stp2y - CybAI news

06 Aug

Why AI’s Tom Cruise problem means it is ‘doomed to fail’

stp2y0 CommentsChat-GPT

In 2021, linguist Emily Bender and computer scientist Timnit Gebru published a paper that described the then-nascent field of language models as one of “stochastic parrots”. A language model, they wrote, “is a system for haphazardly stitching together sequences of linguistic forms it has observed in its vast training data, according to probabilistic information about how they combine, but without any reference to meaning.”The phrase stuck. AI can still get better, even if it is a stochastic parrot, because the more training data it has, the better it will seem. But does something like ChatGPT actually display anything like intelligence,…

06 Aug

This electric plane promises to fly 90 passengers 500 miles on a single charge

stp2y0 CommentsRAG models

Elysian, an electric plane startup based in the Netherlands, thinks it has found a way to change that; a novel new battery-powered jet that it says could fly 90 passengers up to 500 miles on a single charge."Part of the reason why the aerospace industry has been so skeptical about large electric aircraft is because to electrify a contemporary plane, you had to make very wild assumptions about how battery technology will develop," Elysian co-CEO Daniel Rosen Jacobson said.The startup has attempted to get around this problem, he said, by rethinking what an airplane should look like to squeeze every…

06 Aug

VLG-CBM: Training Concept Bottleneck Models with Vision-Language Guidance

stp2y0 CommentsViral News

[Submitted on 18 Jul 2024] View a PDF of the paper titled VLG-CBM: Training Concept Bottleneck Models with Vision-Language Guidance, by Divyansh Srivastava and 2 other authors View PDF HTML (experimental) Abstract:Concept Bottleneck Models (CBMs) provide interpretable prediction by introducing an intermediate Concept Bottleneck Layer (CBL), which encodes human-understandable concepts to explain models' decision. Recent works proposed to utilize Large Language Models (LLMs) and pre-trained Vision-Language Models (VLMs) to automate the training of CBMs, making it more scalable and automated. However, existing approaches still fall short in two aspects: First, the concepts predicted by CBL often mismatch the input image,…

06 Aug

The best Apple AirTag accessories for 2024

stp2y0 CommentsNewsAirtags, apple, best tech, Commerce, evergreen, feature, gear, shopping, streamshopping, tech accessories, thebuyersguide, wearables, yir2023

It can be hard to keep track of all your belongings with our busy lives, work, kids — you name it. The Apple AirTag is an affordable little Bluetooth tracker that can be placed on or inside your bag or luggage, attached to your pet’s collar or your keys, or anything else you might have a habit of losing. But to make the most of your AirTag, it’s worth investing in the best AirTag accessories, like a keychain, holder or case. Convenience is key here, and we’ve tested dozens of AirTag accessories to help you choose the best AirTag holders…

06 Aug

Mitsubishi Heavy Shares Soar as Turbine Engine Sales Help Profit

stp2y0 CommentsRAG modelsAsia, BRICS, china, Emerging Markets, Hisato Kozawa, Industrials, Japanese Yen Spot, Markets, MITSUBISHI HEAVY INDUSTRIES, science, Technology

Mitsubishi Heavy Industries Ltd. shares jumped the most on record after the manufacturing conglomerate posted a 61% increase in quarterly profit, helped by strong sales of gas turbine engines.The Japanese company reported a business profit of ¥83.5 billion ($572 million) for the three months ended June 30, according to a statement Tuesday, after selling six large-scale gas turbines in the US and Asia. That beat analyst estimates of ¥57 billion. Net sales were ¥1.1 trillion, slightly higher than what the market was looking for. Source link lol

06 Aug

Transforming Slot Schema Induction with Generative Dialogue State Inference

stp2y0 CommentsViral News

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. Source link lol

06 Aug

Web-to-App Attribution Measurement Using Mogua

stp2y0 CommentsNewsandroiddev, coding, community, development, engineering, inclusive, ios, javascript, mobile, software

This article introduces how to quickly view the effectiveness of different marketing channels through the Mogua SDK and Dashboard statistics panel. Mogua is a lightweight deep linking solution designed for indie apps like mobile games and tool apps, used for web-to-app attribution statistics. Mogua collects device data from both web and app, matching them to enable parameter passing from the webpages to the app. This ensures app users are correctly attributed to different marketing channels. More details on how the SDK can refer to this article. To use Mogua's products, you need to first create an application and integrate the…

06 Aug

Cognite Appoints Bill Hendricks as President of Cognite Americas

stp2y0 CommentsAI

Momentum in the Americas for Cognite’s Data and Gen AI solutions is driving expanded coverage and customer success investments Cognite,the globally recognized authority in Data and AI for industry, today announced that Bill Hendricks has been appointed President of Cognite Americas. Hendricks’ appointment fuels Cognite’s momentum, leading the charge in global Industrial Data Operations and AI for Industry with Cognite Atlas AI. Cognite’s offerings enable Energy, Process, and Manufacturing companies to leverage AI to advance their transformative digital journey. Hendricks is based in Cognite’s Houston office, where he will support the company’s growth and strengthen its position as a leading AI-powered…

06 Aug

Riverlane raises $75M for quantum computing error correction

stp2y0 CommentsChat-GPT

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Riverlane has raised $75 million in funding to build its quantum error correction technology. The Series C funding will help the company deliver on its ambitious quantum error correction (QEC) roadmap. If you’re not familiar with why a company can raise money for correcting quantum computing errors, take note. Microsoft Copilot tells me it’s like a cosmic spellcheck for quantum computers. It’s crucial because quantum computing is like a high-stakes game of Jenga played on a wobbly table. The building blocks…

06 Aug

Mechanistic Anomaly Detection Research Update

stp2y0 CommentsGenAI

In December 2023, the Eleuther team published Eliciting Latent Knowledge from Quirky Language Models. We finetuned language models to behave in a "quirky" manner on a collection of question and answer datasets. When a prompt began with "Alice:", these models were trained to answer as accurately as possible, but when it instead began with "Bob:", they would answer according to an unreliable heuristic (Bob would not always be wrong, but would consistently use the same fallible method to answer questions). One problem we investigated was detecting when the model was behaving in an "Alice"-like way vs when it was behaving…