Meta AI, the most interesting thing you can do with Ray-Ban Meta glasses, will soon be available to more people. The company's Live Translation feature is rolling out to all the product's markets, and Live AI (where you can hold a free-flowing conversation about what you're looking at) will soon be available in the US and Canada. In addition, glasses owners in the EU can finally use Meta AI with their high-tech specs. Live translation, previously available in early access, is now rolling out in every region where Ray-Ban Meta glasses are available. Handy for trips abroad or chats with locals who speak a different language, the AI-powered feature speaks a translation in your preferred language in real time.

This paper introduces a new IncidentAI dataset for safety prevention. Different from prior corpora that usually contain a single task, our dataset comprises three tasks: named entity recognition, cause-effect extraction, and information retrieval. The dataset is annotated by domain experts who have at least six years of practical experience as high-pressure gas conservation managers. We validate the contribution of the dataset in the scenario of safety prevention. Preliminary results on the three tasks show that NLP techniques are beneficial for analyzing incident reports to prevent future failures. The dataset facilitates future research in NLP and incident management communities. The access to the dataset is also provided (the IncidentAI dataset is available at: https://github.com/Cinnamon/incident-ai-dataset).

Deep neural networks are vulnerable to adversarial attacks. Ideally, a robust model shall perform well on both the perturbed training data and the unseen perturbed test data. It is found empirically that fitting perturbed training data is not hard, but generalizing to perturbed test data is quite difficult. To better understand adversarial generalization, it is of great interest to study the adversarial Rademacher complexity (ARC) of deep neural networks. However, how to bound ARC in multi-layers cases is largely unclear due to the difficulty of analyzing adversarial loss in the definition of ARC. There have been two types of attempts of ARC. One is to provide the upper bound of ARC in linear and one-hidden layer cases. However, these approaches seem hard to extend to multi-layer cases. Another is to modify the adversarial loss and provide upper bounds of Rademacher complexity on such surrogate loss in multi-layer cases. However, such variants of Rademacher complexity are not guaranteed to be bounds for meaningful robust generalization gaps (RGG). In this paper, we provide a solution to this unsolved problem. Specifically, we provide the first bound of adversarial Rademacher complexity of deep neural networks. Our approach is based on covering numbers. We provide a method to handle the robustify function classes of DNNs such that we can calculate the covering numbers. Finally, we provide experiments to study the empirical implication of our bounds and provide an analysis of poor adversarial generalization.

We're on a mission to make it possible for every person, team, and company to be able to tailor their software to solve any problem and take on any challenge. Computers may be our most powerful tools, but most of us can't build or modify the software we use on them every day. At Notion, we want to change this with focus, design, and craft. We've been working on this together since 2016, and have customers like Pixar, Mitsubishi, Figma, Plaid, Match Group, and thousands more on this journey with us. Today, we're growing fast and excited for new teammates to join us who are the best at what they do.

Find open roles in Artificial Intelligence (AI), Machine Learning (ML), Natural Language Processing (NLP), Computer Vision (CV), Data Engineering, Data Analytics, Big Data, and Data Science in general, filtered by job title or popular skill, toolset and products used.

LYNX Technik will showcase its new Instant Dialog Cleaner yellobrik module--which is designed to isolate speech & dialog by removing complex background noises in live broadcasts as well as recorded audio--at IBC2022 on Stand 10.A10. The IDC1411 incorporates powerful Deep Neural Networks (DNN) artificial intelligence (AI) technology from Audionamix and resolves many audio challenges that arise in broadcast and professional AV environments, according to the company. With the help of this technology, the IDC 141 analyzes audio and removes (in real-time) background noises, e.g., road and air traffic, nature sounds, music and other ambient noises. This audio cleaning process separates the dialog and speech from the noises, providing a crisp audio signal that only contains the spoken word. The AI technology automatically detects the voice regardless of the surrounding noises and preserves the real integrity of the dialog, unlike other available noise reduction tools, which the company says are often not adequate.

I am delighted to present my new blog – AI Business Transformation Playbook for Executives. originally posted here. I get into the nuts-and-bolts of AI Systems Solutioning in this rather lengthy blog but the “First Ten Plays” at the end summarizes the key steps. I look forward to your thoughts and comments. – PG “AI, IoT &… Read More »AI Business Transformation Playbook for Executives

Daniel Shearly answers questions about data-centricity, trusted data, and how data-driven intelligence can help in better decision-making for future-proofing businesses. What specific trends do you see that will shape the future of its adoption? Every organization has its own reasons for adopting AI which have historically ranged from a real desire to answer complex problems and uncover insights from large data sets that would be impossible to process through traditional statistical methods to driving operational efficiencies and even just a company's desire to appear cutting edge. As AI is now much more mainstream it's less of a buzzword and badge of innovation and more of something to be applied practically to real-world problems. The main trends I see emerging are the broader adoption of it by companies big and small as knowledge and talent build and barriers to adoption decrease.

When I was the Vice President of Advertiser Analytics at Yahoo!, I painfully learned that my targeted user personas (Media Planners & Buyers and Campaign Managers) didn't want more data in helping them optimize their marketing, campaign, and advertising spend across the Yahoo! Heck, they didn't even want analytics! The aspirations for these personas were to become VPs of Digital Marketing or Directors of Social Media Marketing or Digital Advertising Executives. The last thing they wanted was to become data analysts who had to churn through massive data sets to uncover the audience, messaging, and campaign insights needed to achieve their marketing, campaign, and advertising objectives. In fact, that's what they wanted Yahoo! to do for them!

Summary: The Transcarent Analytics and Reporting team empowers the company to make data-driven decisions and provides critical business insights that help us execute in a world-class fashion. We are seeking a talented and motivated technical expert to accelerate our efforts to drive trust, adoption, and democratization of insights externally to our clients and members as well as internally to our stakeholders and leaders. The Analytics Data Engineer will work closely with our Data Engineering, Product, and Solution stakeholders to deliver well-defined, transformed, tested, documented, and code-reviewed analytical datasets that would be single source of truth for reliable, and consistent results. The ideal candidate will have broad skills in database design, be comfortable dealing with large and complex data sets, have experience building self-service dashboards, and applying analytical rigor to inform business decisions. They will work very efficiently to deliver the right solutions and continuously think of how to best automate and expand the delivery of the analytical solutions.