By
Paula ParisiOctober 28, 2024
OpenAI is taking a new approach to generating media that it says is 50 times faster than the models commonly used today. Called sCM, the approach is a “consistency model,” a variation on the diffusion method used by many leading systems. OpenAI claims its new model is ideal for training for large scale datasets and generating video, audio and images that are of “comparable sample quality to leading diffusion models.” Such models often require hundreds of steps, creating challenges when it comes to real-time applications. OpenAI aims to change this with a faster system that requires less power. Continue reading OpenAI: sCM Generates Media 50x Faster Than Other Models
By
Paula ParisiSeptember 9, 2024
Gracenote, the Nielsen content solutions division, has launched Gracenote Watch Prompts, an AI-powered dataset that equips global video platforms and services with programming facts intended to help influence consumer viewing behavior. Designed to be paired with user preference and consumption data, the new Watch Prompts aim at delivering personalized film and TV promotion, resulting in increased tune-in. According to Nielsen, 74 percent of U.S. consumers last year either didn’t know or only had a vague idea as to what they wanted to watch when starting a streaming session, “meaning a large majority are making on-the-fly viewing decisions.” Continue reading Gracenote Watch Prompts Aim to Help Streaming TV Viewers
By
Paula ParisiOctober 9, 2023
“Robots are great specialists, but poor generalists,” according to Google DeepMind, which says models are typically trained for individual tasks, and changing a single variable can mean starting again from scratch. Now the London-based Alphabet subsidiary thinks it’s come up with a way to combine knowledge across robotics for a general purpose machine helper. In conjunction with 33 academic labs, Google DeepMind has pooled data from 22 different robot types to create the Open X-Embodiment dataset. Simultaneously, the group releases the RT-1-X robotics transformer (RT) model derived from RT-1. Continue reading DeepMind and Academics Advance General Purpose Robots
By
Paula ParisiAugust 29, 2023
Hugging Face has collected $235 million in an investment round that includes contributions from Amazon, IBM, Google, Nvidia, Salesforce, AMD, Intel and Qualcomm. The New York-based startup creates and distributes open-source tools for artificial intelligence development, carving an AI-centric niche similar to the more general programming approach taken by the Microsoft-owned GitHub. The incoming cash infusion — earmarked for talent recruitment — gives Hugging Face a lofty $4.5 billion valuation that experts say indicates momentum for open source in what has to date been a highly competitive AI sector. Continue reading Big Tech Firms Propel Hugging Face to $4.5 Billion Valuation
By
Paula ParisiApril 7, 2023
Meta Platforms has published a new AI technology, the Segment Anything Model (SAM) that the company claims can identify objects it hasn’t seen before. Acting on a text prompt, SAM will highlight items in a photo or video, picking out all the cats, for instance, or flowers. It can also execute other functions, such as generating a 3D construct using a single 2D image or extrapolating from things viewed in a mixed reality headset. Segment Anything can work in concert with other models, potentially minimizing the need for voluminous data sets for training. Continue reading New Meta AI Can Detect Objects It Has Not Been Trained On
By
Debra KaufmanMay 13, 2021
During its Think conference this week, IBM debuted Project CodeNet, an open-source dataset for benchmarking around AI for code. Project CodeNet consists of 14 million code examples, which makes it about 10 times larger than the most similar dataset, which has 52,000 examples. Project CodeNet also offers 500 million lines of code and 55 programming languages including C++, Java, Python, Go, COBOL, Pascal and Fortran, making it a Rosetta Stone for AI systems to automatically translate code into other programming languages. Continue reading IBM CodeNet Enables AI Translation of Computer Languages
By
Debra KaufmanJuly 26, 2019
SiriusXM’s Pandora unveiled Voice Mode, its voice assistant, to Pandora app users on all platforms. This broad rollout took place after a January debut of Voice Mode on about one million iOS and Android devices. The shortcut to Voice Mode can be found in the upper right of the Pandora app screen. Pandora chief product officer Chris Phillips noted that Voice Mode introduces “an even more natural and conversational way for listeners to discover new music … like getting recommendations from a friend who really knows you.” Continue reading Pandora Rolls Out Voice Mode Assistant For All Platforms
By
Debra KaufmanNovember 2, 2018
Facebook’s video clips get over 8 billion views a day on average, but people with bad Internet connections or disabilities often don’t have access to them. That led Facebook to create VideoStory, which the company described in a research paper as “A Dataset for Telling the Stories of Social Media Videos.” The paper, to be delivered at the Conference on Empirical Methods in Natural Language Processing, noted that, “automatically telling the stories using multi-sentence descriptions of videos would allow bridging this gap.” Continue reading Facebook’s VideoStory Relies on AI to Automate Storytelling
By
Debra KaufmanFebruary 26, 2018
T-Commerce, which allows consumers to buy items they see on TV shows directly from their set, is not new. Predicted since the 1980s, it’s appeared as Enhanced TV, T-Commerce and Shopification. The idea that a viewer could simply click on a sweater worn by her favorite sitcom character and purchase it is heady, but the difficulty of making items clickable frame-by-frame was (and is) a massive stumbling block. At the HPA Tech Retreat, RealNetworks described its T-Commerce solution, powered by computer vision. Continue reading HPA 2018: Real Networks Explores the Future of T-Commerce
By
Rob ScottOctober 16, 2013
A new website is building a dataset to help determine whether consumers are increasingly turning to piracy when movies are not made available legally via streaming, digital rental or digital purchase. The site — PiracyData.org — lists the top 10 most pirated movies on BitTorrent as reported by TorrentFreak each week, and researches the Can I Stream It? service to determine whether each title is available legally. The authors suggest that shorter windows would help counter piracy. Continue reading Do Movie Fans Turn to Piracy Due to a Lack of Legal Options?
By
Rob ScottOctober 10, 2013
Verizon is planning to roll out a public beta of its Verizon Cloud “Infrastructure as a Service” platform and cloud-based object storage in the fourth quarter of this year. Verizon Cloud, which includes two primary components — Verizon Cloud Compute (the IaaS platform) and Verizon Cloud Storage (an object-based service) — will target enterprises, mid-size companies and development shops. The launch will initially involve a few hundred new users per month. Continue reading Verizon to Launch Public Beta of Cloud Platform This Year