Abstract
Deep neural networks (DNN) are used to analyze images, videos, signals and texts require a lot of memory and intensive computing power. For example, the very successful GPT4 model contains more than a few trillion parameters. Although such models are of great impact, but they have been used very little in real-world applications, including industrial Internet of Things, self-driving cars, algorithmic health monitoring for use in limited mobile or edge devices. The requirement to run large models on resource-constrained peripherals has led to significant research interest in compressing DNN models. Signal processing researchers have traditionally advocated data (image/video/audio) compression, and by the way, many of these techniques are used for DNN compression. For example, source coding is a basic technique that has been widely used to compress various DNN models. In this paper, we present our views on the use of signal processing methods for DNN model compression.

Abstract
Effective anomaly detection in telecommunication networks is essential for securing digital transactions and supporting the sustainability of our global information ecosystem. However, the volume of data in such high-speed distributed environments imposes strict latency and scalability requirements on anomaly detection systems. This study focuses on distributed heavy hitter detection in telephone networks - a critical component of network traffic analysis and fraud detection. We propose a federated version of the Lossy Counting algorithm and compare it to its centralized version. Our experimental results reveal that the federated approach can detect considerably more unique heavy hitters than the centralized method while enhancing privacy. Furthermore, Federated Lossy Counting does not need a large amount of centralized processing power since it can leverage the networked infrastructure with minimal impact on bandwidth and computing power. © 2024 The Authors.

Abstract
Long-tailed data are a special type of multiclass imbalanced data with a very large amount of minority/tail classes that have a very significant combined influence. Long-tailed learning (LTL) aims to build high-performance models on datasets with long-tailed distributions that can identify all the classes with high accuracy, in particular the minority/tail classes. It is a cutting-edge research direction that has attracted a remarkable amount of research effort in the past few years. In this article, we present a comprehensive survey of the latest advances in long-tailed visual learning. We first propose a new taxonomy for LTL, which consists of eight different dimensions, including data balancing, neural architecture, feature enrichment, logits adjustment, loss function, bells and whistles, network optimization, and posthoc processing techniques. Based on our proposed taxonomy, we present a systematic review of LTL methods, discussing their commonalities and alignable differences. We also analyze the differences between imbalance learning and LTL. Finally, we discuss prospects and future directions in this field. © 2012 IEEE.

Abstract
The classic literature about innovation conveys innovation strategy the leading and starting role to generate business growth due to technology development and more effective managerial practices. The advent of Artificial Intelligence (AI) however reverts this paradigm in the context of Industry 5.0. The focus is moving from how innovation fosters AI to how AI fosters innovation. Therefore, our research question can be stated as follows: What factors influence the effect of AI on Innovation Capacity in the context of Industry 5.0? To address this question we conduct a scoping review of a vast body of literature spanning engineering, human sciences, and management science. We conduct a keyword-based literature search completed by bibliographic analysis, then classify the resulting 333 works into 3 classes and 15 clusters which we critically analyze. We extract 3 hypotheses setting associations between 4 factors: company age, AI maturity, manufacturing strategy, and innovation capacity. The review uncovers several debates and research gaps left unsolved by the existing literature. In particular, it raises the debate whether the Industry5.0 promise can be achieved while Artificial General Intelligence (AGI) remains out of reach. It explores diverging possible futures driven toward social manufacturing or mass customization. Finally, it discusses alternative AI policies and their incidence on open and internal innovation. We conclude that the effect of AI on innovation capacity can be synergic, deceptive, or substitutive depending on the alignment of the uncovered factors. Moreover, we identify a set of 12 indicators enabling us to measure these factors to predict AI's effect on innovation capacity. These findings provide researchers with a new understanding of the interplay between artificial intelligence and human intelligence. They provide practitioners with decision metrics for a successful transition to Industry 5.0.

Abstract
Recommendation systems play a crucial role in modern e-commerce and streaming services. However, the limited availability of public datasets hampers the rapid development of more efficient and accurate recommendation algorithms within the research community. This work introduces a stream-based data generator designed to generate user preferences for a set of items while accommodating progressive changes in user preferences. The underlying principle involves using user/item embeddings to derive preferences by exploring the proximity of these embeddings. Whether randomly generated or learned from a real finite data stream, these embeddings serve as the basis for generating new preferences. We investigate how this fundamental model can adapt to shifts in user behavior over time; in our framework, changes correspond to alterations in the structure of the tripartite graph, reflecting modifications in the underlying embeddings. Through an analysis of real-life data streams, we demonstrate that the proposed model is effective in capturing actual preferences and the changes that they can exhibit over time. Thus, we characterize these changes and develop a generalized method capable of simulating realistic data, thereby generating streams with similar yet controllable drift dynamics.

Abstract