Characters portrayed on-screen with autism, from Sheldon Cooper to Sherlock Holmes, have traditionally been characterized by their unique social skills.
While experts often focus on social skills when diagnosing individuals with autism, recent research suggests that other traits may be better indicators of the condition. These traits include repetitive behaviors, special interests, and differences in sensation.
Recent research has shown that special interests may play a more significant role in diagnosing autism.
Special interest may be more important to obtain an autism diagnosis – Credit: Mikset via Getty
“Our research has the potential to have a significant impact,” says Jack Stanley, a doctoral student in biochemistry and machine learning at McGill University in Montreal. BBC Science Focus.
From the perspective of the autism community, this study could lead to a reevaluation of longstanding clinical standards for diagnosing autism.
Diagnosing individuals with autism typically involves clinical observations where healthcare professionals assess potential autism traits. However, this process relies heavily on subjective judgment and intuition.
McGill researchers utilized a large-scale language model (LLM) to quantitatively understand how clinicians diagnose autism based on over 4,000 reports from clinicians assessing patients with the condition.
“Our goal was not to replace clinicians with LLM, but to better understand the key factors in diagnosing autism,” explains Stanley.
The study revealed that repetitive behaviors and special interests were more crucial in predicting autism diagnoses than social skills, which contrasts with current clinical guidelines.
As the study did not differentiate between genders, the findings may vary for individuals of different genders.
Researchers hope that this study will prompt healthcare professionals to reassess the most relevant factors in diagnosing autism.
read more:
About our experts:
Jack Stanley is a doctoral student at the Mira AI Institute in Montreal, specializing in applying machine learning to biological and medical challenges. He completed an Honors BSC in Statistics and Biochemistry at the University of Toronto before joining McGill.
It’s been a tough week for the Grand St. Seven, a group of technology stocks that have played a leading role in the U.S. stock market, buoyed by investor excitement about breakthroughs in artificial intelligence.
Last year, Microsoft, Amazon, Apple, chipmaker Nvidia, Google parent Alphabet, Facebook owner Meta and Elon Musk’s Tesla accounted for half of the S&P 500’s gains. But doubts about returns on AI investments, mixed quarterly earnings, investor attention shifting elsewhere and weak U.S. economic data have hurt the group over the past month.
Things came to a head this week when the shares of the seven companies entered a correction, with their combined share prices now down more than 10% from their peak on July 10.
Here we answer some questions about Seven and the AI boom.
Why did AI stocks fall?
First, there are concerns that the huge investments being made by Microsoft, Google and others in AI will pay off. These have been growing in recent months. Goldman Sachs analysts The memo was published In June, the Wall Street bank released a report titled “Gen AI: Too Much Spending, Too Little Reward?” which asked whether $1 trillion in investment in AI over the next few years “will ever pay off,” while an analysis by Sequoia Capital, an early investor in ChatGPT developer OpenAI, estimated that tech companies would need $600 billion in rewards to recoup their AI investments.
Gino said “The Magnificent Seven” is also hit by these concerns.
“There are clearly concerns about the return on the AI investments that they’re making,” he said, adding that big tech companies have “done a good job explaining” their AI strategies, at least in their most recent financial results.
Another factor at play is investor hope that the Federal Reserve, the U.S. central bank, may cut interest rates as soon as next month. The prospect of lower borrowing costs has boosted investors’ support for companies that could benefit, such as small businesses, banks and real estate companies. This is an example of “sector rotation,” in which investors move money between different parts of the stock market.
Concerns about the Big 7 are affecting the S&P 500, given that a small number of tech stocks make up much of the index’s value.
“Given the growing concentration of this group within U.S. stocks, this will have broader implications,” said Henry Allen, macro strategist at Deutsche Bank AG.Concerns about a weakening U.S. economy also hit global stock markets on Friday.
What happened to tech stocks this week?
As of Friday morning, the seven stocks were down 11.8% from last month’s record highs, but had been dipping in and out of correction territory — a drop of 10% or more from a recent high — in recent weeks amid growing doubts.
Quarterly earnings this week were mixed. Microsoft’s cloud-computing division, which plays a key role in helping companies train and run AI models, reported weaker-than-expected growth. Amazon, the other cloud-computing giant, also disappointed, as growth in its cloud business was offset by increased spending on AI-related infrastructure like data centers and chips.
But shares of Meta, the owner of advertising-dependent Facebook and Instagram, rose on Thursday as the company’s strong revenue growth offset promises of heavy investment in AI. Apple’s sales also beat expectations on Thursday.
“Expectations for the so-called ‘great seven’ group have perhaps become too high,” Dan Coatsworth, an analyst at investment platform AJ Bell, said in a note this week. “These companies’ success puts them out of reach in the eyes of investors, and any shortfall in greatness leaves them open to harsh criticism.”
A general perception that tech stocks may be overvalued is also playing a role: “Valuations have reached 20-year highs and they needed to come down and take a pause to digest some of the gains of the past 18 months,” says Angelo Gino, a technology analyst at CFRA Research.
The Financial Times reported on Friday that hedge fund Elliott Management said in a note to investors that AI is “overvalued” and that Nvidia, which has been a big beneficiary of the AI boom, is in a “bubble.”
Can we expect to see further advances in AI over the next 12 months?
Further breakthroughs are almost certain, which may reassure investors. The biggest players in the field have a clear roadmap, with the next generation of frontier models already underway to train, and new records are being set almost every month. Last week, Alphabet Inc.’s Google DeepMind announced that its system had set a new record at the International Mathematical Olympiad, a high school-level math competition. The announcement has observers wondering whether the company will be able to tackle long-unsolved problems in the near future.
The question for labs is whether these breakthroughs will generate enough revenue to cover the rapidly growing costs of achieving them: The cost of training cutting-edge AI has increased tenfold every year since the AI boom really began, raising questions about how even well-funded companies such as OpenAI, the Microsoft-backed startup behind ChatGPT, will cover those costs in the long run.
Is generative AI already benefiting the companies that use it?
In many companies, the most successful uses of generative AI (the term for AI tools that can create plausible text, voice, and images from simple prompts) have come from the bottom up: people who have effectively used tools like Microsoft’s Copilot or Anthropic’s Claude to figure out how to work more efficiently, or even eliminate time-consuming tasks from their day entirely. But at the enterprise level, clear success stories are few and far between. Whereas Nvidia got rich selling shovels in the gold rush, the best story from an AI user is Klarna, the buy now, pay later company, which announced in February that its OpenAI-powered assistant can: Resolved two-thirds of customer service requests In the first month.
Dario Maisto, a senior analyst at Forrester, said a lack of economically beneficial uses for generative AI is hindering investment.
“The challenge remains to translate this technology into real, tangible economic benefits,” he said.
This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.
Strictly Necessary Cookies
Strictly Necessary Cookie should be enabled at all times so that we can save your preferences for cookie settings.
