Armen Aghasaryan

• Ph.D.
• Head of Department at Nokia

Modern Artificial Intelligence (AI) technologies, led by Machine Learning (ML), have gained unprecedented momentum over the past decade. Following this wave of ``AI summer'', the network research community has also embraced AI/ML algorithms to address many problems related to network operations and management. However, compared to their counterpart...

Modern artificial intelligence (AI) technologies, led by machine learning (ML), have gained unprecedented momentum over the past decade. Following this wave of "AI summer," the network research community has also embraced AI/ML algorithms to address many problems related to network operations and management. However, compared to their counterparts...

Maintaining and managing ever more complex telecommunication networks is an increasingly difficult task, which often challenges the capabilities of human experts. There is a consensus both in academia and in the industry on the need to enhance human capabilities with sophisticated algorithmic tools for decision-making, with the aim of transitioning...

Complex systems, such as communication networks, generate thousands of new data points about the system state every minute. Even if faults are rare events, they can easily propagate, which makes it challenging to distinguish root causes of errors from effects among the thousands of highly correlated alerts appearing simultaneously in high volumes o...

The digital system of the future will face the growing challenge of controlling the system behavior in complex dynamically evolving environments. In this paper, we examine the applicability of a new management paradigm based on a reinforcement learning approach, where no preliminary specification of the system model is required. The learning agent...

In the era of growing digitalization, dynamic resource management becomes one of the critical problems in many application fields where, due to the permanently evolving environment, the trade-off between cost and system performance needs to be continuously adapted. While traditional approaches based on prior system specification or model learning a...

The digital system of the future will face the growing challenge of controlling the system behavior in complex dynamically evolving environments. In this paper, we examine the applicability of a new management paradigm based on Reinforcement Learning approach, where no preliminary specification of the system model is required. In contrast, the lear...

In this paper, we present an approach for automated profiling of cloud-based distributed applications. The failure dependencies within or between application nodes can be methodically elucidated by dynamically applying a series of unitary perturbations on the underlying computing resources. Each such perturbation in a node acts as a stimulus which...

The “privacy versus personalization” dilemma refers to the situation in which it is necessary for users to disclose their sensitive personal data in order to benefit from collaborative personalized services. Solving this dilemma is a challenge because generating collaborative filtering recommendations requires access to the set of all user profiles...

We address the problem of alarm correlation in large distributed systems. The key idea is to make use of the concurrence of events in order to separate and simplify the state estimation in a faulty system. Petri nets and their causality semantics are used to model concurrency. Special partially stochastic Petri nets are developed, that establish so...

We present the design, implementation, and evaluation of a decentralized framework for enabling privacy in Web-scale recommendation services. Our framework, which comprises of a decentralized middleware that is hosted and run by federated entities, is designed to support collaborative-filtering and content-based recommendations. We design a novel d...

The Locality Sensitive Hashing (LSH) technique of scalably finding nearest-neighbors can be adapted to enable discovering similar users while preserving their privacy. The key idea is to compute the user profile on the end-user device, apply LSH on the local profile, and use the LSH cluster identifier as the interest group identifier of a user. By...

When people expose their private life in online social networks, this doesn't mean that they do not care about their privacy, but they do lack tools to evaluate the risks and to protect their data. To address this issue, we have previously designed the FORPS system (Friends Oriented Reputation Privacy Score) that evaluates the dangerousness of peop...

With the boom of social media, it has become increasingly easier for ordinary people to not only post their own content but share other people's content on the Internet. In this paper, we conceptualize a growing problem of moving user data - once a user posts some content on the Internet, the data is largely out of her control, the content can be f...

In this paper, we present a privacy control mechanism called PDE (Privacy Data Envelope) allowing users to protect their privacy sensitive content travelling over social and communication networks. Our solution is based on privacy policies expressed by the user and associated with his content. This approach makes use of a decentralized architecture...

The Friends-Oriented Reputation Privacy Score (FORPS) system provides a smart and simple way to help end-users managing their privacy in a social network. It aims to prevent a non-desirable propagation of personal sensitive data. FORPS built privacy sensitivity profile by understanding what are the category of themes, the category of objects and th...

This demonstration illustrates the concept of Privacy Data Envelopes (PDE) where each piece of information identified by the user as privacy-sensitive is embodied (or "enveloped") into a data structure that in addition to the initial raw data carries privacy-related properties and policies. We show a scenario where such envelopes cross the boundari...

The WellCom platform enables the creation, delivery, and management of advanced personalized and interactive multimedia applications and services in a distributed home environment. End users obtain easy and seamless access to interactive and personalized television (TV) services as well as TV-related applications through their mobile terminals. The...

Most of the existing personalization systems such as content recommenders or targeted ads focus on individual users and ignore the social situation in which the services are consumed. However, many human activities are social and involve several individuals whose tastes and expectations must be taken into account by the system. When a group profile...

Today most of existing personalization systems (e.g. content recommenders, or targeted ad) focus on individual users and ignore the social situation in which the services are consumed. However, many human activities are social and involve several individuals whose tastes and expectations must be taken into account by the service providers. When a g...

Mastering knowledge of the user profile is one of the technical cornerstones for service providers who handle a large amount of end user service consumption data and are well positioned to dynamically infer user interest domains. This paper presents a holistic approach to service personalization by offering a means to gather a user's consumption da...

In this chapter, we present a recommending system that has been developed for filtering TV content provided to users on their mobile devices. This recommender is fully based on ontologies which are used to formalize both the user and her/his interests, and the audiovisual content. The developed ontologies allow matchmaking between user and content...

More and more systems allow user personalization and provide item recommendations, intended to fit individual user interests. In a traditional VoD system, for example, the recommendations are oriented towards a single user even though he is not watching the video alone. Hence, there is a need to have recommendations for a set of users, a group. Col...

Mastering knowledge of user profile is one of the technical cornerstones for service providers who handle a large amount of service consumption data and are well positioned to dynamically infer user interests. This paper presents a technology allowing to gather usage data from different multimedia services, create and track users profiles in real-t...

We present here a recommending system that has been developed for filtering TV content provided to mobile devices users. This recommender is fully based on ontologies, which are used to formalize both the user and her/his interests, and the audiovisual content. The developed ontologies allow matchmaking between user and content at different levels,...

Access to relevant information, adapted to user's needs, preferences and environment, is a challenge in many applications running in content delivery platforms, like IPTV, VoD and mobile Video. In order to provide users with personalized content, applications use various techniques such as content recommendation, content filtering, preference-drive...

As services become increasingly personalized, mastering knowledge of user profiles is becoming a key requirement for service providers who hold large amounts of end user service consumption data. This paper describes a profiling engine that automatically learns user profiles (user preferences, interest domains, and behaviors) by aggregating traces...

In this paper, we present a beginning work on a behavioural profiling approach within a multiservice environment where the service usage and content consumptions data are collected on different service delivery platforms and/or on the user terminals. Therefore, it is necessary to be able to aggregate these potentially heterogeneous data so that a f...

Multi-media services and other critical multi-site services (e.g. VPN) are becoming mainstream, and require a guaranteed Quality of Service (QoS). Services need to be established across several Autonomous Systems (ASes), often to connect end-users. Thus, provisioning and control of end-to-end QoS requirements arise as ones of the main challenges in...

Quality of Service (QoS) has been a major concern in the field of network management, even more so for emerging dynamic multimedia applications (Video on Demand, Telefony over IP etc.) that are becoming mainstream. This problem is particularly sensitive in the context of exchanges accross multiple independent and heterogeneous domains (X-domain), w...

This document presents a generic model capturing the essential structural and behavioral characteristics of network components in the light of fault management. The generic model is described by means of UML notations, and can be compiled to obtain rules for a Viterbi distributed diagnoser. This paper presents the results of the continued efforts o...

Distributed architectures for network management have been the subject of a large research eort, but distributed algorithms that im- plement the corresponding functions have been much less investigated. In this paper we describe novel algorithms for model-based distributed fault diagnosis.

This paper describes a formalism to model the behavior of telecommunications net-works when a fault occurs and how the effects are propagated across equipment. The objective of such a formalism, which is derived from UML diagram sequences, is to ease the construction and the update of a model corresponding to the supervised network; this model can...

The aim of this paper is to investigate new tracks to tackle the open issues of alarm correlation for complex networks (e.g. IP/SDH or IP/WDM). The suggested approach consist of transforming, by means of a neural network the equipment alarm messages into a signal, and then, to use signal-processing methods in order to extract the relevant informati...

This paper is dedicated to a mixed approach of two formalisms used in telecommunication networks and system monitoring. The first one is based on the chronicle recognition that detects relevant pieces of evolution, whereas the second one, is based on the Petri nets and is take into account the complete behavior of the system. We propose a cooperati...

We address the problem of alarm correlation in large distributed systems. The key idea is to make use of the concurrence of events in order to separate and simplify the state estimation in a faulty network. Petri nets and their causality semantics are used to model concurrency. Special partially stochastic Petri nets are developed, that establish s...

For pt.I see ibid., p.720-5 (1997). We present an original construction of stochastic Petri nets (PN) dedicated to large distributed discrete event systems. Its main characteristic is to provide statistically independent behaviors to concurrent (parallel) processes of the system. We end up with “hybrid” model where only some events are randomized,...

This paper presents a new use of safe Petri nets in the field of distributed discrete event systems, with application to telecommunication network management. This study has in its long range objectives to provide a generic supervisor, which can be easily distributed on a set of sensors. Petri nets are used to provide both a model and an algorithm...

: This report presents a new use of safe Petri nets in the field of distributed Discrete Event Dynamic Systems, with application to telecommunication network management. This study has in its long range objectives to provide a generic supervisor, which can be easily distributed on a set of sensors. Petri nets are used to provide both a model and an...

We have performed robust stability/performance μ-analysis for a beam figure control system to check if the pre-designed SVD-control met design specifications for any possible damping factor within a specified range. Such an elaborate technique was used due to failure of robustness analysis with unstructured uncertainty model. However, even the μ-an...

Policy-based management integrates two very different information models, an IP packet based model, and a system/device/network resource based model. The Internet standards or drafts incorporate the two models, however, the IP packet based model is much more used in today's policy-based management systems. We propose a more extensive usage of the s...