Google has announced PaLM 2: its latest AI language model and competitor to rival systems like OpenAI\u2019s GPT-4. <\/p>\n
\u201cPaLM 2 models are stronger in logic and reasoning, thanks to broad training in logic and reasoning,\u201d said Google CEO Sundar Pichai onstage at the company\u2019s I\/O conference. \u201cIt\u2019s also trained on multilingual text spanning over 100 languages.\u201d <\/p>\n
PaLM 2 is much better at a range of text-based tasks, Google senior research director Slav Petrov told journalists in a roundtable prior to the model\u2019s announcement at Google\u2019s I\/O conference, including reasoning, coding, and translation. \u201cIt is significantly improved compared to PaLM 1 [which was announced in April 2022],\u201d said Petrov. <\/p>\n
As an example of its multilingual capabilities, Petrov showed how PaLM 2 is able to understand idioms in different languages, giving the example of the German phrase \u201cIch verstehe nur Bahnhof,\u201d which literally translates to \u201cI only understand train station\u201d but is better understood as \u201cI don\u2019t understand what you\u2019re saying\u201d or, as an English idiom, \u201cit\u2019s all Greek to me.\u201d <\/p>\n
In a research paper describing PaLM 2\u2019s capabilities, Google\u2019s engineers claimed the system\u2019s language proficiency is \u201csufficient to teach that language\u201d and noted this is in part due to a greater prevalence of non-English texts in its training data. <\/p> Like other large language models, which take huge amounts of time and resources to create, PaLM 2 is less a single product than a family of products \u2014\u00a0with different versions that will be deployed in consumer and enterprise settings. The system is available in four sizes, named Gecko, Otter, Bison, and Unicorn, from smallest to largest, and has been fine-tuned on domain-specific data to perform certain tasks for enterprise customers. <\/p>\n Think of these adaptations like taking a basic truck chassis and adding a new engine or front bumper to accomplish certain tasks or work better in specific terrain. There\u2019s a version of PaLM trained on health data (Med-PaLM 2), which Google says can answer questions similar to those found on the US Medical Licensing Examination to an \u201cexpert\u201d level and another version trained on cybersecurity data (Sec-PaLM 2) that can \u201cexplain the behavior of potential malicious scripts and help detect threats in code,\u201d said Petrov. Both of these models will be available via Google Cloud, initially to select customers.<\/p>\n Within Google\u2019s own domain, PaLM 2 is already being used to power 25 features and products, including Bard, the company\u2019s experimental chatbot. Updates available through Bard include improved coding capabilities and greater language support. It\u2019s also being used to power features in Google Workspace apps like Docs, Slides, and Sheets. <\/p>\n Notably, Google says the lightest version of PaLM 2, Gecko, is small enough to run on mobile phones, processing 20 tokens per second \u2014\u00a0roughly equivalent to around 16 or 17 words. Google did not say what hardware was used to test this model, only that it was running \u201con the latest phones.\u201d Nevertheless, the miniaturization of such language models is significant. Such systems are expensive to run in the cloud, and being able to use them locally would have other benefits, like improved privacy. The problem is that smaller versions of language models are inevitably less capable than their larger brethren.<\/p> With PaLM 2, Google will be hoping to close the \u201cAI gap\u201d between the company and competitors like Microsoft, which has been aggressively pushing AI language tools into its suite of Office software. Microsoft now offers AI features that help summarize documents, write emails, generate slides for presentations, and much more. Google will need to keep parity with the company or risk being perceived as slow to implement its AI research. <\/p>\n Although PaLM 2 is certainly a step forward for Google\u2019s work on AI language models, it suffers from problems and challenges common to the technology more broadly.<\/p>\n For example, some experts are beginning to question the legality of training data used to create language models. This data is usually scraped from the internet and often includes copyright-protected text and pirated ebooks. Tech companies creating these models have generally responded by refusing to answer questions about where they source their training data from. Google has continued this tradition in its description of PaLM 2, noting only that the system\u2019s training corpus is comprised of \u201ca diverse set of sources: web documents, books, code, mathematics, and conversational data,\u201d without offering further detail. <\/p>\n There are also problems inherent to the output of language models like \u201challucinations,\u201d or the tendency of these systems to simply make up information. Speaking to The Verge<\/em>, Google VP of research Zoubin Ghahramani says that, in this regard, PaLM 2 was an improvement on earlier models \u201cin the sense that we\u2019re putting a huge amount of effort into continually improving metrics of groundedness and attribution\u201d but noted that the field as a whole \u201cstill has a ways to go\u201d in combating false information generated by AI.<\/p>\n","protected":false},"excerpt":{"rendered":" Google has announced PaLM 2: its latest AI language model and competitor to rival systems like OpenAI\u2019s GPT-4. \u201cPaLM 2 models are stronger in logic and reasoning, thanks to broad training in logic and reasoning,\u201d said Google CEO Sundar Pichai onstage at the company\u2019s I\/O conference. \u201cIt\u2019s also trained on multilingual text spanning over 100 […]<\/p>\n","protected":false},"author":1,"featured_media":3405,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[93,95,96,81,84,87,90],"class_list":{"0":"post-3403","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-google-fiber","8":"tag-android-phone-guide","9":"tag-android-phone-news","10":"tag-android-phone-reviews","11":"tag-google","12":"tag-google-guide","13":"tag-google-news","14":"tag-google-reviewsandroid-phone"},"_links":{"self":[{"href":"https:\/\/eufad.com\/index.php?rest_route=\/wp\/v2\/posts\/3403","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/eufad.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/eufad.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/eufad.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/eufad.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=3403"}],"version-history":[{"count":1,"href":"https:\/\/eufad.com\/index.php?rest_route=\/wp\/v2\/posts\/3403\/revisions"}],"predecessor-version":[{"id":5158,"href":"https:\/\/eufad.com\/index.php?rest_route=\/wp\/v2\/posts\/3403\/revisions\/5158"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/eufad.com\/index.php?rest_route=\/wp\/v2\/media\/3405"}],"wp:attachment":[{"href":"https:\/\/eufad.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=3403"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/eufad.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=3403"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/eufad.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=3403"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}