On one screen, an urban professional in Oslo taps through ultra-secure banking apps, relies on an AI-powered personal assistant, and streams media seamlessly over high-speed 5G using their iPhone. On the other screen, a farmer in Malawi scrolls through a modest Android phone -- likely costing less than a week's wages -- just to read the news, check tomorrow's weather, and send WhatsApp messages over a patchy mobile connection. These very different experiences highlight the divide between the Global North and the Global South. These terms refer not only to geographic locations but also to the world's wealthiest and most industrialized regions -- such as Europe, North America, and parts of East Asia -- and economically developing nations across much of Africa, Latin America, South Asia, and Oceania. Technology symbolizes innovation, convenience, and seamless connectivity in the Global North.

Issues with app quality have resulted in mobile users' unwillingness to settle for "best effort" 5G network connectivity, with some generative artificial intelligence (Gen AI) users willing to pay a premium for guaranteed performance. About 35% of 5G users will consider paying more for differentiated connectivity that guarantees better performance for essential tasks, revealed a study released by Ericsson. This number is 1.5 times more in densely populated markets such as Thailand, India, and Brazil, compared to the global average, according to the online survey, which polled more than 23,000 smartphone users, of whom 17,000 were 5G smartphone users from 16 markets, including Australia, France, Singapore, South Korea, Thailand, China, the UK, and the US. Also: 5G to hit 5.6 billion subscribers in 2029, but 4G will remain dominant in three regions "These are mobile-first markets, where reliance on mobile connectivity is particularly strong," Ericsson said. "In these markets, elevated demand for differentiated connectivity could be attributed to the need for improved capacity, rather than general network inadequacy." The telecom equipment manufacturer estimates that the market potential for differentiated connectivity is about 35% of 5G users across the 16 global markets.

When product testers told us that our robo-cars "lacked humanity" and felt like "soulless, uncanny harbingers of doom," we listened, and updated our software to make your driving experience feel more human than ever before. Let's just say, the next time you curse out a maniac swerving recklessly in front of you, it may not be a person you want to run off the road. When it comes to defensive driving, we believe that the best defense is a good offense--and thanks to your valuable feedback, nothing is more offensive than our human-inspired driverless vehicles. New advanced driver-assistance settings range from "Driving Too Slow in the Fast Lane" to "Driving Too Fast in the Slow Lane," and all selections include meaningless lane changes that never actually save any time. There's even an option to run red lights and drive over the speed limit if you're running a little late, there's an emergency, or you simply want to passive-aggressively communicate anger to a difficult passenger.

Accurate spatio-temporal prediction is crucial for the sustainable development of smart cities. However, current approaches often struggle to capture important spatio-temporal relationships, particularly overlooking global relations among distant city regions. Most existing techniques predominantly rely on Convolutional Neural Networks (CNNs) to capture global relations. However, CNNs exhibit neighbourhood bias, making them insufficient for capturing distant relations. To address this limitation, we propose ST-SampleNet, a novel transformer-based architecture that combines CNNs with self-attention mechanisms to capture both local and global relations effectively. Moreover, as the number of regions increases, the quadratic complexity of self-attention becomes a challenge. To tackle this issue, we introduce a lightweight region sampling strategy that prunes non-essential regions and enhances the efficiency of our approach. Furthermore, we introduce a spatially constrained position embedding that incorporates spatial neighbourhood information into the self-attention mechanism, aiding in semantic interpretation and improving the performance of ST-SampleNet. Our experimental evaluation on three real-world datasets demonstrates the effectiveness of ST-SampleNet. Additionally, our efficient variant achieves a 40% reduction in computational costs with only a marginal compromise in performance, approximately 1%.

The proliferation of the Internet of Things (IoT) has heightened the vulnerability to cyber threats, making it imperative to develop Anomaly Detection Systems (ADSs) capable of adapting to emerging or novel attacks. Prior research has predominantly concentrated on offline unsupervised learning techniques to protect ADSs, which are impractical for real-world applications. Furthermore, these studies often rely heavily on the assumption of known legitimate behaviors and fall short of meeting the interpretability requirements in security contexts, thereby hindering their practical adoption. In response, this paper introduces Adaptive NAD, a comprehensive framework aimed at enhancing and interpreting online unsupervised anomaly detection within security domains. We propose an interpretable two-layer anomaly detection approach that generates dependable, high-confidence pseudo-labels. Subsequently, we incorporate an online learning mechanism that updates Adaptive NAD using an innovative threshold adjustment method to accommodate new threats. Experimental findings reveal that Adaptive NAD surpasses state-of-the-art solutions by achieving improvements of over 5.4% and 23.0% in SPAUC on the CIC-Darknet2020 and CIC-DoHBrw-2020 datasets, respectively. The code for Adaptive NAD is publicly available at https://github.com/MyLearnCodeSpace/Adaptive-NAD.

I saw two visions of the future in Los Angeles last weekend. First, a Waymo Jaguar I-PACE pulled over to pick me up on a busy street in downtown L.A., spinning lidar sensors mounted on the hood like a second set of side mirrors. We inched comfortably through stop-and-go Saturday afternoon traffic and made an impressive left turn ahead of two lanes of oncoming cars as I said my prayers in the passenger seat. On the other hand, the robot lost its nerve trying to turn right across a crosswalk. As pedestrians cleared and the light turned from green to yellow to red, the Waymo remained fixed to the spot.

A sleek, gold car pulls up to a bustling corner market, and a middle-aged couple alights. A woman eases a suitcase into the same vehicle's spacious trunk. Later, a doodle and its master watch rocket videos in the front seat as the car eases around the neighborhood. That's the vision shown off by Tesla CEO Elon Musk last week during a presentation broadcast from a set at Warner Bros. Studio, outside of Los Angeles. Some 20 prototypes cruised the movie lot as a series of mocked-up images showed scenes of the idyllic tomorrow these sleek people-movers could usher us into.

Many information systems use tags and keywords to describe and annotate content. These allow for efficient organization and categorization of items, as well as facilitate relevant search queries. As such, the selected set of tags for an item can have a considerable effect on the volume of traffic that eventually reaches an item. In tagging systems where tags are exclusively chosen by an item's owner, who in turn is interested in maximizing traffic, a principled approach for assigning tags can prove valuable. In this paper we introduce the problem of optimal tagging, where the task is to choose a subset of tags for a new item such that the probability of browsing users reaching that item is maximized.

Kamil Khan, Colorado State University; Sudeep Pasricha, Colorado State University Abstract: In emerging high-performance Network-on-Chip (NoC) architectures, efficient power management is crucial to minimize energy consumption. We propose a novel framework called CAFEEN that employs both heuristic-based fine-grained and machine learning-based coarse-grained power-gating for energy-efficient NoCs. CAFEEN uses a fine-grained method to activate only essential NoC buffers during lower network loads. It switches to a coarse-grained method at peak loads to minimize compounding wake-up overhead using multi-agent reinforcement learning. Results show that CAFEEN adaptively balances power-efficiency with performance, reducing total energy by 2.60 for single application workloads and 4.37 for multiapplication workloads, compared to state-of-the-art NoC power-gating frameworks.

However, REs purely rely on expert knowledge and cannot learn from massive ubiquitous network data for automatic management. Today, neural networks (NNs) have shown superior accuracy and flexibility, thanks to their ability to learn from rich labeled data. Nevertheless, NNs are often incompetent in cold-start scenarios and too complex for deployment on network devices. In this paper, we propose Metis, a general framework that converts REs to network device affordable models for superior accuracy and throughput by taking advantage of REs' expert knowledge and NNs' learning ability. In Metis, we convert REs to byte-level recurrent neural networks (BRNNs) without training.