Research

(under construction - probably forever)

Processing & Learning on Networks

My group focuses principally on the fundamental study of networked- and topologically-based data processing and learning techniques. The meta-level research question we try to answer is how to analyze and process data with an irregular structure to learn meaningful representations to better understand their behavior. For this, we use a mix of techniques from signal processing, machine learning, mathematical modeling, and network theory. The main research thrusts include:

Graph signal processing: We investigate signal processing techniques for graph data so as to capture their structure analogously as we do for time series and images.
Graph Neural Networks: We study (deep) representation learning techniques to extract and process representations from graph data.
Topological and Higher-Order Networks: We study machine learning techniques for data whose structure can be represented by topological objects or higher-order networks such as simplicial/cell complexes and hypergraphs.
Data-driven Applications on Networks: Many of the above fundamental research is inspired by different network-based applications where we show impact either via a proof of concept or investigate them in detail.

Most of the research considers the temporal component in the data and/or topology as quite often we have that the data, their interpretation, and hidden pattern vary over time. Additionally, we integrate different forms of domain knowledge as inductive biases so as to have a more tractable world model to analyze. I find the latter particularly useful in data-driven solutions so as to develop more effective algorithms as well as to characterize the role the data topology has in their inner-working mechanisms. Below is more detail on the above research thrusts and respective publications.

Graph Neural Networks

Graph neural networks (GNNs) are machine learning techniques for graph-structured data that aim to learn hierarchical representation following the neural network principle. Due to the versatility of the graphs to represent multiple technological and natural complex systems, as well as the challenge of processing the information associated with them, GNNs provide a powerful data-driven solution to different tasks. They have for instance been used to detect fake news spreading on Twitter, provide recommendations on Pinterest and Linkedin, help discover new drugs from molecular graphs, as well as estimate time of arrival on Google Maps. Likewise, they have shown unprecedented performance in physical networks such as water, transportation, and power networks. In this research thrust, we focus on how: (i-Architecture Analysis) to develop and study GNN architectures; (ii- Uncertain Environment) to characterize GNNs' behavior when the underlying topology or data are uncertain due to estimation errors or adversarial attacks;(iii-Dynamic Graphs) to incorporate time-varying graphs into GNNs; (iv-Stability Analysis) to understand the role of the different GNN components in uncertain environment. More specifically:

Architecture Analysis: We focus on developing, characterizing, and unifying GNN architectures to better learn representations from graph data. We focus on a graph-based filtering perspective and an additional inductive bias perspective to provide a complementary perspective to the message passing one. This perspective allows unifying many of the different state-of-the-art alternatives, understanding their advantages and limitations, as well as providing principled guidelines to develop others based on desired properties for the task at hand.
Uncertain Environment: In real applications graphs and data are rarely ideal and prone to outliers, estimation errors, or adversarial attacks. This makes the setting where GNNs operate rather uncertain and we are motivated to understand how this uncertainty affects the learned representations as well as how to develop more robust architectures.
Dynamic Setting: Inspired by the fact that graphs and data change over time, we study the impact of this dynamicity on GNNs to adapt learning to the dynamic topology and investigate the impact of the temporal variation on the learned representations.
Stability Analysis: Both the uncertain environment and dynamic graph changes are needed as the first step in characterizing the behavior of GNNs when blind to this setting. This falls under the umbrella of stability/transferability analysis of GNNs to understand our handle to improve their inner working mechanisms.

Selected Publications

Architecture Analysis

E. Isufi, F. Gama, D. I. Shuman and S. Segarra, Graph Filters for Signal Processing and Machine Learning on Graphs, in the IEEE Transactions on Signal Processing, Dec. 2023 [PDF] (overview article)
E. Isufi, F. Gama and A. Ribeiro, EdgeNets: Edge Varying Graph Neural Networks, in the IEEE Transactions on Pattern Analysis and Machine Intelligence, Sep. 2021. [PDF]
F. Gama, E. Isufi, G. Leus and A. Ribeiro, Graphs, Convolutions and Neural Networks: From Graph Filters to Graph Neural Networks, IEEE Signal Processing Magazine; Special Issue on Graph Signal Processing: Foundations and Emerging Directions. 2020. [PDF]
Z. Gao, A. Prorok and E. Isufi, On the Trade-Off between Stability and Representional Capacity in Graph Neural Networks, submitted to IEEE Transactions on Signal Processing, Dec. 2023 [PDF]
L. Ruiz, F. Gama, A. Ribeiro and E. Isufi, Nonlinear State-Space Generalizations of Graph Convolutional Neural Networks, in Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Toronto, Ontario, Canada, Jun. 2021. [PDF]

Uncertain Environment

Z. Gao and E. Isufi, Learning Stochastic Graph Neural Networks with Constrained Variance, in the IEEE Transactions on Signal Processing, Jan. 2023. [PDF]
Z. Gao, E. Isufi and A. Ribeiro, Stochastic Graph Neural Networks, in the IEEE Transactions on Signal Processing, 2021. [PDF]
A. Möllers, A. Immer, E. Isufi and V. Fortuin, Uncertainty in Graph Contrastive Learning with Bayesian Neural Networks, 5th Symposium on Advances in Approximate Bayesian Inference, collocated with ICML, July 2023.

Dynamic Setting

M. Sabbaqi and E. Isufi, Graph-Time Convolutional Neural Networks: Architecture and Theoretical Analysis, in the IEEE Transactions on Pattern Analysis and Machine Intelligence, Aug. 2023. [PDF]
A. Cavallo, M. Sabbaqi, and E. Isufi, Spatiotemporal Covariance Neural Networks, Joint European Conference on Machine Learning and Knowledge Discovery in Databass (ECML-PKDD), Vilnius, Lithuania, Sep. 2024. [PDF]
M. Sabbaqi, R. Taormina, A. Hanjalic and E. Isufi, Graph-Time Convolutional Autoencoders, Proceedings of the First Learning on Graphs Conference (Log 2022), PMLR 198, Virtual, Dec. 2022
E. Isufi and G. Mazzola, Graph-Time Convolutional Neural Networks, in Proceedings of the IEEE Data Science and Learning Workshop, Toronto, Ontario, Canada, Jun. 2021. [Code] [PDF] (audience choice award)

Stability Analysis

Z. Gao, A. Prorok and E. Isufi, On the Trade-Off between Stability and Representional Capacity in Graph Neural Networks, submitted to IEEE Transactions on Signal Processing, Dec. 2023 [PDF]
Z. Gao, E. Isufi and A. Ribeiro, Stability of Graph Convolutional Neural Networks to Stochastic Perturbations, Elsevier Signal Processing, Special Issue on Processing and Learning over Networks, Jun. 2021. [PDF]
Z. Gao and E. Isufi, Learning Stable Graph Neural Networks via Spectral Regularization, IEEE Asilomar Conference on Signals, Systems and Computations, Pacific Grove, USA, Nov. 2022. [PDF] (invited paper)
R. Levie, E. Isufi and G. Kutyniok, On the Transferability of Spectral Graph Filters, in Proceedings of the 2019 IEEE International Conference on Sampling Theory and Applications (SAMPTA), Bordeaux, France, Jul. 2019. [PDF]

Graph Signal Processing

Graph signal processing (GSP) aims to develop techniques for processing data over networks by generalizing classical concepts commonly used for time series and images such as the Fourier transform, filtering, sampling, and interpolation, among others. For example, in a social network users represent the nodes of the graphs, their connections such as friendships or followers the edges, and the data signal may be user preferences for a particular topic. We then want to analyze how: (i-Fourier) this preference varies across the network; (ii-Filter) removes artifacts and noise in the data; (iii-Sampling and interpolation) collect the information from a few nodes (e.g., on which users to run a survey) and interpolate the missing values on the other nodes. While GSP covers a broad spectrum of fundamental research aspects, I have been mainly trying to answer the following research questions:

Graph Filters: How can we develop more effective filtering techniques for graph data? Being filters one of the cornerstone techniques in signal processing and deep learning (e.g., CNNs rely on 2D convolutional filtering) I have been focusing on understanding the data behavior w.r.t. the underlying graph as well as on analyzing filter design techniques.
Graph-Time Signal Processing: How can we process time-varying data on graphs? Even on a static graph such as an infrastructure (water, power, road) network the data on the nodes change over time. We would like then to take into account the joint spatiotemporal dependency on the data to process the information they carry.
Dynamic Graphs: How can we process (time-varying) data over dynamic graphs? And how do the dynamics in the topology impact our data processing techniques? These questions target the important yet challenging aspect of network data processing, which is to handle dynamics in the topology.
Graph Learning: How can we learn the underlying data topology from data? Oftentimes we either observe a part of the graph or do not observe at all. Then, we would like to infer this graph structure from the measurements to provide insights for the downstream task.

Selected Publications

Graph Filters

1. E. Isufi, F. Gama, D. I. Shuman and S. Segarra, Graph Filters for Signal Processing and Machine Learning on Graphs, in the IEEE Transactions on Signal Processing, Dec. 2023 [PDF] (overview article)
2. F. Gama, E. Isufi, G. Leus and A. Ribeiro, Graphs, Convolutions and Neural Networks: From Graph Filters to Graph Neural Networks, IEEE Signal Processing Magazine; Special Issue on Graph Signal Processing: Foundations and Emerging Directions. 2020. [PDF] (magazine article)
3. L. Ben Saad, B. Beferull-Lozano and E. Isufi, Quantization Analysis and Robust Design for Distributed Graph Filters, in the IEEE Transactions on Signal Processing, Dec. 2021. [PDF]
4. M. Coutino, E. Isufi and G. Leus, Advances in Distributed Graph Filtering, IEEE Transactions on Signal Processing, vol. 67 (9), pp. 2320 - 2333, 2019. [PDF]
5. J. Liu, E. Isufi and G. Leus, Filter Design for Autoregressive Moving Average Graph Filters, IEEE Transactions on Signal and Information Processing over Networks, vol. 5 (1), pages 47-60, 2019. [PDF]
6. E. Isufi, A. Loukas, A. Simonetto and G. Leus, Autoregressive Moving Average Graph Filtering, IEEE Transactions on Signal Processing, vol.67 (2), pp. 274-288, 2017. [PDF]
7. R. Levie, E. Isufi and G. Kutyniok, On the Transferability of Spectral Graph Filters, in Proceedings of the 2019 IEEE International Conference on Sampling Theory and Applications (SAMPTA), Bordeaux, France, Jul. 2019. [PDF]

Graph-Time Signal Processing

E. Isufi, P. Banelli, P. Di Lorenzo and G. Leus, Observing and Tracking Bandlimited Graph Processes, Elsevier Signal Processing, 2020. [PDF]
E. Isufi, A. Loukas, N. Perraudin and G. Leus, Forecasting Time Series with VARMA Recursions on Graphs, IEEE Transactions on Signal Processing, 2019. [PDF]
P. Di Lorenzo, P. Banelli, E. Isufi, S. Barbarossa and G. Leus, Adaptive Graph Signal Processing: Algorithms and Optimal Sampling Strategies, IEEE Transactions on Signal Processing, vol. 66 (13), pp. 3584 - 3598, 2018. [PDF]
A. Natali, E. Isufi and G. Leus, Forecasting Multi-dimensional Graph Processes over Graphs, in Proceedings of the 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Barcelona, Spain, May. 2020. [PDF]
E. Isufi, P. Banelli and G. Leus, 2-Dimensional Finite Impulse Response Graph-Temporal Filters, in Proceedings of the IEEE Global Conference on Signal and Information Processing (GlobalSIP), Washington DC, USA, Dec. 2016. [PDF]

Dynamic Graphs

B. Das and E. Isufi, Online Graph Filtering Over Expanding Graphs, in the IEEE Transactions on Signal Processing, Sep. 2024 [PDF]
F. Gama, E. Isufi, A. Ribeiro and G. Leus, Controllability of Bandlimited Graph Processes Over Random Time-Varying Graphs, IEEE Transactions on Signal Processing, 2019.
E. Isufi, A. S. U. Mahabir and G. Leus, Blind Graph Topology Change Detection, IEEE Signal Processing Letters, vol. 15 (5), pp. 655 - 659, 2018. [PDF]
E. Isufi, A. Loukas, A. Simonetto and G. Leus, Filtering Random Graph Processes Over Random Time-Varying Graphs, IEEE Transactions on Signal Processing, vol.65 (16), pp. 4406-4421, 2017. [PDF]
M. Sabbaqi and E. Isufi, Inferring Time Varying Signals over Uncertain Graphs, IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), South Korea, Apr. 2024.
B. Das and E. Isufi, Tensor Graph Decomposition for Temporal Networks, IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), South Korea, Apr. 2024.
B. Das and E.Isufi, Online Vector Autoregressive Models over Expanding Graphs, IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), Greece, June. 2023. (invited paper)
B. Das and E. Isufi, Graph Filtering Over Expanding Graphs, in the IEEE Data Science and Learning Workshop (DSLW), Singapore, May. 2022. [PDF]

Graph Learning

R. Money, J. Krishnan, B. Beferull-Lozano and E. Isufi, Scalable and Privacy-aware Online Learning of Nonlinear Structural Equation Models, in the IEEE Open Journal on Signal Processing, Jan. 2023. [PDF]
B. Das, A. Hanjalic and E. Isufi, Task-Aware Connectivity Learning for Incoming Nodes on Growing Graphs, in the IEEE Transactions on Signal and Information Processing over Networks, Aug. 2022. [PDF]
A. Natali, E. Isufi, M. Coutino and G. Leus, Learning Time-Varying Graphs from Online Data, in the IEEE Open Journal on Signal Processing, May. 2022 [PDF]. (top 25 donwloaded articles)
M. Coutino, E. Isufi, T. Maehara and G. Leus, State-Space Network Topology Identification from Partial Observations, IEEE Transactions on Signal and Information Processing over Networks; Special Issue on Network Topology Identification, 2020. [PDF]

Higher-Order Networks

Higher-order networks (HONS) such as hypergraphs, simplicial, and cell complexes, can represent multiway relations in the data, which graphs fail to represent. They are particularly relevant to model flows on networks which can be seen as data associated with the edges (pairs of nodes). One particularity in this setting is that we may have present interdependent data on multiple levels such as data associated with nodes, edges, and triples of nodes. For example, in a water network, the pressure can be seen as node data and the flows as edge data, which intrinsically influence each other. Our main goal here is to understand this interdependency to develop data processing techniques that leverage HONS as an inductive bias. The two broader research thrusts are:

HONS Signal Processing: We focus on analyzing and processing HONS data from a signal processing perspective by relying on the convolution principle and the spectral representation of these data via the Hodge theory. In particular, we want to understand how can we leverage the interdependency between the different levels to develop signal-processing algorithms in a principled way and to learn the underlying topology from data.
HONS Representation Learning: The core idea of this research thrust is to use HONS as inductive biases to learn joint representations from the multiway data. Relying on the Hodge theory, we aim to build learning techniques that on the one hand bridge with the HONS principles in network science and signal processing and on the other hand offer a degree of mathematical tractability.

Selected Publications

HONS Signal Processing

R. Money, J. Krishnan, B. Beferull-Lozano and E. Isufi, Online Missing Data Imputation of Edge Flows, in the IEEE Signal Processing Letters, Nov. 2022. [PDF]
M. Yang, E. Isufi, M. T. Schaub and G. Leus, Simplicial Convolutional Filters, in the IEEE Transactions on Signal Processing, Sep. 2022. [PDF]
R. Money, J. Krishnan, B. Beferull-Lozano and E. Isufi, Evolution Backcasting of Edge Flows from Partial Observations Using Simplicial Vector Autoregressive Models, IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), South Korea, Apr. 2024.
A. Buciulea, E. Isufi, G. Leus and A. G. Marques, Learning the Topology of a Simplicial Complex Using Simplicial Signals: A Greedy Approach, IEEE Sensor Array and Multichannel Signal Processing Workshop 2024 (SAM), United States, Jul. 2024.
A. Buciulea, E. Isufi, G. Leus and A. G. Marques, Learning Graphs and Simplicial Complexes from Data, IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), South Korea, Apr. 2024.
J. Krishnan, R. Money, B. Beferull-Lozano and E. Isufi, Simplicial Vector Autoregressive Model for Streaming Edge Flows, IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), Greece, June. 2023. (top 3% recognition award)
M. Yang and E. Isufi, Simplicial Trend Filtering, IEEE Asilomar Conference on Signals, Systems and Computations, Pacific Grove, USA, Nov. 2022. (invited paper)
E. Isufi and M. Yang, Convolutional Filters for Simplicial Complexes, in the IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), Singapore, May. 2022. [PDF]
G. Leus, M. Yang, M. Coutino and E. Isufi, Topological Volterra Filters, in Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Toronto, Ontario, Canada, Jun. 2021. [PDF]

HONS Representation Learning

M. Yang, V. Borovitskiy, and E. Isufi, Hodge-Compositional Edge Gaussian Processes, 27th International Conference on Artificial Intelligence and Statistics (AISTATS), Valenca, Spain, May 2024 [PDF]
M. Yang and E. Isufi, Convolutional Learning on Simplicial Complexes, Jan. 2023 [PDF].
C. Liu, G. Leus and E. Isufi, Unrolling of Simplicial ElasticNet for Edge Flow Signal Reconstruction, in the IEEE Open Journal on Signal Processing, Dec. 2023. [PDF]
A. Möllers, A. Immer, V. Fortuin and E. Isufi, Hodge-Aware Contrastive Learning, IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), South Korea, Apr. 2024. (invited paper)
M. Yang, E. Isufi and G. Leus, Simplicial Convolutional Neural Networks, in the IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), Singapore, May. 2022. [PDF] (invited paper)

Applications

Networks and their higher-order generalization are quite versatile tools for modeling complex systems and irregular data structures. While in most cases they can be seen as one of the many tools to represent the data structure, with respective advantages and limitations, we find them particularly relevant in situations where the data generation process is intrinsically linked to the network structure. Examples of the latter are infrastructure networks such as water distribution networks, mobility networks, or power networks. In another setting, we have evidence that graph-based abstraction of the data structure often carries specific meaning for the application at hand such as in flood modeling and recommender systems. The fundamental research we conduct is often inspired by challenges in these applications. In most cases, we provide a proof-of-concept in these domains by comparing them with baseline techniques but in other cases, we deeply investigate their impact with domain-specific collaborators. More specifically, we look at the following application domain challenges.

Infrastructure Networks: One key challenge in infrastructure networks is state estimation from a few noisy data. Within Aidrolab we look at this challenge for water networks but we have also investigated it for power systems and are looking at transportation networks. Typically this is done with physic/model-based numerical solutions, which are either slow or inaccurate. Intersecting graph-based models with physical information provides a powerful alternative in this setting to exploit the advantages of both domains. We look at this intersection to improve the accuracy in state estimation, forecast the state evolution, as well as speed up the computation time.
Flood Modeling: Modeling the evolution of a flood is of critical importance in nowadays where drastic weather changes are more frequent. In most cases, this task is approached from a computational fluid dynamic perspective, and numerical solutions are called into action. While these solutions are quite accurate, they are extremely slow and are not transferable (i.e., for every new boundary condition or setup they need to be re-run). Physic-inspired graph-based machine learning can be a viable alternative in this setting. We look particularly at how to leverage the physical information within GNNs as an inductive bias to learn the underlying model governing the flood evolution and, at the same time, reduce the computational by orders of magnitude.

Recommender Systems: RecSys is one of the most fertile domains for graph-based data processing. Here user-item interactions can be naturally represented via graphs and group/bulk recommendations call for higher-order representations. Graphs have resulted powerfully in this setting as they provide some abstraction to the task to overcome many of its challenges including the extremely high data sparsity, user biases, multi-criteria recommendations, and the cold-start problem. We draw quite some inspiration from RecSys challenges to study more fundamental graph-based data processing techniques but have also investigated a few of them in more detail.
Sensor Networks: A lot of the earlier research in the GSP thrust has taken inspiration from sensor networks, where the connectivity network acts also as a data exchange platform in distributed processing. We have analyzed and developed GSP algorithms tailored to sensor networks that overcome many of their challenges including broken communication links, limited communication resources, and anomalies in the data.

Selected Publications

Infrastructure Networks

B. Habib, E. Isufi, W. van Breda, A. Jongepier and J. L. Cremer, Deep Statistical Solver for Distribution System State Estimation, in the IEEE Transactions on Power Systems, June. 2023 [PDF]
B. Kerimov, R. Bentivoglio, J. A. G. Diaz, E. Isufi, F. Tscheikner-Grati, D. B. Steffelbauer, R. Taormina, Assessing the Performance and Transferability of Graph Neural Network Metamodels for Water Distribution Systems, Journal of Hydroinformatics, Sep. 2023. [PDF]
A. S. Roca, A. G. Díaz, E. Isufi and R. Taormina, EPANET Metamodels with Deep Unrolling of the Global Gradient Algorithm, WSDA / CCWI Joint Conference, 2023.
A. Garzon, R. Bentivoglio, E. Isufi, Z. Kapelan and R. Taormina, Modeling Water Distribution Systems with Graph Neural Networks, European Geoscience Union (EGU) General Assembly, 2021.
R. Taormina and E. Isufi, Geometric Deep Learning for Modeling, Prediction and Forecasting in Urban Water Systems, Advancing Earth and Space Science (AGU) Fall Meeting, Dec. 2020. (invited paper)
[Msc thesis] Albert Solà Roca (2023) GGANET: Algorithm Unrolling for Water Distribution Networks Metamodeling.

Proof of concept -- foundamental research inspired by application challenges and corroborated in a baseline setup

C. Liu, G. Leus and E. Isufi, Unrolling of Simplicial ElasticNet for Edge Flow Signal Reconstruction, in the IEEE Open Journal on Signal Processing, Dec. 2023. [PDF]
R. Money, J. Krishnan, B. Beferull-Lozano and E. Isufi, Online Missing Data Imputation of Edge Flows, in the IEEE Signal Processing Letters, Nov. 2022. [PDF]
M. Sabbaqi and E. Isufi, Graph-Time Convolutional Neural Networks: Architecture and Theoretical Analysis, in the IEEE Transactions on Pattern Analysis and Machine Intelligence, Aug. 2023. [PDF]
R. Money, J. Krishnan, B. Beferull-Lozano and E. Isufi, Evolution Backcasting of Edge Flows from Partial Observations Using Simplicial Vector Autoregressive Models, IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), South Korea, Apr. 2024.
J. Krishnan, R. Money, B. Beferull-Lozano and E. Isufi, Simplicial Vector Autoregressive Model for Streaming Edge Flows, IEEE International Conference on Acoustic, Speech and Signal Processing, (ICASSP), Greece, June. 2023. (top 3% recognition award)
M. Sabbaqi, R. Taormina, A. Hanjalic and E. Isufi, Graph-Time Convolutional Autoencoders, Proceedings of the First Learning on Graphs Conference (Log 2022), PMLR 198, Virtual, Dec. 2022

Flood Modeling

R. Bentivoglio, E. Isufi, S. N. Jonkman and R. Taormina, Rapid Spatio-Temporal Flood Modeling via Hydraulics-Based Graph Neural Networks, in the Hydrology and Earth System Sciences. Nov. 2023 [PDF] [Video]
R. Bentivoglio, E. Isufi, S. N. Jonkman and R. Taormina, Deep Learning Methods for Flood Mapping: A Review of Existing Applications and Future Research Directions, in theHydrology and Earth System Sciences. Jul. 2022 [PDF].
R. Bentivoglio, E. Isufi, S. N. Jonkman and R. Taormina, Multi-scale Graph Neural Networks for Flood Modeling, Natural Hazards and Earth System Sciences (NHESS). Aug. 2024. [PDF]

Recommender Systems

E. Isufi, M. Pocchiari and A. Hanjalic, Accuracy-Diversity Trade-off in Recommender Systems via Graph Convolutions, Elsevier Information Processing and Management; Special Issue on Advances in Graph Representation Learning for Large-scale Multimedia Analytics, Mar. 2021. [PDF]
R. Kalisvaart, M. Mansoury, A. Hanjalic and E. Isufi, Towards Carbon Footprint-Aware Recommender Systems for Greener Item Recommendation, submitted to ACM Transactions on Recommender Systems, Mar. 2025. [PDF]
[MSc thesis] Simon Dahrs (2022) Pure Cold Start Recommendation by Learning on Stochastically Expanded Graphs

Proof of concept -- foundamental research inspired by application challenges and corroborated in a baseline setup

Z. Gao and E. Isufi, Learning Stochastic Graph Neural Networks with Constrained Variance, in the IEEE Transactions on Signal Processing, Jan. 2023. [PDF]
B. Das, A. Hanjalic and E. Isufi, Task-Aware Connectivity Learning for Incoming Nodes on Growing Graphs, in the IEEE Transactions on Signal and Information Processing over Networks, Aug. 2022. [PDF]
B. Das and E. Isufi, Online Filtering over Expanding Graphs, IEEE Asilomar Conference on Signals, Systems and Computations, Pacific Grove, USA, Nov. 2022. (finalist best student paper award)
B. Das and E. Isufi, Graph Filtering Over Expanding Graphs, in the IEEE Data Science and Learning Workshop (DSLW), Singapore, May. 2022. [PDF] (best student paper award)

Sensor Networks

Proof of concept -- foundamental research inspired by application challenges and corroborated in a baseline setup

Z. Gao and E. Isufi, Learning Stochastic Graph Neural Networks with Constrained Variance, in the IEEE Transactions on Signal Processing, Jan. 2023. [PDF]
L. Ben Saad, B. Beferull-Lozano and E. Isufi, Quantization Analysis and Robust Design for Distributed Graph Filters, in the IEEE Transactions on Signal Processing, Dec. 2021. [PDF]
Z. Gao, E. Isufi and A. Ribeiro, Stability of Graph Convolutional Neural Networks to Stochastic Perturbations, Elsevier Signal Processing, Special Issue on Processing and Learning over Networks, Jun. 2021. [PDF]
Z. Gao, E. Isufi and A. Ribeiro, Stochastic Graph Neural Networks, in the IEEE Transactions on Signal Processing, 2021. [PDF]
M. Yang, M. Coutino, G. Leus and E. Isufi, Node-Adaptive Regularization for Graph Signal Reconstruction, IEEE Open Journal of Signal Processing, Jan. 2021. [PDF]
E. Isufi, A. Loukas, N. Perraudin and G. Leus, Forecasting Time Series with VARMA Recursions on Graphs, IEEE Transactions on Signal Processing, 2019. [PDF]
M. Coutino, E. Isufi and G. Leus, Advances in Distributed Graph Filtering, IEEE Transactions on Signal Processing, vol. 67 (9), pp. 2320 - 2333, 2019. [PDF]
P. Di Lorenzo, P. Banelli, E. Isufi, S. Barbarossa and G. Leus, Adaptive Graph Signal Processing: Algorithms and Optimal Sampling Strategies, IEEE Transactions on Signal Processing, vol. 66 (13), pp. 3584 - 3598, 2018. [PDF]
E. Isufi, A. Loukas, A. Simonetto and G. Leus, Filtering Random Graph Processes Over Random Time-Varying Graphs, IEEE Transactions on Signal Processing, vol.65 (16), pp. 4406-4421, 2017. [PDF]
E. Isufi, A. Loukas, A. Simonetto and G. Leus, Autoregressive Moving Average Graph Filtering, IEEE Transactions on Signal Processing, vol.67 (2), pp. 274-288, 2017. [PDF]

Google Sites

Report abuse