CybAI news - - Page 2357

27 May

6 NATO countries are planning to build a ‘drone wall’ to defend themselves against Russia, reports say

stp2y0 CommentsRAG models

Six NATO countries are joining forces to create a "drone wall" to defend themselves against Russia, AFP news agency reported.Over the weekend, ministers from Finland, Norway, Poland, Estonia, Latvia, and Lithuania convened to discuss a coordinated system to help defend against Russian provocations and smuggling attempts, per the Financial Times."This is a completely new thing -- a drone wall stretching from Norway to Poland, and the goal is to use drones and other technologies to protect our borders," Lithuania's interior minister Agnė Bilotaitė reportedly told the Baltic News Service.Further details on when the wall could be in place and how…

27 May

Dissecting Query-Key Interaction in Vision Transformers

stp2y0 CommentsViral News

arXiv:2405.14880v1 Announce Type: new Abstract: Self-attention in vision transformers has been thought to perform perceptual grouping where tokens attend to other tokens with similar embeddings, which could correspond to semantically similar features in an image. However, contextualization is also an important and necessary computation for processing signals. Contextualization potentially requires tokens to attend to dissimilar tokens such as those corresponding to backgrounds or different objects, but this effect has not been reported in previous studies. In this study, we investigate whether self-attention in vision transformers exhibits a preference for attending to similar tokens or dissimilar tokens, providing evidence of perceptual…

27 May

Gemma: Introducing new state-of-the-art open models

stp2y0 CommentsNewsNone

Responsible by designGemma is designed with our AI Principles at the forefront. As part of making Gemma pre-trained models safe and reliable, we used automated techniques to filter out certain personal information and other sensitive data from training sets. Additionally, we used extensive fine-tuning and reinforcement learning from human feedback (RLHF) to align our instruction-tuned models with responsible behaviors. To understand and reduce the risk profile for Gemma models, we conducted robust evaluations including manual red-teaming, automated adversarial testing, and assessments of model capabilities for dangerous activities. These evaluations are outlined in our Model Card.We’re also releasing a new Responsible…

27 May

Alteryx Inspire focuses on innovation in enterprise-grade AI – SiliconANGLE

stp2y0 CommentsChat-GPTRyan Stevens, SiliconANGLE, What to expect during the “Alteryx Inspire” event: Join theCUBE May 15

The role of data and innovation in shaping the future of generative artificial intelligence is critical. All year on theCUBE, a major focus of coverage has been around workflows, data and intellectual property, especially given the recent wave of gen AI, according to Dave Vellante, chief analyst of theCUBE Research. “This AI thing is real. It’s AI everywhere … powering all the trends, all the markets … all the way through the stack,” said Vellante, during the recent “Supercloud 6: AI Innovators” event. The trends at play that are driving innovation in enterprise-grade AI represent a perfect wave for a company such…

27 May

North Korea Says Satellite Launch Failed After Rocket Exploded

stp2y0 CommentsRAG models

A North Korean rocket meant to put a reconnaissance satellite into orbit exploded in mid-air on Monday, the country’s Korean Central News Agency said, citing the space administration. Source link lol

27 May

Extracting Prompts by Inverting LLM Outputs

stp2y0 CommentsViral News

arXiv:2405.15012v1 Announce Type: new Abstract: We consider the problem of language model inversion: given outputs of a language model, we seek to extract the prompt that generated these outputs. We develop a new black-box method, output2prompt, that learns to extract prompts without access to the model's logits and without adversarial or jailbreaking queries. In contrast to previous work, output2prompt only needs outputs of normal user queries. To improve memory efficiency, output2prompt employs a new sparse encoding techique. We measure the efficacy of output2prompt on a variety of user and system prompts and demonstrate zero-shot transferability across different LLMs. Source link…

27 May

Netflix’s Standard with Ads plan grows to 40 million subscribers – gHacks Tech News

stp2y0 CommentsNews

There was a lot of outcry when Netflix announced the introduction of ads on its platform. The streaming service introduced an ad-powered plan in late 2022 in select regions. At $6.99 per month, it was considerably cheaper than non-ad plans, even back in 2022. In fact, adding another user slot to one of the ad-free Netfllix plans is more expensive than subscribing to the ad-powered plan. Netflix did a lot to push the ad-powered plan. First called Basic with ads, it soon was rebranded to Standard with ads, to highlight its advantages over Netflix's basic plan. This was not the…

27 May

AI pioneer LeCun to next-gen AI builders: ‘Don’t focus on LLMs’

stp2y0 CommentsChat-GPT

Join us in returning to NYC on June 5th to collaborate with executive leaders in exploring comprehensive methods for auditing AI models regarding bias, performance, and ethical compliance across diverse organizations. Find out how you can attend here. AI pioneer Yann LeCun kicked off an animated discussion today after telling the next generation of developers not to work on large language models (LLMs). “This is in the hands of large companies, there’s nothing you can bring to the table,” Lecun said at VivaTech today in Paris. “You should work on next-gen AI systems that lift the limitations of LLMs.” The…

27 May

Machine listening: Making speech recognition systems more inclusive

stp2y0 CommentsGenAICommunications; Artificial Intelligence; Information Technology; Computer Science; STEM Education; Racial Disparity; Bioethics; Media and Entertainment

Interactions with voice technology, such as Amazon's Alexa, Apple's Siri, and Google Assistant, can make life easier by increasing efficiency and productivity. However, errors in generating and understanding speech during interactions are common. When using these devices, speakers often style-shift their speech from their normal patterns into a louder and slower register, called technology-directed speech. Research on technology-directed speech typically focuses on mainstream varieties of U.S. English without considering speaker groups that are more consistently misunderstood by technology. In JASA Express Letters, published on behalf of the Acoustical Society of America by AIP Publishing, researchers from Google Research, the University…

27 May

How to create a mind

stp2y0 CommentsNews

This essay explores how cognitive science serves as a blueprint for AI agents, giving us a framework to understand AI developments, pinpoint system gaps, and contrast human and AI minds. We walk through how the key components - perception (data inputs), working memory (context windows), procedural & declarative long-term memory (databases), motor functions (tools), and the orchestrator - all work together. Understanding our minds is a quest that has been pursued since antiquity. Greek philosopher Aristotle explored internal mental processes such as perception, thinking, and memory, laying the groundwork for cognitive psychology. Meanwhile in the East, Indian philosopher Gotama analyzed…