Researchers at Amazon have trained what they are calling the largest text-to-speech model ever created, which they claim is exhibiting “emergent” qualities — the ability to inherently improve itself at speaking complex sentences naturally. Called BASE TTS, for Big Adaptive Streamable TTS with Emergent abilities, the new model could pave the way for more human-like interactions with AI, reports suggest. Trained on 100,000 hours of public domain speech data, BASE TTS offers “state-of-the-art naturalness” in English as well as some German, Dutch and Spanish. Text-to-speech models are used in developing voice assistants for smart devices and apps and accessibility. Continue reading Amazon Claims ’Emergent Abilities’ for Text-to-Speech Model

Instagram co-founders Kevin Systrom and Mike Krieger are back with a personalized news feed called Artifact that that uses artificial intelligence to pattern users’ interests and the friends that most likely want to discuss them with you. The new app — whose name combines articles, facts and artificial intelligence — opened a public waiting list this week and is available on iOS and Android. The Verge calls it “TikTok for text,” adding that “you might also call it Google Reader reborn as a mobile app or maybe even a surprise attack on Twitter.” Continue reading Instagram Creators Launch Artifact, Called a ‘TikTok for Text’

Google has released a research paper on a new text-to-image generator called Imagen, which combines the power of large transformer language models for text with the capabilities of diffusion models in high-fidelity image generation. “Our key discovery is that generic large language models (e.g. T5), pretrained on text-only corpora, are surprisingly effective at encoding text for image synthesis,” the company said. Simultaneously, Google is introducing DrawBench, a benchmark for text-to-image models it says was used to compare Imagen with other recent technologies including VQGAN+CLIP, latent diffusion models, and OpenAI’s DALL-E 2. Continue reading Google’s Imagen AI Model Makes Advances in Text-to-Image

Facebook, the University of Lorraine and the University College London published a paper that detailed how they researched a way to use artificial intelligence to create complex game worlds. Building video game environments has always been time-consuming, but this group’s approach — using content from the fantasy text-based multiplayer adventure “LIGHT” — demonstrated how machine learning algorithms can learn to “assemble different elements, arranging locations and populating them with characters and objects.” Continue reading Facebook and Academics Use AI to Generate Game Worlds

Google has unveiled GPipe, an open-sourced library that makes training deep neural networks more efficient under the TensorFlow framework Lingvo for sequence modeling. According to Google AI software engineer Yanping Huang, “in GPipe … we demonstrate the use of pipeline parallelism to scale up DNN training,” noting that larger DNN models “lead to better task performance.” Huang and his colleagues published a paper on “Efficient Training of Giant Neural Networks Using Pipeline Parallelism.” Continue reading Google GPipe Library Speeds Deep Neural Network Training

New Windows personal computers will be introduced at CES next week. One called “PC Plus” will run both Windows 8.1 and Android mobile apps according to Creative Strategies’ Tim Bajarin, who says the initiative would be backed by Intel. PC Plus will be revealed by multiple original equipment manufacturers in a move many say should “scare” Microsoft. The PC market has seen a decline in the past two years, inspiring risks like this by OEMs. Continue reading PC Plus Running Windows and Android To Be Revealed at CES