by data analysis that is completed through natural language processing (NLP) and machine learning (ML). The data output then helps users achieve the...

New computational model can detect cognitive impairment from audio recordings of neuropsychological tests

It takes a lot of time and money to diagnose Alzheimer’s disease. After running a lengthy face-to-face neuropsychological exam, the doctor must copy, review, and analyze each response in detail. But researchers at Boston University have developed a new tool that could automate the process and eventually allow it to move online. Their machine learning-powered computational model can detect cognitive impairment from audio recordings of neuropsychological-tests; no in-person appointment required. Their findings were published in Alzheimer’s & Dementia: Journal of the Alzheimer’s Association . “This approach brings us one step closer to early intervention,” said Ioannis Paschalidis, co-author of the paper and BU College of Engineering Distinguished Professor of Engineering. He said faster early detection of Alzheimer’s could prompt larger clinical trials that focus on individuals in the early stages of the disease and potentially enable clinica...