Semantic-Aware Collaborative Deep Reinforcement Learning Over Wireless Cellular Networks

Lotfi, Fatemeh; Semiari, Omid; Saad, Walid

doi:10.1109/icc45855.2022.9839122

Cited by 22 publications

(6 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Collaborative RL is a form of RL in which multiple agents are present in the system, and they collaborate to determine the optimum policy. In [90], collaborative deep RL (CDRL) is used to train a group of heterogeneous agents over a wireless cellular network. First, the algorithm selects the best subset of semantically relevant DRL agents for collaboration.…”

Section: Loss Function Expression Descriptionmentioning

confidence: 99%

“…Looking to implement RL-based semantic communication for a specific application, [91] proposes a DRL framework for air-to-ground URLLC communcation using unmanned aerial vehicles (UAV's). Similar to [90], this work proposes the use of a multi-agent DRL framework, coined graph attention exchange network (GAXNet), for semantic communication. Self-attention is used to determine the attention a UAV gives to other UAV's in the network, and based on this attention, training is performed in a centralized manner.…”

Section: Loss Function Expression Descriptionmentioning

confidence: 99%

See 1 more Smart Citation

Engineering Semantic Communication: A Survey

Wheeler

Natarajan

2023

IEEE Access

View full text Add to dashboard Cite

As the global demand for data has continued to rise exponentially, some have begun turning to the idea of semantic communication as a means of efficiently meeting this demand. Pushing beyond the boundaries of conventional communication systems, semantic communication focuses on the accurate recovery of the meaning conveyed from source to receiver, as opposed to the accurate recovery of transmitted symbols. In this survey, we aim to provide a comprehensive view of the history and current state of semantic communication and the techniques for engineering this higher level of communication. A survey of the current literature reveals four broad approaches to engineering semantic communication. We term the earliest of these approaches classical semantic information, which seeks to extend informationtheoretic results to include semantic information. A second approach makes use of knowledge graphs to achieve semantic communication, and a third utilizes the power of modern deep learning techniques to facilitate this communication. The fourth approach focuses on the significance of information, rather than its meaning, to achieve efficient, goal-oriented communication. We discuss each of these four approaches and their corresponding studies in detail, and provide some challenges and opportunities that pertain to each approach. Finally, we introduce a novel approach to semantic communication, which we term contextbased semantic communication. Inspired by the way in which humans naturally communicate with one another, this context-based approach provides a general, optimization-based design framework for semantic communication systems. Together, this survey provides a useful guide for the design and implementation of semantic communication systems.

show abstract

Section: Loss Function Expression Descriptionmentioning

confidence: 99%

Section: Loss Function Expression Descriptionmentioning

confidence: 99%

Engineering Semantic Communication: A Survey

Wheeler

Natarajan

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Therefore, leveraging the advantage of the attention layer makes sense. Inspired by [20][21][22][23], a selective inclusion/exclusion attention mechanism is applied to all projections of the embedding vector, anchor, and negative vectors to prevent outliers' contribution to the DML loss function. As mentioned before, this module also defines adaptive margins.…”

Section: B Classifiermentioning

confidence: 99%

Triplet Loss-less Center Loss Sampling Strategies in Facial Expression Recognition Scenarios

Hossein¹,

Lotfi²,

Atyabi³

et al. 2023

Preprint

View full text Add to dashboard Cite

Facial expressions convey massive information and play a crucial role in emotional expression. Deep neural network (DNN) accompanied by deep metric learning (DML) techniques boost the discriminative ability of the model in facial expression recognition (FER) applications. DNN, equipped with only classification loss functions such as Cross-Entropy cannot compact intra-class feature variation or separate inter-class feature distance as well as when it gets fortified by a DML supporting loss item. The triplet center loss (TCL) function is applied on all dimensions of the sample's embedding in the embedding space. In our work, we developed three strategies: fully-synthesized, semi-synthesized, and prediction-based negative sample selection strategies. To achieve better results, we introduce a selective attention module that provides a combination of pixel-wise and element-wise attention coefficients using high-semantic deep features of input samples. We evaluated the proposed method on the RAF-DB, a highly imbalanced dataset. The experimental results reveal significant improvements in comparison to the baseline for all three negative sample selection strategies.

show abstract

“…This semantic-aware scheme has been demonstrated [272] to attain lower latency with higher reliability than the state-of-the-art QMIX scheme (see [273]). Also in the context of collaborative deep RL (CDRL), the authors of [274] propose a semantic-aware CDRL framework that enables knowledge to be efficiently transferred among heterogeneous agents that are distributed across a resource-constrained wireless cellular network and have semantically related tasks. They, therefore, introduce a new heterogeneous federated DRL algorithm [274] for selecting the best subset of semantically-associated DRL agents for collaboration.…”

Section: Algorithmic Developments In Semantic-aware Communication And...mentioning

confidence: 99%

“…Also in the context of collaborative deep RL (CDRL), the authors of [274] propose a semantic-aware CDRL framework that enables knowledge to be efficiently transferred among heterogeneous agents that are distributed across a resource-constrained wireless cellular network and have semantically related tasks. They, therefore, introduce a new heterogeneous federated DRL algorithm [274] for selecting the best subset of semantically-associated DRL agents for collaboration. This CDRL algorithm has been shown to offer an 83% improvement in maximum reward compared with baseline methods [274].…”

Section: Algorithmic Developments In Semantic-aware Communication And...mentioning

confidence: 99%

Semantic Communication: A Survey on Research Landscape, Challenges, and Future Directions

Getu,

Kaddoum,

Bennis

2023

Preprint

View full text Add to dashboard Cite

<p> Amid the global rollout of fifth-generation (5G) services, researchers in academia, industry, and national labo?ratories have been developing proposals for the sixth generation (6G). Despite the many 6G proposals, the materialization of 6G as presently envisaged is fraught with many fundamental interdisciplinary, multidisciplinary, and transdisciplinary (IMT) challenges. To alleviate some of these challenges, semantic com?munication (SemCom) has emerged as promising 6G technology enabler. SemCom is designed to transmit only semantically?relevant information and hence help to minimize power usage, bandwidth consumption, and transmission delay. Thus, SemCom embodies a paradigm shift that can change the status quo that wireless connectivity is an opaque data pipe carrying messages whose context-dependent meaning have been ignored. On the other hand, 6G is critical for the materialization of major SemCom use cases. These paradigms of 6G for SemCom and SemCom for 6G call for a tighter integration and marriage of 6G and SemCom. To facilitate this integration and marriage, this comprehensive survey paper first provides the fundamentals of semantics and semantic information, semantic representation, semantic information, and semantic entropy. It then builds on this understanding and details the state-of-the-art research landscape of SemCom; exposes the fundamental and major challenges of SemCom; and offers promising future research directions for SemCom theories, algorithms, and realization. Accordingly, this survey article stimulates major streams of research on SemCom theories, algorithms, and implementation. </p>

show abstract

Semantic-Aware Collaborative Deep Reinforcement Learning Over Wireless Cellular Networks

Cited by 22 publications

References 7 publications

Engineering Semantic Communication: A Survey

Engineering Semantic Communication: A Survey

Triplet Loss-less Center Loss Sampling Strategies in Facial Expression Recognition Scenarios

Semantic Communication: A Survey on Research Landscape, Challenges, and Future Directions

Contact Info

Product

Resources

About