Hyper-relational Knowledge Graphs (HRKGs) extend traditional KGs beyond binary relations, enabling the representation of contextual, provenance, and temporal information in domains, such as historical events, sensor data, video content, and narratives. HRKGs can be structured using several Metadata Representation Models (MRMs), including Reification (REF), Singleton Property (SGP), and RDF-star (RDR). However, the effects of different MRMs on KG Embedding (KGE) and Link Prediction (LP) models remain unclear. This study evaluates MRMs in the context of LP tasks, identifies the limitations of existing evaluation frameworks, and introduces a new task that ensures fair comparisons across MRMs. Furthermore, we propose a framework that effectively reflects the knowledge representations of the three MRMs in latent space. Experiments on two types of datasets reveal that REF performs well in simple HRKGs, whereas SGP is less effective. However, in complex HRKGs, the differences among MRMs in the LP tasks are minimal. Our findings contribute to an optimal knowledge representation strategy for HRKGs in LP tasks.

A recent paper proposes Dynamic Tanh (DyT) as a drop-in replacement for layer normalization (LN). Although the method is empirically well-motivated and appealing from a practical point of view, it lacks a theoretical foundation. In this work, we shed light on the mathematical relationship between layer normalization and dynamic activation functions. In particular, we derive DyT from LN and show that a well-defined approximation is needed to do so. By dropping said approximation, an alternative activation function is obtained, which we call Dynamic Inverse Square Root Unit (DyISRU). DyISRU is the exact counterpart of layer normalization, and we demonstrate numerically that it indeed resembles LN more accurately than DyT does.

Autonomous navigation robots can increase the independence of blind people but often limit user control, following what is called in Japanese an "omakase" approach where decisions are left to the robot. This research investigates ways to enhance user control in social robot navigation, based on two studies conducted with blind participants. The first study, involving structured interviews (N=14), identified crowded spaces as key areas with significant social challenges. The second study (N=13) explored navigation tasks with an autonomous robot in these environments and identified design strategies across different modes of autonomy. Participants preferred an active role, termed the "boss" mode, where they managed crowd interactions, while the "monitor" mode helped them assess the environment, negotiate movements, and interact with the robot. These findings highlight the importance of shared control and user involvement for blind users, offering valuable insights for designing future social navigation robots.

While large language model (LLM)-based chatbots have been applied for effective engagement in credit dialogues, their capacity for dynamic emotional expression remains limited. Current agents primarily rely on passive empathy rather than affective reasoning. For instance, when faced with persistent client negativity, the agent should employ strategic emotional adaptation by expressing measured anger to discourage counterproductive behavior and guide the conversation toward resolution. This context-aware emotional modulation is essential for imitating the nuanced decision-making of human negotiators. This paper introduces an EQ-negotiator that combines emotion sensing from pre-trained language models (PLMs) with emotional reasoning based on Game Theory and Hidden Markov Models. It takes into account both the current and historical emotions of the client to better manage and address negative emotions during interactions. By fine-tuning pre-trained language models (PLMs) on public emotion datasets and validating them on the credit dialogue datasets, our approach enables LLM-based agents to effectively capture shifts in client emotions and dynamically adjust their response tone based on our emotion decision policies in real-world financial negotiations. This EQ-negotiator can also help credit agencies foster positive client relationships, enhancing satisfaction in credit services.

This is INSANEEEEEEEEEEEEEE! Hands down, this is one of the first real AI innovations we've seen this year that lives up to the AI hype. OpenAI just dropped native image generation in GPT-4o, and I don't say this lightly, but... This ish right here is a game changer (I said that using my best Katt Williams impression). Also: I tried ChatGPT's new image generator, and it shattered my expectations You don't need professional design skills, expensive software, or even an ounce of artistic talent to produce jaw-dropping visuals. All you really need is your imagination and the right prompt.

The attention put a rocket under his music career. While his first two albums were DIY affairs, recorded in a couple of days and self-released, his latest, The Crux, was created in New York's fabled Electric Lady studios. Released on Friday, it's packed full of off-kilter lyrics and squiggly synth lines that burrow into your brain. The first two singles, Delete Ya and Basic Being Basic have already been radio hits, and the rest of the album pulls on influences as diverse as Electric Light Orchestra, New Order, Cake, Hall & Oates and Bruce Springsteen (coincidentally, all bands that would work perfectly on the Stranger Things soundtrack). There are a couple of knockouts – including the crunchy garage rock of Gap Toothed Smile, and the choppy New Wave anthem Link – but the point of the album is its diversity.

An AI overhaul may be on the horizon for Apple's Health app. In the Power On newsletter, Bloomberg's Mark Gurman reports that Apple is working on a much more comprehensive version of its Health app under the code name Project Mulberry, with plans to integrate an AI agent that would somewhat "replicate" a doctor and act as a personal health coach. In addition to making lifestyle recommendations based on users' health data, the app will reportedly include educational videos from real doctors about an array of health topics. The Health app will also put a new emphasis on food tracking, and may even offer form correction tips for workouts using the device's camera, Gurman reports. The service, unofficially being referred to as Health, could arrive with iOS 19.4, which Gurman says is expected to be released next spring or summer. In the meantime, Apple reportedly has doctors on staff whose data is being used to train the AI agent, and it's planning to open a studio near Oakland, California where they can film content.

Russian drones have struck Ukraine's eastern city of Kharkiv, killing two people and wounding dozens, the city's mayor Ihor Terekhov has said. "For the second time in a week, the enemy launched a combined attack, launching seven'Shaheed' at residential areas, hospitals, and the city's infrastructure," Terekhov said in a Telegram message on Sunday, referring to Iranian-made Shahed drones. The swarm of drones also targeted a military hospital, a shopping centre and apartment blocks, he said. Five of the 35 people wounded in the attack overnight were children. At least 13 have been hospitalised, including a teenage girl who is in serious condition.

Non-paying Gemini users can now play around with Google's newest model, the experimental version of Gemini 2.5 Pro. The company announced this weekend that it's making Gemini 2.5 Pro (experimental) free for everyone to use, albeit with tighter rate limits for non-subscribers. Google introduced Gemini 2.5 Pro just last week, touting it as its "most intelligent AI model" yet, and rolled it out to Gemini Advanced users first. While free users can now try it out too, Google added that "Gemini Advanced users have expanded access and a significantly larger context window." Gemini 2.5 Pro (experimental) is the first of Google's Gemini 2.5 "thinking" models, which are said to deliver more accurate results through reasoning.

The impact of AI has been felt across industries from Hollywood to publishing – and now it's come for modelling. H&M announced last week that it would create AI "twins" of 30 models with the intention of using them in social media posts and marketing imagery if the model gives her permission. In a statement, Jörgen Andersson, the chief creative officer at H&M, described the idea as "something that will enhance our creative process and how we work with marketing but fundamentally not change our human-centric approach in any way". The retail giant has worked with successful models including Vilma Sjöberg and Mathilda Gvarliani, who model for Vogue and brands including Chanel. As part of the agreement, each model would be able book her twin on shoots for other brands – meaning they could, in image anyway, be in two places at the same time.