Leandros Tassiulas
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Leandros Tassiulas has a total of 183 co-authors. Here are the top co-authors by number of shared publications:
Thanasis Korakis
University of Thessaly,Department of Electrical and Computer Engineering,Greece
10 shared publications- George Iosifidis
Department of Software Technology, Delft University of Technology, Delft, The Netherlands
8 shared publications - Iordanis Koutsopoulos
Department of Informatics, Athens University of Economics and Business, Athens, Greece
7 shared publications - Ali Maatouk
Yale University, New Haven, USA
6 shared publications - Jianwei Huang
The Chinese University of Hong Kong
6 shared publications - Rex Ying
Yale University, New Haven, CT, USA
6 shared publications
Latest Publications
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 9183-9194
Fin-RATE: A Real-world Financial Analytics and Tracking Evaluation Benchmark for LLMs on SEC Filings
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 9183-9194
Open AccessFin-RATE: A Real-world Financial Analytics and Tracking Evaluation Benchmark for LLMs on SEC Filings
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 3655-3666
NumCache: KV Cache Compression and Retrieval for Financial Document QA
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 3655-3666
Open AccessNumCache: KV Cache Compression and Retrieval for Financial Document QA
- Open Access
- 1 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 86424-86441
Tele-LLMs: A Series of Specialized Large Language Models for Telecommunications
The emergence of large language models (LLMs) has significantly impacted various fields, from natural language processing to sectors like medicine and finance. However, despite their rapid proliferation, the applications of LLMs in telecommunications remain limited, often relying on general-purpose models that lack domain-specific specialization. This lack of specialization results in underperformance, particularly when dealing with telecommunications-specific technical terminology and their associated mathematical representations. This paper addresses this gap by first creating and disseminating Tele-Data, a comprehensive dataset of telecommunications material curated from relevant sources, and Tele-Eval, a large-scale question-and-answer dataset tailored to the domain. Through extensive experiments, we explore the most effective training techniques for adapting LLMs to the telecommunications domain, ranging from examining the division of expertise across various telecommunications aspects to employing parameter-efficient techniques.We also investigate how models of different sizes behave during adaptation and analyze the impact of their training data on this behavior. Leveraging these findings, we develop and open-source Tele-LLMs, the first series of language models ranging from 1B to 8B parameters, specifically tailored for telecommunications. Our evaluations demonstrate that these models outperform their general-purpose counterparts on Tele-Eval and telecommunications-related literature tasks while retaining their previously acquired capabilities, thus avoiding the catastrophic forgetting phenomenon.
- 1 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 86424-86441
Open AccessTele-LLMs: A Series of Specialized Large Language Models for Telecommunications
The emergence of large language models (LLMs) has significantly impacted various fields, from natural language processing to sectors like medicine and finance. However, despite their rapid proliferation, the applications of LLMs in telecommunications remain limited, often relying on general-purpose models that lack domain-specific specialization. This lack of specialization results in underperformance, particularly when dealing with telecommunications-specific technical terminology and their associated mathematical representations. This paper addresses this gap by first creating and disseminating Tele-Data, a comprehensive dataset of telecommunications material curated from relevant sources, and Tele-Eval, a large-scale question-and-answer dataset tailored to the domain. Through extensive experiments, we explore the most effective training techniques for adapting LLMs to the telecommunications domain, ranging from examining the division of expertise across various telecommunications aspects to employing parameter-efficient techniques.We also investigate how models of different sizes behave during adaptation and analyze the impact of their training data on this behavior. Leveraging these findings, we develop and open-source Tele-LLMs, the first series of language models ranging from 1B to 8B parameters, specifically tailored for telecommunications. Our evaluations demonstrate that these models outperform their general-purpose counterparts on Tele-Eval and telecommunications-related literature tasks while retaining their previously acquired capabilities, thus avoiding the catastrophic forgetting phenomenon.
- Open Access
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2806-2817
LitBench: A Graph-Centric Large Language Model Benchmarking Tool For Literature Tasks
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 2806-2817
Open AccessLitBench: A Graph-Centric Large Language Model Benchmarking Tool For Literature Tasks
- Open Access
- 21 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 14766-14776
Modeling RIS for Multi-User Resource Allocation Under Pricing-Aware Policies
Reconfigurable Intelligent Surfaces (RISs) transform the inherently stochastic wireless channel into a programmable, software-controlled medium, ensuring stable and reliable network performance. However, system-level orchestration that simultaneously satisfies operator pricing policies and user-specific performance guarantees remains largely unexplored. We introduce LEVEL, a low-runtime algorithm that allocates a shared RIS among heterogeneous users by directly constraining each user’s deviation from its performance objective, while ensuring that the resulting allocation aligns with operator-defined pricing tiers and service agreements. LEVEL couples physics-consistent codebook modeling with a lightweight per-element optimization that executes in the millisecond regime, enabling real-time use in beyond-5G/6G networks and scalable execution across deployments with multiple RIS units. Extensive full-wave emulations in an indoor multi-RIS setting and large-scale simulations show that LEVEL consistently outperforms existing multiplexing schemes across all pricing tiers while preserving tier-consistent fairness, quantified via a strong correlation between user tier and achieved efficiency. The results establish a practical path for policy-aware, fair, and predictable RIS sharing in future programmable wireless environments.
- 21 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 14766-14776
Open AccessModeling RIS for Multi-User Resource Allocation Under Pricing-Aware Policies
Reconfigurable Intelligent Surfaces (RISs) transform the inherently stochastic wireless channel into a programmable, software-controlled medium, ensuring stable and reliable network performance. However, system-level orchestration that simultaneously satisfies operator pricing policies and user-specific performance guarantees remains largely unexplored. We introduce LEVEL, a low-runtime algorithm that allocates a shared RIS among heterogeneous users by directly constraining each user’s deviation from its performance objective, while ensuring that the resulting allocation aligns with operator-defined pricing tiers and service agreements. LEVEL couples physics-consistent codebook modeling with a lightweight per-element optimization that executes in the millisecond regime, enabling real-time use in beyond-5G/6G networks and scalable execution across deployments with multiple RIS units. Extensive full-wave emulations in an indoor multi-RIS setting and large-scale simulations show that LEVEL consistently outperforms existing multiplexing schemes across all pricing tiers while preserving tier-consistent fairness, quantified via a strong correlation between user tier and achieved efficiency. The results establish a practical path for policy-aware, fair, and predictable RIS sharing in future programmable wireless environments.
- 19 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 25 (6) , 7640-7656
Age Optimal Sampling for Unreliable Channels Under Unknown Channel Statistics
In this paper, we study a system in which a sensor forwards status updates to a receiver through an error-prone channel, while the receiver sends the transmission results back to the sensor via a reliable channel. Both channels are subject to random delays. To evaluate the timeliness of the status information at the receiver, we use the Age of Information (AoI) metric. The objective is to design a sampling policy that minimizes the expected time-average AoI, even when the channel statistics (e.g., delay distributions) are unknown. We first review the threshold structure of the optimal offline policy under known channel statistics and then reformulate the design of the online algorithm as a stochastic approximation problem. We propose a Robbins-Monro algorithm to solve this problem and demonstrate that the optimal threshold can be approximated almost surely. Moreover, we prove that the cumulative AoI regret of the online algorithm increases with rate$\mathcal {O}(\ln K)$ , where$K$ is the number of successful transmissions. In addition, our algorithm is shown to be minimax order optimal, in the sense that for any online learning algorithm, the cumulative AoI regret up to the$K$ -th successful transmissions grows with the rate at least$\Omega (\ln K)$ in the worst case delay distribution. Finally, we improve the stability of the proposed online learning algorithm through a momentum-based stochastic gradient descent algorithm. Simulation results validate the performance of our proposed algorithm.- 19 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 25 (6) , 7640-7656
Age Optimal Sampling for Unreliable Channels Under Unknown Channel Statistics
In this paper, we study a system in which a sensor forwards status updates to a receiver through an error-prone channel, while the receiver sends the transmission results back to the sensor via a reliable channel. Both channels are subject to random delays. To evaluate the timeliness of the status information at the receiver, we use the Age of Information (AoI) metric. The objective is to design a sampling policy that minimizes the expected time-average AoI, even when the channel statistics (e.g., delay distributions) are unknown. We first review the threshold structure of the optimal offline policy under known channel statistics and then reformulate the design of the online algorithm as a stochastic approximation problem. We propose a Robbins-Monro algorithm to solve this problem and demonstrate that the optimal threshold can be approximated almost surely. Moreover, we prove that the cumulative AoI regret of the online algorithm increases with rate$\mathcal {O}(\ln K)$ , where$K$ is the number of successful transmissions. In addition, our algorithm is shown to be minimax order optimal, in the sense that for any online learning algorithm, the cumulative AoI regret up to the$K$ -th successful transmissions grows with the rate at least$\Omega (\ln K)$ in the worst case delay distribution. Finally, we improve the stability of the proposed online learning algorithm through a momentum-based stochastic gradient descent algorithm. Simulation results validate the performance of our proposed algorithm.- 11 December 2025
- Published by Elsevier in Computer Networks
- Vol. 275
AGORAN: An agentic open marketplace for 6G RAN automation
- 11 December 2025
- Published by Elsevier in Computer Networks
- Vol. 275
AGORAN: An agentic open marketplace for 6G RAN automation
- Open Access
- 2 December 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies
- Vol. 9 (4) , 1-27
EC-Sense: Radio E nergy C apture for Ultra-Low-Power Wireless Soil Moisture Sens ing
Understanding spatial and temporal variations in soil moisture is critical for sustainable agriculture. Contactless approaches such as ground-penetrating radar and satellite sensing are either expensive or offer limited resolution, making them less scalable. In-situ sensors provide higher resolution but often involve challenging data retrieval processes and wired infrastructure. In this work, we propose EC-Sense, a wire-free system for in-situ soil moisture sensing that employs ultra-low-power sensor Tags buried underground and reference Tags placed on the surface. EC-Sense introduces a novel sensing modality based on energy capture time (EC Time), defined as the time taken by a sensor Tag to harvest sufficient energy to activate a response. EC Time serves as a proxy for path loss in soil, which correlates strongly with soil moisture. By leveraging differential path loss—estimated from the EC Times of both surface and buried Tags—EC-Sense isolates soil-induced attenuation from environmental effects above ground. This differential sensing approach, combined with a decoupled sensing and communication architecture, enables EC-Sense to overcome limitations of existing wideband RF-based sensors. Despite using an active radio, our Tags consume only 3.7 μW on average and achieve a projected lifetime of 10 years under realistic conditions. We deployed EC-Sense in an open agricultural field for over a week and measured soil moisture at multiple depths daily, achieving 98% accuracy compared to ground truth. We further evaluate EC-Sense across three representative soil types—sandy loam, silt loam, and silty clay loam—and demonstrate reliable sensing at depths of 40 cm, 30 cm, and 25 cm, respectively, at field capacity.
- 2 December 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies
- Vol. 9 (4) , 1-27
Open AccessEC-Sense: Radio E nergy C apture for Ultra-Low-Power Wireless Soil Moisture Sens ing
Understanding spatial and temporal variations in soil moisture is critical for sustainable agriculture. Contactless approaches such as ground-penetrating radar and satellite sensing are either expensive or offer limited resolution, making them less scalable. In-situ sensors provide higher resolution but often involve challenging data retrieval processes and wired infrastructure. In this work, we propose EC-Sense, a wire-free system for in-situ soil moisture sensing that employs ultra-low-power sensor Tags buried underground and reference Tags placed on the surface. EC-Sense introduces a novel sensing modality based on energy capture time (EC Time), defined as the time taken by a sensor Tag to harvest sufficient energy to activate a response. EC Time serves as a proxy for path loss in soil, which correlates strongly with soil moisture. By leveraging differential path loss—estimated from the EC Times of both surface and buried Tags—EC-Sense isolates soil-induced attenuation from environmental effects above ground. This differential sensing approach, combined with a decoupled sensing and communication architecture, enables EC-Sense to overcome limitations of existing wideband RF-based sensors. Despite using an active radio, our Tags consume only 3.7 μW on average and achieve a projected lifetime of 10 years under realistic conditions. We deployed EC-Sense in an open agricultural field for over a week and measured soil moisture at multiple depths daily, achieving 98% accuracy compared to ground truth. We further evaluate EC-Sense across three representative soil types—sandy loam, silt loam, and silty clay loam—and demonstrate reliable sensing at depths of 40 cm, 30 cm, and 25 cm, respectively, at field capacity.
- Open Access
- 30 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 14-20
Towards Unsupervised Drift Detection in Programmable Data-Planes
- 30 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 14-20
Open AccessTowards Unsupervised Drift Detection in Programmable Data-Planes
- Open Access
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 3728-3739
LitFM: A Retrieval Augmented Structure-aware Foundation Model For Citation Graphs
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 3728-3739
Open AccessLitFM: A Retrieval Augmented Structure-aware Foundation Model For Citation Graphs