Research Interests

My research lies in the areas of mobile computing, mobile sensing, machine learning, resource-efficient computing, and the investigation of the interraction between the computing and the society. In particular I am interested in:


AMC - "Bringing Resource Efficiency to Smartphones with Approximate Computing" (ARRS, PI, 2020-2021 200,000 EUR)

RICERCANDO - Rapid Interpretation and Cross-Experiment Root-Cause Analysis in Network Data with Orange (H2020, co-PI, 2016-2018 150,000 EUR)

AKOS - Feasability study: Crowdsourced measurements mining for Net neutrality violation detection (Industrial, PI, 2018, 20,000 EUR)

Mobile computing for efficient information dissemination in emergency situations (US-Slovenia bilateral project, PI, 2020-2021, 3,000 EUR)

Current Projects

Bringing Resource Efficiency to Smartphones with Approximate Computing (funded by ARRS)

The breakdown of Moore's law and Dennard scaling critically threatens the future growth of mobile computing. It indicates that further packing of computing resources cannot be sustained, should we wish to preserve the portability and energy frugality of our devices. Yet, as mobile computing affirms its central role in our everyday life, we entrust an ever increasing number of complex computational tasks to it.

In this project we aim to set foundations for a new mobile computing paradigm termed - approximate mobile computing (AMC). ACM is based on our insight that, depending on the context, a result need not be perfectly accurate in order to fulfil a user’s needs: when playing a video game outdoors a gamer tolerates imperfect 3D rendering, just as a person looking for recommendations does not mind getting a slightly shuffled order of nearby restaurants. Bringing ACM to realisation requires that the underlying foundation - the technology for enabling accuracy-adaptable computation - is implemented.

In the project we focus on 1) identification of approximate computing techniques (ACTs) that are suitable for mobile devices, specifically, smartphones; 2) implementation of selected ACTs on low level of the mobile computing stack, such as compiler and supporting library level, and 3) exposing control knobs for the dynamic tuning of ACTs, so that the result accuracy can be traded for resource usage at runtime. Results of our work will provide a basis for future experimentation with AMC, in particular with respect to its context-aware adaptation and further experimentation with resource savings that AMC brings.

Involved researchers:
  • Dr. Octavian Machidon
  • Dr. Davor Sluga
  • Dr. Alina Machidon
  • Ivan Majhen
  • Amol Jawale
  • Dr. Veljko Pejovic

Security in the IoT Domain (in collaboration with JRC)

The Internet of Things (IoT) encompasses devices capable of sensing their surroundings, actioning upon machine learning-based reasoning, and wirelessly communicating to each other. Such devices bring "smartness" to the setting in which they are deployed, thus leading to smart homes, offices, factories, cars, and others. However, a complex mix of interacting devices, the sensors they host, and the authentication/authorisation policies they conform to results in a series of potential security and privacy threats that are difficult to tackle with conventional security approaches. In particular, a need arises to:
  • Consider security beyond a single device.
  • Examine different aspects of security, from technical implementations, to human behaviour aspects.
  • Develop comprehensive secure practices and easy-to-use policies for IoT environments.
The European Commission’s Joint Research Center (JRC) at Ispra, Italy and the Faculty of Computer and Information Science (FRI), University of Ljubljana, Slovenia jointly work on tackling the above challenges. Our research concentrates on machine learning modelling and analysis of sensory data and wireless network traffic, with the goal of deducing activities and interactions in the smart environment, profiling users, inferring private and sensitive information. Further, we develop possible countermeasures and mitigation techniques to the identified security threats. Finally, we aim to devise an automated system for security policy adaptation governing the behavior of the smart devices.

Involved researchers (FRI side):
  • Daniel Pellarini
  • Andraz Krasovec
  • Dr. Veljko Pejovic

Mobile User Attention Management

Attention is the most precious of the resources we have. From communication services (chat, SMS), to online social networks (Facebook, Twitter), and warning systems (e.g. "battery low"), different actors compete for our attention. In our research we aim to understand mechanisms that guide user attention and to develop technical solutions for attention management.
Our goal is to advance human-computer interaction to the point defined by Mark Weiser's 1991 quote:
machines that fit the human environment instead of forcing humans to enter theirs will make using computing as refreshing as taking a walk in the woods
From the practical point of view, attention management systems built towards the above vision will enable us to:
  • Reduce user frustration and application churn rate.
  • Increase user engagement with the delivered content.
  • Increase compliance with suggestions in mHealth and behaviour change intervention applications
  • Improve safety in ubiquitous computing environments (e.g. by not interrupting while a user is driving).
  • Reduce unwanted data traffic.
  • Open doors for anticipatory computing, where future activities could be intelligently steered through predictive attention management.
In our research, through mobile sensing, we have thoroughly investigated factors that impact a user's interruptibility (see our UbiComp'14, UbiComp'15, and CHI'16 papers, for example). Further, we built an Android library InterruptMe that uses sensors on a smartphone to recognize the user's context and infer if a user is interruptible or not. InterruptMe is available as open source software.

In our latest line of research we are investigating the ability of mobile and wireless sensors to recognize a user's cognitive load, as the information on cognitive load can greatly enhance attention management systems (see our UbiComp'18 survey for details). Our preliminary analysis with wearable and wireless methods for cognitive load inference shows the potential for unobtrusive recognition of a user's task engagement. We are currently working on in-depth inference of the actual level of cognitive load, as well as, on inferring cognitive load in real-world situations for automotive industry.
Activity - Interruptibility
Our Wi-Mind systems for wireless cognitive load inference.

Past Projects

Mobile broadband measurement analysis (funded by EU Horizon2020 and AKOS)

Mobile broadband (MBB) networks have revolutionised the way we communicate, yet our understanding and means of troubleshooting such complex systems remain modest. In our research we develop RICERCANDO (Rapid Interpretation and Cross-Experiment Root-Cause Analysis in Network Data with Orange), a tool for integrative exploration, visualization and interpretation of MBB data and meta-data across multiple experiments. RICERCANDO supports human experts in the process of detecting and understanding the root-cause of the network problems and performance degradation.

RICERCANDO was developed by an interdisciplinary team that includes data mining (prof. Blaz Zupan, Jernej Kernc) and networking experts (prof. Ricciato, prof. Pejovic, Ivan Majhen, and Dr. Miha Janez).

All tools developed as a part of RICERCANDO project are free and open-source, and can be found here:
We also invite you to find out more in our RICERCANDO paper

Finally, we used our tools, including RICERCANDO, to uncover Net neutrality violations. We analyzed almost two years worth of distributed crowdsourced measurement data from Agency for communication networks and services of the Republic of Slovenia (AKOS) and found examples of Net neutrality violation practices.