The internet activity records (IARs) of a mobile cellular network posses significant information which can be exploited to identify the network's efficacy and the mobile users' behavior. In this work, we extract useful information from the IAR data and identify a healthy predictability of spatio-temporal pattern within the network traffic. The information extracted is helpful for network operators to plan effective network configuration and perform management and optimization of network's resources. We report experimentation on spatiotemporal analysis of IAR data of the Telecom Italia. Based on this, we present mobile traffic partitioning scheme. Experimental results of the proposed model is helpful in modelling and partitioning of network traffic patterns.

--Action recognition greatly benefits motion understanding in video analysis. Recurrent networks such as long short-term memory (LSTM) networks are a popular choice for motion-aware sequence learning tasks. Recently, a convolutional extension of LSTM was proposed, in which input-to-hidden and hidden-to-hidden transitions are modeled through convolution with a single kernel. This implies an unavoidable tradeoff between effectiveness and efficiency. Herein, we propose a new enhancement to convolutional LSTM networks that supports accommodation of multiple convolutional kernels and layers. This resembles a Network-in-LSTM approach, which improves upon the aforementioned concern. In addition, we propose an attention-based mechanism that is specifically designed for our multi-kernel extension. We evaluated our proposed extensions in a supervised classification setting on the UCF-101 and Sports-1M datasets, with the findings showing that our enhancements improve accuracy. We also undertook qualitative analysis to reveal the characteristics of our system and the convolutional LSTM baseline. CTION recognition is a challenging-yet-essential task in modern computer vision that is typically performed on video clips. Videos are now frequently encountered in our everyday lives on social media platforms such as Instagram, Facebook, and Y ouTube. Many applications can benefit from action recognition; for example, autonomous driving, security and surveillance, and sports analysis. Unlike static images, videos have an inherently spatiotemporal nature. The motion of subjects, such as persons, animals, or objects, carries significant information on the current action.

There is a growing interest in the wireless communications community to complement the traditional model-based design approaches with data-driven machine learning (ML)-based solutions. While conventional ML approaches rely on the assumption of having the data and processing heads in a central entity, this is not always feasible in wireless communications applications because of the inaccessibility of private data and large communication overhead required to transmit raw data to central ML processors. As a result, decentralized ML approaches that keep the data where it is generated are much more appealing. Owing to its privacy-preserving nature, federated learning is particularly relevant for many wireless applications, especially in the context of fifth generation (5G) networks. In this article, we provide an accessible introduction to the general idea of federated learning, discuss several possible applications in 5G networks, and describe key technical challenges and open problems for future research on federated learning in the context of wireless communications.

With the proliferation of training data, distributed machine learning (DML) is becoming more competent for large-scale learning tasks. However, privacy concern has to be attached prior importance in DML, since training data may contain sensitive information of users. Most existing privacy-aware schemes are established based on an assumption that the users trust the server collecting their data, and are limited to provide the same privacy guarantee for the entire data sample. In this paper, we remove the trustworthy servers assumption, and propose a privacy-preserving ADMM-based DML framework that preserves heterogeneous privacy for users' data. The new challenging issue is to reduce the accumulation of privacy losses over ADMM iterations as much as possible. In the proposed privacy-aware DML framework, a local randomization approach, which is proved to be differentially private, is adopted to provide users with self-controlled privacy guarantee for the most sensitive information. Further, the ADMM algorithm is perturbed through a combined noise-adding method, which simultaneously preserves privacy for users' less sensitive information and strengthens the privacy protection of the most sensitive information. Also, we analyze the performance of the trained model according to its generalization error. Finally, we conduct extensive experiments using synthetic and real-world datasets to validate the theoretical results and evaluate the classification performance of the proposed framework.

Influence maximization (IM) is the problem of finding a set of $s$ nodes in a network with maximum influence. We consider models such as the classic Independent Cascade (IC) model of Kempe, Kleinberg, and Tardos \cite{KKT:KDD2003} where influence is defined to be the expectation over {\em simulations} (random sets of live edges) of the size of the reachability set of the seed nodes. In such models the influence function is unbiasedly approximated by an average over a set of i.i.d.\ simulations. A fundamental question is to bound the IM {\em sample complexity}: The number of simulations needed to determine an approximate maximizer. An upper bound of $O( s n \epsilon^{-2} \ln \frac{n}{\delta})$, where $n$ is the number of nodes in the network and $1-\delta$ the desired confidence applies to all models. We provide a sample complexity bound of $O( s \tau \epsilon^{-2} \ln \frac{n}{\delta})$ for a family of models that includes the IC model, where $\tau$ (generally $\tau \ll n$) is the step limit of the diffusion. Algorithmically, we show how to adaptively detect when a smaller number of simulations suffices. We also design an efficient greedy algorithm that computes a $(1-1/e-\epsilon)$-approximate maximizer from simulation averages. Influence maximization from simulation-averages is practically appealing as it is robust to dependencies and modeling errors but was believed to be less efficient than other methods in terms of required data and computation. Our work shows that we can have both robustness and efficiency.

Outlier detection is a fundamental task in data mining and has many applications including detecting errors in databases. While there has been extensive prior work on methods for outlier detection, modern datasets often have sizes that are beyond the ability of commonly used methods to process the data within a reasonable time. To overcome this issue, outlier detection methods can be trained over samples of the full-sized dataset. However, it is not clear how a model trained on a sample compares with one trained on the entire dataset. In this paper, we introduce the notion of resilience to sampling for outlier detection methods. Orthogonal to traditional performance metrics such as precision/recall, resilience represents the extent to which the outliers detected by a method applied to samples from a sampling scheme matches those when applied to the whole dataset. We propose a novel approach for estimating the resilience to sampling of both individual outlier methods and their ensembles. We performed an extensive experimental study on synthetic and real-world datasets where we study seven diverse and representative outlier detection methods, compare results obtained from samples versus those obtained from the whole datasets and evaluate the accuracy of our resilience estimates. We observed that the methods are not equally resilient to a given sampling scheme and it is often the case that careful joint selection of both the sampling scheme and the outlier detection method is necessary. It is our hope that the paper initiates research on designing outlier detection algorithms that are resilient to sampling.

We present here the Temporal Clustering Algorithm (TCA), an incremental learning algorithm applicable to problems of anticipatory computing in the context of the Internet of Things. This algorithm was tested in a specific prediction scenario of consumption of an electric water dispenser typically used in tropical countries, in which the ambient temperature is around 30-degree Celsius. In this context, the user typically wants to drinking iced water therefore uses the cooler function of the dispenser. Real and synthetic water consumption data was used to test a forecasting capacity on how much energy can be saved by predicting the pattern of use of the equipment. In addition to using a small constant amount of memory, which allows the algorithm to be implemented at the lowest cost, while using microcontrollers with a small amount of memory (less than 1Kbyte) available on the market. The algorithm can also be configured according to user preference, prioritizing comfort, keeping the water at the desired temperature longer, or prioritizing energy savings. The main result is that the TCA achieved energy savings of up to 40% compared to the conventional mode of operation of the dispenser with an average success rate higher than 90% in its times of use.

Reinforcement learning agents are prone to undesired behaviors due to reward mis-specification. Finding a set of reward functions to properly guide agent behaviors is particularly challenging in multi-agent scenarios. Inverse reinforcement learning provides a framework to automatically acquire suitable reward functions from expert demonstrations. Its extension to multi-agent settings, however, is difficult due to the more complex notions of rational behaviors. In this paper, we propose MA-AIRL, a new framework for multi-agent inverse reinforcement learning, which is effective and scalable for Markov games with high-dimensional state-action space and unknown dynamics. We derive our algorithm based on a new solution concept and maximum pseudolikelihood estimation within an adversarial reward learning framework. In the experiments, we demonstrate that MA-AIRL can recover reward functions that are highly correlated with ground truth ones, and significantly outperforms prior methods in terms of policy imitation.

Recently, the posit numerical format has shown promise for DNN data representation and compute with ultra-low precision ([5..8]-bit). However, majority of studies focus only on DNN inference. In this work, we propose DNN training using posits and compare with the floating point training. We evaluate on both MNIST and Fashion MNIST corpuses, where 16-bit posits outperform 16-bit floating point for end-to-end DNN training.

The recent decades have seen a surge of interests in distributed computing. Existing work focus primarily on either distributed computing platforms, data query tools, or, algorithms to divide big data and conquer at individual machines etc. It is, however, increasingly often that the data of interest are inherently distributed, i.e., data are stored at multiple distributed sites due to diverse collection channels, business operations etc. We propose to enable learning and inference in such a setting via a general framework based on the distortion minimizing local transformations. This framework only requires a small amount of local signatures to be shared among distributed sites, eliminating the need of having to transmitting big data. Computation can be done very efficiently via parallel local computation. The error incurred due to distributed computing vanishes when increasing the size of local signatures. As the shared data need not be in their original form, data privacy may also be preserved. Experiments on linear (logistic) regression and Random Forests have shown promise of this approach. This framework is expected to apply to a general class of tools in learning and inference with the continuity property.