Toggle light / dark theme

Blog post: https://openai.com/blog/new-and-improved-embedding-model/
Abstract: We are excited to announce a new embedding model which is significantly more capable, cost effective, and simpler to use. The new model, text-embedding-ada-002, replaces five separate models for text search, text similarity, and code search, and outperforms our previous most capable model, Davinci, at most tasks, while being priced 99.8% lower.

#nlproc #nlp #artificialintelligence #ml #deeplearning

According to Baidu CEO Robin Li Baidu will release its AI generative language model Ernie bot in March. However, specific dates regarding when it will be available remains unknown, but it could come “very soon”. It is very likely that only Baidu’s users (those owning a Baidu account) will have access to Ernie.

Although Ernie has been broadly adopted by Baidu services, it has yet to open for the public.

Li previously announced that the company will launch Ernie Bot in March. The large language model was first introduced in 2019 and has been updated to the third generation. As OpenAI’s ChatGPT draws broad attention, Baidu is also jumping on the bandwagon and preparing to launch Ernie.

Has the quest for room temperature superconductivity finally succeeded? Researchers at the University of Rochester (U of R), who previously were forced to retract a controversial claim of room temperature superconductivity at high pressures, are back with an even more spectacular claim. This week in they report a new material that superconducts at room temperature—and not much more than ambient pressures.

“If this is correct, it’s completely revolutionary,” says James Hamlin, a physicist at the University of Florida who was not involved with the work. A room temperature superconductor would usher in a century-long dream. Existing superconductors require expensive and bulky chilling systems to conduct electricity frictionlessly, but room temperature materials could lead to hyperefficient electricity grids and computer chips, as well as the ultrapowerful magnets needed for levitating trains and fusion power.

But given the U of R group’s recent retraction, many physicists won’t be easily convinced. “I think they will have to do some real work and be really open for people to believe it,” Hamlin says. Jorge Hirsch, a physicist at the University of California, San Diego, and a vociferous critic of the earlier work, is even more blunt. “I doubt [the new result], because I don’t trust these authors.”

Ziqiang zhang*, long zhou*, chengyi wang, sanyuan chen, yu wu, shujie liu, zhuo chen, yanqing liu, huaming wang, jinyu li, lei he, sheng zhao, furu wei.

Microsoft

Abstract. We propose a cross-lingual neural codec language model, VALL-E X, for cross-lingual speech synthesis. Specifically, we extend VALL-E and train a multi-lingual conditional codec language model to predict the acoustic token sequences of the target language speech by using both the source language speech and the target language text as prompts. VALL-E X inherits strong in-context learning capabilities and can be applied for zero-shot cross-lingual text-to-speech synthesis and zero-shot speech-to-speech translation tasks. Experimental results show that it can generate high-quality speech in the target language via just one speech utterance in the source language as a prompt while preserving the unseen speaker’s voice, emotion, and acoustic environment. Moreover, VALL-E X effectively alleviates the foreign accent problems, which can be controlled by a language ID.