CalcGraph: taming the high costs of deep learning using models

in Software and Systems Modeling (2022)

Models based on differential programming, like deep neural networks, are well established in research and able to outperform manually coded counterparts in many applications. Today, there is a rising ... [more ▼]

Models based on differential programming, like deep neural networks, are well established in research and able to outperform manually coded counterparts in many applications. Today, there is a rising interest to introduce this flexible modeling to solve real-world problems. A major challenge when moving from research to application is the strict constraints on computational resources (memory and time). It is difficult to determine and contain the resource requirements of differential models, especially during the early training and hyperparameter exploration stages. In this article, we address this challenge by introducing CalcGraph, a model abstraction of differentiable programming layers. CalcGraph allows to model the computational resources that should be used and then CalcGraph’s model interpreter can automatically schedule the execution respecting the specifications made. We propose a novel way to efficiently switch models from storage to preallocated memory zones and vice versa to maximize the number of model executions given the available resources. We demonstrate the efficiency of our approach by showing that it consumes less resources than state-of-the-art frameworks like TensorFlow and PyTorch for single-model and multi-model execution. [less ▲]

Near Real-Time Electric Load Approximation in Low Voltage Cables of Smart Grids with Models@run.time

in 31st Annual ACM Symposium on Applied Computing (SAC'16) (2016, April)

Micro-generations and future grid usages, such as charging of electric cars, raises major challenges to monitor the electric load in low-voltage cables. Due to the highly interconnected nature, real-time ... [more ▼]

Micro-generations and future grid usages, such as charging of electric cars, raises major challenges to monitor the electric load in low-voltage cables. Due to the highly interconnected nature, real-time measurements are problematic, both economically and technically. This entails an overload risk in electricity networks when cables must be disconnected for maintenance reasons or are accidentally damaged. Therefore, it is of great interest for electricity grid providers to anticipate the load in networks and quicker detect failures. However, computing the electric load in cables requires computational intensive power flow calculations and live consumption measurements. Today’s view of the grid is usually based on on-field documentation of cables, fuses, and measurements by technicians and therefore often outdated. Thus, the electric load is usually only simulated in case of major topology variations. However, live measurements of smart meters provide new opportunities. In this paper we present a novel approach for a near real-time electric load approximation by deriving in live the current electric topology and cable loads from smart meter data. We leverage the models@run.time paradigm to combine live measurements with topology characteristics of the grid. Our approach enables to approximate the load in cables, not only for the current grid topology, but also to simulate topology changes for maintenance purposes. We showed that this allows a near real-time approximation while remaining very accurate (average deviation of 1.89% compared to offline power-flow calculation tools). Developed with a grid operator, this approach will be integrated in a monitoring and warning system and as an embeddable solution for on-field simulation. [less ▲]

Privacy Challenges in Ambient Intelligence Systems

in Journal of Ambient Intelligence and Smart Environments (2016)

Today, privacy is a key concept. It is also one which is rapidly evolving with technological advances, and there is no consensus on a single definition for it. In fact, the concept of privacy has been ... [more ▼]

Today, privacy is a key concept. It is also one which is rapidly evolving with technological advances, and there is no consensus on a single definition for it. In fact, the concept of privacy has been defined in many different ways, ranging from the “right to be left alone” to being a “commodity” that can be bought and sold. In the same time, powerful Ambient Intelligence (AmI) systems are being developed, that deploy context-aware, personalised, adaptive and anticipatory services. In such systems personal data is vastly collected, stored, and distributed, making privacy preservation a critical issue. The human- centred focus of AmI systems has prompted the introduction of new kinds of technologies, e.g. Privacy Enhancing Technologies (PET), and methodologies, e.g. Privacy by Design (PbD), whereby privacy concerns are included in the design of the system. One particular application field, where privacy preservation is of critical importance is Ambient Assisted Living (AAL). Emerging from the continuous increase of the ageing population, AAL focuses on intelligent systems of assistance for a better, healthier and safer life in their living environment. In this paper, we first build on our previous work, in which we introduced a new tripartite categorisation of privacy as a right, an enabler, and a commodity. Second, we highlight the specific privacy issues raised in AAL. Third, we review and discuss current approaches for privacy preservation. Finally, drawing on lessons learned from AAL, we provide insights on the challenges and opportunities that lie ahead. Part of our methodology is a statistical analysis performed on the IEEE publications database. We illustrate our work with AAL scenarios elaborated in cooperation with the city of Luxembourg. [less ▲]

Stream my Models: Reactive Peer-to-Peer Distributed Models@run.time

in Lethbridge, Timothy; Cabot, Jordi; Egyed, Alexander (Eds.) 2015 ACM/IEEE 18th International Conference on Model Driven Engineering Languages and Systems (MODELS) (2015, September)

The models@run.time paradigm promotes the use of models during the execution of cyber-physical systems to represent their context and to reason about their runtime behaviour. However, current modeling ... [more ▼]

The models@run.time paradigm promotes the use of models during the execution of cyber-physical systems to represent their context and to reason about their runtime behaviour. However, current modeling techniques do not allow to cope at the same time with the large-scale, distributed, and constantly changing nature of these systems. In this paper, we introduce a distributed models@run.time approach, combining ideas from reactive programming, peer-to-peer distribution, and large-scale models@run.time. We define distributed models as observable streams of chunks that are exchanged between nodes in a peer-to-peer manner. lazy loading strategy allows to transparently access the complete virtual model from every node, although chunks are actually distributed across nodes. Observers and automatic reloading of chunks enable a reactive programming style. We integrated our approach into the Kevoree Modeling Framework and demonstrate that it enables frequently changing, reactive distributed models that can scale to millions of elements and several thousand nodes. [less ▲]

Adaptive Blurring of Sensor Data to balance Privacy and Utility for Ubiquitous Services

in The 30th Annual ACM Symposium on Applied Computing (2015, April)

Given the trend towards mobile computing, the next generation of ubiquitous “smart” services will have to continuously analyze surrounding sensor data. More than ever, such services will rely on data ... [more ▼]

Given the trend towards mobile computing, the next generation of ubiquitous “smart” services will have to continuously analyze surrounding sensor data. More than ever, such services will rely on data potentially related to personal activities to perform their tasks, e.g. to predict urban traffic or local weather conditions. However, revealing personal data inevitably entails privacy risks, especially when data is shared with high precision and frequency. For example, by analyzing the precise electric consumption data, it can be inferred if a person is currently at home, however this can empower new services such as a smart heating system. Access control (forbid or grant access) or anonymization techniques are not able to deal with such trade-off because whether they completely prohibit access to data or lose source traceability. Blurring techniques, by tuning data quality, offer a wide range of trade-offs between privacy and utility for services. However, the amount of ubiquitous services and their data quality requirements lead to an explosion of possible configurations of blurring algorithms. To manage this complexity, in this paper we propose a platform that automatically adapts (at runtime) blurring components between data owners and data consumers (services). The platform searches the optimal trade-off between service utility and privacy risks using multi-objective evolutionary algorithms to adapt the underlying communication platform. We evaluate our approach on a sensor network gateway and show its suitability in terms of i) effectiveness to find an appropriate solution, ii) efficiency and scalability. [less ▲]

Conviviality-driven access control policy

in Requirements Engineering (2015), 20(4), 363-382

Nowadays many organizations experience security incidents due to unauthorized access to information. To reduce the risk of such incidents, security policies are often employed to regulate access to ... [more ▼]

Nowadays many organizations experience security incidents due to unauthorized access to information. To reduce the risk of such incidents, security policies are often employed to regulate access to information. Such policies, however, are often too restrictive, and users do not have the rights necessary to perform assigned duties. As a consequence, access control mechanisms are perceived by users as a barrier and thus bypassed, making the system insecure. In this paper, we draw a bridge between the social concept of conviviality and access control. Conviviality has been introduced as a social science concept for ambient intelligence and multi-agent systems to highlight soft qualitative requirements like user-friendliness of systems. To bridge the gap between conviviality and security, we propose a methodological framework for updating and adapting access control policies based on conviviality recommendations. Our methodology integrates and extends existing techniques to assist system designers in the derivation of access control policies from socio-technical requirements of the system, while taking into account the conviviality of the system. We illustrate our framework using the Ambient Assisted Living use case from the HotCity of Luxembourg. © 2014, Springer-Verlag London. [less ▲]

A Rule-based Contextual Reasoning Platform for Ambient Intelligence environments

in Theory, Practice, and Applications of Rules on the Web (2013, July 01)

The special characteristics and requirements of intelligent environments impose several challenges to the reasoning processes of Ambient Intelligence systems. Such systems must enable heterogeneous ... [more ▼]

The special characteristics and requirements of intelligent environments impose several challenges to the reasoning processes of Ambient Intelligence systems. Such systems must enable heterogeneous entities operating in open and dynamic environments to collectively rea- son with imperfect context information. Previously we introduced Con- textual Defeasible Logic (CDL) as a contextual reasoning model that addresses most of these challenges using the concepts of context, map- pings and contextual preferences. In this paper, we present a platform integrating CDL with Kevoree, a component-based software framework for Dynamically Adaptive Systems. We explain how the capabilities of Kevoree are exploited to overcome several technical issues, such as com- munication, information exchange and detection, and explain how the reasoning methods may be further extended. We illustrate our approach with a running example from Ambient Assisted Living. [less ▲]

Introducing Conviviality as a New Paradigm for Interactions among IT Objects

in Proceedings of the Workshop on AI Problems and Approaches for Intelligent Environments (2012, August), 907

The Internet of Things allows people and objects to seamlessly interact, crossing the bridge between real and virtual worlds. Newly created spaces are heterogeneous; social relations naturally extend to ... [more ▼]

The Internet of Things allows people and objects to seamlessly interact, crossing the bridge between real and virtual worlds. Newly created spaces are heterogeneous; social relations naturally extend to smart objects. Conviviality has recently been introduced as a social science concept for ambient intelligent systems to highlight soft qualitative requirements like user friendliness of systems. Roughly, more opportunities to work with other people increase the conviviality. In this paper, we first propose the conviviality concept as a new interaction paradigm for social exchanges between humans and Information Technology (IT) objects, and extend it to IT objects among themselves. Second, we introduce a hierarchy for IT objects social interactions, from low-level one-way interactions to high-level complex interactions. Then, we propose a mapping of our hierarchy levels into dependence networks-based conviviality classes. In particular, low levels without cooperation among objects are mapped to lower conviviality classes, and high levels with complex cooperative IT objects are mapped to higher conviviality classes. Finally, we introduce new conviviality measures for the Internet of Things, and an iterative process to facilitate cooperation among IT objects, thereby the conviviality of the system. We use a smart home as a running example. [less ▲]