This week at CES, Google presented an early look at new software and hardware upgrades coming to Google TV devices. The new features include the integration of Gemini, Google's AI model, to the Google Assistant, as well as a new ambient experience. New smart TVs with Google TV will also gain far-field mics and proximity sensors to support the new software perks. If you've used a Google TV or Google streaming device, you may have already used the "hey Google" prompt to search for shows to watch. With the addition of Gemini, those "conversations" should now feel more natural.

Smart glasses are traditionally long on promise, short on delivery, especially at these sorts of consumer electronics shindigs. There's always a steady stream of companies promising we're on the cusp of having our very own Gary-from-Veep attached to our faces before fading away. The weight of promises Halliday has laid upon the table is a sign of braggadocio, but it'll take a while before we know if it's deserved or not. Halliday has turned up at CES 2025 in Las Vegas with a pair of eponymous smart glasses filled to the brim with technology. There's a waveguide display in the right eyecup that will project the equivalent of a 3.5-inch screen into the wearer's view.

People feel attached to places that are meaningful to them, which psychological research calls "place attachment." Place attachment is associated with self-identity, self-continuity, and psychological well-being. Even small cues, including videos, images, sounds, and scents, can facilitate feelings of connection and belonging to a place. Telepresence robots that allow people to see, hear, and interact with a remote place have the potential to establish and maintain a connection with places and support place attachment. In this paper, we explore the design space of robotic telepresence to promote place attachment, including how users might be guided in a remote place and whether they experience the environment individually or with others. We prototyped a telepresence robot that allows one or more remote users to visit a place and be guided by a local human guide or a conversational agent. Participants were 38 university alumni who visited their alma mater via the telepresence robot. Our findings uncovered four distinct user personas in the remote experience and highlighted the need for social participation to enhance place attachment. We generated design implications for future telepresence robot design to support people's connections with places of personal significance.

Personalized fashion recommendation is a difficult task because 1) the decisions are highly correlated with users' aesthetic appetite, which previous work frequently overlooks, and 2) many new items are constantly rolling out that cause strict cold-start problems in the popular identity (ID)-based recommendation methods. These new items are critical to recommend because of trend-driven consumerism. In this work, we aim to provide more accurate personalized fashion recommendations and solve the cold-start problem by converting available information, especially images, into two attribute graphs focusing on optimized image utilization and noise-reducing user modeling. Compared with previous methods that separate image and text as two components, the proposed method combines image and text information to create a richer attributes graph. Capitalizing on the advancement of large language and vision models, we experiment with extracting fine-grained attributes efficiently and as desired using two different prompts. Preliminary experiments on the IQON3000 dataset have shown that the proposed method achieves competitive accuracy compared with baselines.

Large Language Model (LLM) has transformative potential in various domains, including recommender systems (RS). There have been a handful of research that focuses on empowering the RS by LLM. However, previous efforts mainly focus on LLM as RS, which may face the challenge of intolerant inference costs by LLM. Recently, the integration of LLM into RS, known as LLM-Enhanced Recommender Systems (LLMERS), has garnered significant interest due to its potential to address latency and memory constraints in real-world applications. This paper presents a comprehensive survey of the latest research efforts aimed at leveraging LLM to enhance RS capabilities. We identify a critical shift in the field with the move towards incorporating LLM into the online system, notably by avoiding their use during inference. Our survey categorizes the existing LLMERS approaches into three primary types based on the component of the RS model being augmented: Knowledge Enhancement, Interaction Enhancement, and Model Enhancement. We provide an in-depth analysis of each category, discussing the methodologies, challenges, and contributions of recent studies. Furthermore, we highlight several promising research directions that could further advance the field of LLMERS.

If one of your New Year's Resolutions was to get back on the dating scene, today is the day to finally bite the bullet. January 5 is'Dating Sunday' - the first Sunday in January, which is annually recognised as the busiest day for dating apps. On this day, apps including Hinge, Tinder, and Bumble see substantial increases in activity, with users more actively seeking and initiating conversations. 'As we move into the first days of the new year, many people find themselves reflecting on the past year and visioning for the year ahead,' said Moe Ari Brown, Hinge's Love and Connection expert. 'The new year marks a shift from one chapter to the next and offers us the opportunity for a reboot.

As Integrated Development Environments (IDEs) increasingly integrate Artificial Intelligence, Software Engineering faces both benefits like productivity gains and challenges like mismatched user preferences. We propose Hyper-Dimensional (HD) vector spaces to model Human-Computer Interaction, focusing on user actions, stylistic preferences, and project context. These contributions aim to inspire further research on applying HD computing in IDE design.

Good mental health enables individuals to cope with the normal stresses of life. In Germany, approximately one-quarter of the adult population is affected by mental illnesses. Teletherapy and digital health applications are available to bridge gaps in care and relieve healthcare professionals. The acceptance of these tools is a strongly influencing factor for their effectiveness, which also needs to be evaluated for AI-based conversational agents (CAs) (e. g. ChatGPT, Siri) to assess the risks and potential for integration into therapeutic practice. This study investigates the perspectives of both the general population and healthcare professionals with the following questions: 1. How frequently are CAs used for mental health? 2. How high is the acceptance of CAs in the field of mental health? 3. To what extent is the use of CAs in counselling, diagnosis, and treatment acceptable? To address these questions, two quantitative online surveys were conducted with 444 participants from the general population and 351 healthcare professionals. Statistical analyses show that 27 % of the surveyed population already confide their concerns to CAs. Not only experience with this technology but also experience with telemedicine shows a higher acceptance among both groups for using CAs for mental health. Additionally, participants from the general population were more likely to support CAs as companions controlled by healthcare professionals rather than as additional experts for the professionals. CAs have the potential to support mental health, particularly in counselling. Future research should examine the influence of different communication media and further possibilities of augmented intelligence. With the right balance between technology and human care, integration into patient-professional interaction can be achieved.

We present HiRMed (Hierarchical RAG-enhanced Medical Test Recommendation), a novel tree-structured recommendation system that leverages Retrieval-Augmented Generation (RAG) for intelligent medical test recommendations. Unlike traditional vector similarity-based approaches, our system performs medical reasoning at each tree node through a specialized RAG process. Starting from the root node with initial symptoms, the system conducts step-wise medical analysis to identify potential underlying conditions and their corresponding diagnostic requirements. At each level, instead of simple matching, our RAG-enhanced nodes analyze retrieved medical knowledge to understand symptom-disease relationships and determine the most appropriate diagnostic path. The system dynamically adjusts its recommendation strategy based on medical reasoning results, considering factors such as urgency levels and diagnostic uncertainty. Experimental results demonstrate that our approach achieves superior performance in terms of coverage rate, accuracy, and miss rate compared to conventional retrieval-based methods. This work represents a significant advance in medical test recommendation by introducing medical reasoning capabilities into the traditional tree-based retrieval structure.

Micro-video recommendation is attracting global attention and becoming a popular daily service for people of all ages. Recently, Graph Neural Networks-based micro-video recommendation has displayed performance improvement for many kinds of recommendation tasks. However, the existing works fail to fully consider the characteristics of micro-videos, such as the high timeliness of news nature micro-video recommendation and sequential interactions of frequently changed interests. In this paper, a novel Multi-aggregator Time-warping Heterogeneous Graph Neural Network (MTHGNN) is proposed for personalized news nature micro-video recommendation based on sequential sessions, where characteristics of micro-videos are comprehensively studied, users' preference is mined via multi-aggregator, the temporal and dynamic changes of users' preference are captured, and timeliness is considered. Through the comparison with the state-of-the-arts, the experimental results validate the superiority of our MTHGNN model.