A5103104556- Complex Network Analysis Techniques
- Software-Defined Networks and 5G
- Bayesian Modeling and Causal Inference
- Opinion Dynamics and Social Influence
- Game Theory and Applications
- Evolutionary Game Theory and Cooperation
- Advanced Optical Network Technologies
- Advanced MIMO Systems Optimization
- Cooperative Communication and Network Coding
- Internet Traffic Analysis and Secure E-voting
- Advanced Graph Neural Networks
- Advanced Wireless Network Optimization
- Experimental Behavioral Economics Studies
- Access Control and Trust
- FinTech, Crowdfunding, Digital Finance
- Data Mining Algorithms and Applications
- Network Traffic and Congestion Control
- Auction Theory and Applications
- Advanced Database Systems and Queries
- Quantum many-body systems
- Reinforcement Learning in Robotics
- Mobile Agent-Based Network Management
- Artificial Intelligence in Games
- Multi-Agent Systems and Negotiation
- Caching and Content Delivery
Concordia University
2023-2024
Computer Research Institute of Montréal
2022-2023
Norwegian University of Science and Technology
2019
Soongsil University
2013-2014
University of Michigan
2008-2012
University of Arkansas at Fayetteville
2006
The Open Radio Access Network (O-RAN) architecture brings openness, intelligence, and virtualization to RANs, allowing multi-vendor existence, achieving economics of scale, enabling intelligent management orchestration. O-RAN components such as near real-time RAN Intelligent Controllers (RICs), Central Units (O-CUs), Distributed Unit (O-DUs) can be considered virtual network functions hosted on the O-Cloud. This allows service providers disaggregate from their hard-ware, dynamic...
Microscopic interpretation of histopathology images underlies many important diagnostic and treatment decisions. While advances in vision-language modeling raise new opportunities for analysis such images, the gigapixel-scale size whole slide (WSIs) introduces unique challenges. Additionally, pathology reports simultaneously highlight key findings from small regions while also aggregating across multiple slides, often making it difficult to create robust image-text pairs. As such, remain a...
With the 5th generation of mobile networking (5G) on our doorstep, optical network operators are reorganizing their infrastructure to deploy different topologies (virtual networks/applications) same demand. This new paradigm, called slicing, can be enabled by segmenting resources based requirements application level. paper investigates a nested decomposition mathematical modeling design reliable 5G slicing problem, i.e., every virtual path is protected against any single link failures...
Modeling information diffusion in networks enables reasoning about the spread of ideas, news, opinion, and technology across a network agents. Existing models generally assume given structure, practice derived from observations agent communication or other interactions. In many realistic settings, however, observing all connections is not feasible. We consider problem modeling when only partially observed, investigate two approaches. The first learns graphical model potentials for...
A dynamic model of a multiagent system defines probability distribution over possible behaviors time. Alternative representations for such models present tradeoffs in expressive power, and accuracy cost inferential tasks interest. In history-dependent representation, behavior at given time is specified as probabilistic function some portion history. Models may be further distinguished based on whether they specify individual or joint behavior. Joint are more expressive, but general grow...
Factored models of multiagent systems address the complexity joint behavior by exploiting locality in agent interactions. History-dependent graphical (hGMMs) further capture dynamics conditioning on history. The challenges modeling real human motivated us to extend hGMM representation distinguishing two types This distinction opens opportunity for learning dependence networks that are different from given structures representing observed We propose a greedy algorithm hGMMs time-series data,...
The Everything is Alive (EiA) agent system architecture modular and scalable. It uses the E2 platform, a derivative of Eclipse core to dynamically add or remove plugin capabilities that have XML WSDL interfaces. This paper describes two can be modularly added EiA agents. described include robot command capability customizes an operate as controller menu-based natural language interface (MBNLI) provides human-level for commanding robot.
Discovering periodic patterns in a customer transaction database is the task of identifying itemsets (sets items or values) that periodically appear sequence transactions.Numerous methods can identify exhibiting behavior.Nonetheless, problem these traditional approaches concept behavior defined very strictly.Indeed, pattern considered to be if amount time number transactions between all pairs its consecutive occurrences less than fixed maxP er (maximum periodicity) threshold.As result,...
Device-to-Device (D2D) communication has become an important technology in future networks, since it can provide advantages such as cell throughput enhancement, user equipment power saving, and increased instantaneous data rate. However, the performance of both D2D cellular be affected by mutual interferences between them. In this paper, we analyze a resource allocation method for links underlaying networks. A novel which allows more than one pair to reuse downlink is proposed. addition,...
Open and competitive marketplaces, like those enabled by the internet, bring enormous value to consumers while fostering innovation growth across a wide variety of businesses. Electronic commerce has forever changed how people trade goods services. Last year alone, US spent more than $150 billion dollars on internet sales [5]. However, developing countries largely have not benefited from these advances, often due low network penetration, lack locally relevant markets, requirements for...
In this paper, we analyze a resource allocation method for device-to-device (D2D) communication underlaying cellular networks. A novel algorithm is proposed to form number of D2D groups such that group pairs can reuse the downlink one user (CU) while guaranting SINR at both receivers and CUs.
I introduce a graphical representation for modeling multiagent systems based on different kinds of reasoning about agent behavior. seek to investigate this model's predictive and representative capabilities across various domains, examine methods learning the structure from interaction data. also propose explore framework's scalability in large real-world scenarios, such as social networks, evaluate its prediction performance with existing network behavior models.
Classical routing heuristics, e.g., Open Shortest Path First, have several significant issues, such as they are not able to generalize or adapt heterogeneous environments including dynamics of topology, traffic patterns, and Quality Service (QoS) requirements. To solutions, network operators recently utilized machine learning algorithms at centralized controllers. However, solutions scalable due many reasons, slow data transfer the central controller in a large network. Distributed...