Some Japanese researchers feel that AI systems trained on foreign languages cannot grasp the intricacies of Japanese language and culture
Japan is building its own versions of ChatGPT — the artificial intelligence chatbot made by US firm OpenAI that became a worldwide sensation after it was unveiled just under a year ago.
English bias LLMs typically use huge amounts of data from publicly available sources to learn the patterns of natural speech and prose. They are trained to predict the next word on the basis of previous words in a piece of text. The vast majority of the text that ChatGPT’s previous model, GPT-3, was trained on was in English.
Whereas English has just 26 letters, written Japanese consists of two sets of 48 basic characters, plus 2,136 regularly used Chinese characters, or kanji. Most kanji have two or more pronunciations, and a further 50,000 or so rarely used kanji exist. Given that complexity, it is not surprising that ChatGPT can stumble with the language.
To gauge how sensitive LLMs are to Japanese culture, a group of researchers launched Rakuda, a ranking of how well LLMs can answer open-ended questions on Japanese topics. Rakuda co-founder Sam Passaglia and his colleagues asked ChatGPT to compare the fluidity and cultural appropriateness of answers to standard prompts. Their use of the tool to rank the results was based on a preprint published in June that showed that GPT-4 agrees with human reviewers 87% of the time1.
However, the Fugaku LLM might be succeded by an even larger one. Japan’s Ministry of Education, Culture, Sports, Science and Technology is funding the creation of a Japanese AI program tuned to scientific needs that will generate scientific hypotheses by learning from published research, speeding up identification of targets for enquiry. The model could start off at 100 billion parameters, which would be just over half the size of GPT-3, and would be expanded over time.
Masafumi Oyamada, senior principal researcher at NEC Data Science Laboratories, says that it can be used “in a wide range of industries, such as finance, transportation and logistics, distribution and manufacturing”. He adds that researchers could put it to work writing code, helping to write and edit papers and surveying existing published papers, among other tasks.
France Dernières Nouvelles, France Actualités
Similar News:Vous pouvez également lire des articles d'actualité similaires à celui-ci que nous avons collectés auprès d'autres sources d'information.
See building-by-building details for 2023 Ohio school report cardsFind out how your local school did on the 2023 Ohio school report cards released Thursday. These report cards are for individual buildings.
Lire la suite »
| Here’s why your teen overshares online, and why that could be goodFor parents, sharing personal details about intimate facets of one’s life — from sexual orientation to a neurodivergent diagnosis — can seem risky. For teens, it’s a different story.
Lire la suite »
I’m a doctor — this is why you fart so much and why some stink moreAn influencer is having a gas.
Lire la suite »
Why Mitt Romney decided to hang it up — and why he thinks Biden should tooThe retiring Utahn and leading GOP anti-Trumpist simply 'can't vote for Donald Trump,' he told POLITICO. But that doesn't mean he supports the current president, either.
Lire la suite »
Now Open: Red Lantern Japanese Cuisine in Kamm's CornersThe owners also operate the nearby Dragon Towers Chinese restaurant.
Lire la suite »