In May, Google New experimental CEOs revealed Artificial intelligence They practiced the texts and images they said would make the Internet Searches much easier. Wednesday, google browser Provide a glimpse into how technology will change the way people search the web.
Starting next year, a unified multitasking model, or MUM, will enable Google users to combine text and image searches using Lens, a smartphone app that’s also integrated into Google Search and other products. So you could, for example, take a picture of a T-shirt with Lens, and then search for “socks with this style.” Searching for “how to fix” on the bike part image will bring up how-to videos or blog posts.
Google will integrate MUM into search results to suggest additional ways for users to explore. If you ask Google how to draw, for example, MUM can detail step-by-step instructions, style tutorials, or how to use homemade materials. Google is also planning in the coming weeks to bring MUM to Youtube Videos in search, where artificial intelligence displays search suggestions under videos based on video texts.
MUM is trained to make inferences about text and images. The integration of MUM into Google search results also represents an ongoing march toward the use of language models that rely on massive amounts of text extracted from the web and a kind of neural network The architecture is called transformers. One of the first such efforts came in 2019, when Google injected a language model called BERT into search results to change web rankings and summarize the text below the results.
Google Vice President Pandu Nayak said BERT represents the biggest change in search results in the better part of a decade, but MUM takes language understanding of AI applied to Google search results to the next level.
For example, MUM uses data from 75 languages rather than English alone, and is trained on images and text rather than text alone. It is 1,000 times larger than BERT when measured in the number of parameters or connections between artificial neurons in a deep learning system.
While Nayak describes MUM as a milestone in language understanding, he also acknowledges that large language paradigms come with known challenges and risks.
BERT and other transformer-based models are proven to suck I found bias in the data used to train them. In some cases, researchers have found that the larger the language model, the worse the amplification of bias and toxic text. People who work to detect and change the racist, sexist, and other problem outputs of large language models say that proofreading the text used to train these models is critical to minimizing harm and that the way the data is filtered can have a negative impact. In April, the Allen Institute for Artificial Intelligence reported that blocklists used in a common dataset that Google used to train its T5 language model could lead to the exclusion of entire blocks, Like people who identify as gay, making it difficult for linguistic models to understand text by or related to these groups.
Last year, several AI researchers at Google, including the former Ethical AI team Timnit Gibru Margaret Mitchell, said they faced opposition from CEOs to their work, which shows that big language models can hurt people. Among Googlers, Gebru’s ouster after a dispute over a paper criticizing the environmental and social costs of large language models has led to allegations of racism, calls for unions, and the need for Stronger whistleblower protection For AI ethics researchers.
In June, five US senators cited multiple incidents of algorithmic bias in Alphabet and Gebru’s ouster among reasons to question whether Google products like Search or the Google workplace are safe for blacks. in a Message “We are concerned that algorithms will rely on data that reinforces negative stereotypes and either exclude people from seeing ads for housing, employment, credit and education or only show predatory opportunities,” the senators wrote to executives.