Current benchmarks for evaluating large language models (LLMs) in social media moderation completely overlook a serious threat: covert advertisements, which disguise themselves as regular posts to deceive and mislead consumers into making purchases, leading to significant ethical and legal concerns. In this paper, we present the CHASM, a first-of-its-kind dataset designed to evaluate the capability of Multimodal Large Language Models (MLLMs) in detecting covert advertisements on social media. CHASM3 is a high-quality, anonymized, manually curated dataset consisting of 4,992 instances, based on real-world scenarios from the Chinese social media platform Rednote. The dataset was collected and annotated under strict privacy protection and quality control protocols. It includes many product experience sharing posts that closely resemble covert advertisements, making the dataset particularly challenging.

Current benchmarks for evaluating large language models (LLMs) in social media moderation completely overlook a serious threat: covert advertisements, which disguise themselves as regular posts to deceive and mislead consumers into making purchases, leading to significant ethical and legal concerns. In this paper, we present the CHASM, a first-of-its-kind dataset designed to evaluate the capability of Multimodal Large Language Models (MLLMs) in detecting covert advertisements on social media. CHASM is a high-quality, anonymized, manually curated dataset consisting of 4,992 instances, based on real-world scenarios from the Chinese social media platform Rednote. The dataset was collected and annotated under strict privacy protection and quality control protocols. It includes many product experience sharing posts that closely resemble covert advertisements, making the dataset particularly challenging.The results show that under both zero-shot and in-context learning settings, none of the current MLLMs are sufficiently reliable for detecting covert advertisements.Our further experiments revealed that fine-tuning open-source MLLMs on our dataset yielded noticeable performance gains. However, significant challenges persist, such as detecting subtle cues in comments and differences in visual and textual structures.We provide in-depth error analysis and outline future research directions. We hope our study can serve as a call for the research community and platform moderators to develop more precise defenses against this emerging threat.

Anker's lifestyle brand Eufy has already swallowed a big chunk of the robot vacuum market and now it's got its sights on your yard. The company has been sharing details of its first two robot mowers since the start of the year, and now they're ready to start selling them. Eufy's E15 and E18 are designed to automate one of the most tedious jobs around the home -- if you're able to pay. I've been testing an E15 for the last few weeks ahead of their retail debut today and I'm fairly impressed. Early robot mowers needed a boundary wire to tell them where they were allowed to mow.

The Legend of Zelda: Echoes of Wisdom feels like a kindred spirit to the 2019 remake of Link's Awakening, both in challenge and in vibes. It's a far cry from the incredibly intricate and complex worlds in Tears of the Kingdom, and while I only played for about 90 minutes (spread over two different parts of the game),I came away from the demo charmed by the gorgeous, tilt-shift art style. Not to mention being quite pleased to finally be playing as Zelda for the first time in the series that bears her damn name. And while plenty of adults will surely enjoy The Legend of Zelda: Echoes of Wisdom, it also feels tailor-made as an entry point for younger players. We already knew about the art style and playing as Zelda -- what was most important about this preview was that I got a chance to see just how Zelda's "echoes" worked in the game itself.

If the video generation of our suggested prompt isn't of satisfactory quality, we could decompose the problem further by for example prompting with "a video of Mario expertly jumping over an chasm". In this way every sequence of actions is associated with language, allowing us to ponder actions at multiple resolutions -- for example with the prompt "Mario expertly jumping over a chasm, and then crouching, moving right twice, left once, and then completing the remainder of the level". Can stable diffusion be used to completely resolve abstract action spaces? If Mario expertly jumps over a chasm, what is the atomic input action sequence to the game for this abstract action? There's something missing -- stable diffusion allows us to go from text to video. But now we need to go from text to text, while keeping the video in perspective.

Digital transformation accelerated during the pandemic, as companies of all sizes and in a range of industries invested in advanced, collaborative technologies in a bid to adapt to the new normal. As a result, many more organizations have now established a solid foundation on which to build a digital future. At the heart of this digital transformation is data – every digital interaction generates vast quantities of it. And it can help businesses identify new opportunities, make better decisions, and improve operational efficiency. However, data can only deliver business benefits if it can be discovered, analyzed, and used to generate actionable insights.

Tom Mustill was kayaking with his friend Charlotte in Monterey Bay, California, when an animal three times the size of the largest Tyrannosaurus Rex hurtled from the water and crashed down on their tiny craft. As the flying humpback whale fell upon them and their kayak was sucked beneath the waves, Mustill assumed he would die. Miraculously he and Charlotte found themselves gasping for breath, clinging to their capsized kayak. How had they survived a smash with a creature three times the weight of a double-decker bus? What happened next was almost as weird.

It's highly unlikely that magic is possible. There are apparently two "laws" of magic -- that of similarity and that of contagion. The former amounts to the belief that the practitioner can bring about an effect by doing something which simulates the effect -- for example, mutilating a voodoo doll which looks very similar to a real person will result in the latter undergoing the same mutilations. The second law amounts to the belief that if two objects were in contact, then the influence of each one on the other remains even if the contact later ceases. Think of the same voodoo doll, but containing within it a lock of hair belonging to a person one wants to harm. It should be obvious that material objects don't work this way, because they cannot transcend themselves.

Amid the constant disruption from unlikely competitors and changes in the industry occurring in faster and shorter cycles, time to market is constantly shrinking. To run a business and navigating this complexity in the present day and age, managers need relevant information and insights that can help understand the intended target audience, and their needs and preferences. Thanks to the availability of multiple sources of market and customer data, analytics and artificial intelligence can be used to effectively respond to market dynamics, and drive revenue, profitability, and customer satisfaction. These new technologies are no longer a privilege for tech firms. An increasing number of companies are leveraging these tools to steer through unsettled waters and enhance their performance.

AI is a game changer. And being a data and analytics guy, I could not be more excited about it. The McKinsey research study "Notes from the AI frontier: Applications and value of deep learning" provided some valuable insights into where and how Artificial Intelligence (i.e., Deep Learning / Neural Networks (CNNs, RNNs, GANs), Reinforcement Learning and Deep Reinforcement Learning) will derive and drive new sources of customer, product and operational value, especially when compared to traditional analytic approaches. AI will add billions of dollars of financial and economic value to ALL industries. A no-brainer if ever one existed.