PaperlayerView exact revision on arXiv

Source-linked AI summary

Performance Optimization for Semantic Communications: An Attention-based Reinforcement Learning Approach

Yining Wang, Mingzhe Chen, Tao Luo, Walid Saad, Dusit Niyato, H. Vincent Poor, Shuguang Cui

arXiv:2208.08239v1cs.ITcs.AI

TL;DR

The paper addresses semantic text transmission under wireless resource and delay constraints, where only part of the semantic information may be sent. It proposes an attention-integrated APPO algorithm for selecting semantic triples and allocating resource blocks, and reports higher total MSS and faster convergence than comparison methods.

  • Problem

    Wireless resource and delay constraints require jointly selecting resource blocks and partial semantic information while maximizing the total MSS of recovered text.

  • Method

    The framework models text meaning as knowledge-graph triples and uses attention-integrated APPO to estimate triple importance, select semantic information, and optimize resource-block allocation.

  • Results

    The proposed APPO improves total MSS over comparison algorithms across evaluated settings and converges after 100 iterations, versus about 300 for APG and 250 for ADQN.

  • Takeaways & Limitations

    The approach enables the base station to allocate limited resource blocks according to semantic importance while transmitting less data and improving total MSS.

Abstract

from arXiv · show

In this paper, a semantic communication framework is proposed for textual data transmission. In the studied model, a base station (BS) extracts the semantic information from textual data, and transmits it to each user. The semantic information is modeled by a knowledge graph (KG) that consists of a set of semantic triples. After receiving the semantic information, each user recovers the original text using a graph-to-text generation model. To measure the performance of the considered semantic communication framework, a metric of semantic similarity (MSS) that jointly captures the semantic accuracy and completeness of the recovered text is proposed. Due to wireless resource limitations, the BS may not be able to transmit the entire semantic information to each user and satisfy the transmission delay constraint. Hence, the BS must select an appropriate resource block for each user as well as determine and transmit part of the semantic information to the users. As such, we formulate an optimization problem whose goal is to maximize the total MSS by jointly optimizing the resource allocation policy and determining the partial semantic information to be transmitted. To solve this problem, a proximal-policy-optimization-based reinforcement learning (RL) algorithm integrated with an attention network is proposed. The proposed algorithm can evaluate the importance of each triple in the semantic information using an attention network and then, build a relationship between the importance distribution of the triples in the semantic information and the total MSS. Compared to traditional RL algorithms, the proposed algorithm can dynamically adjust its learning rate thus ensuring convergence to a locally optimal solution.

I. INTRODUCTION

The paper proposes a semantic communication framework that extracts text meaning into knowledge graphs, evaluates recovered-text quality with MSS, and optimizes partial transmission under wireless constraints using attention-based RL.

  • Contribution: Prior work includes semantic communication systems and evaluation metrics, while this paper jointly considers semantic extraction, transmission, recovery, and performance evaluation.The paper positions this joint framework as its main contribution.
  • Framework: The framework extracts text meaning into a knowledge graph of entity-relation-entity semantic triples, which users convert back into text with graph-to-text generation.The semantic pipeline jointly covers extraction, transmission, and recovery.
  • Performance metric: MSS measures whether recovered information is correct and whether it contains all information from the original text.The metric jointly captures semantic accuracy and semantic completeness.
  • Problem formulation: The BS jointly optimizes resource allocation and the semantic information selected for transmission to maximize total MSS under delay constraints.The optimization addresses partial semantic-information transmission when wireless resources are limited.
  • Proposed algorithm: The attention PPO algorithm evaluates triple importance and relates importance distributions to total MSS when selecting semantic information and allocating resources.This relationship supports effective policies for resource allocation and semantic-information transmission.
  • Reported results: Compared with a standard communication network, APPO reduces transmitted words by up to 41.3% while achieving a 2-fold improvement in total MSS.The paper also analyzes convergence and reports a locally optimal solution guarantee.

B. Transmission Model

The transmission model uses OFDMA resource blocks to serve users, with allocation constrained so each block and each user participate in at most one assignment.

  • Resource-block transmission: OFDMA provides Q downlink orthogonal resource blocks for semantic-information transmission.The model allocates these blocks to serve the users.
  • Resource allocation: Each user's allocation is a binary vector, where α_i,q = 1 means resource block q is assigned to user i.The model assumes each resource block is allocated to only one user and each user occupies only one resource block.
  • Channel model: The downlink capacity depends on each resource block's bandwidth, BS transmit power, interference, noise, and the user channel gain.The channel gain incorporates Rayleigh fading and BS-user distance.
  • Delay constraint: A transmission-delay limit D determines the maximum number of semantic-information tokens that can be sent to each user.The model assumes constant transmit power and allows transmit-power optimization as an extension.

C. Text Recovery

The BS transmits partial semantic information selected under the delay threshold, after which each user uses graph-to-text generation to recover coherent text.

  • Partial semantic information: The BS transmits a selected subset of semantic triples as the partial semantic information for user i.The selected graph must satisfy the transmission-delay constraint.
  • Text recovery: After receiving the partial graph, each user applies a graph-to-text generation model to recover coherent multi-sentence text.The model is assumed to be well-trained and shared among all users.
  • Recovery quantities: The recovered text length is represented by M_i, while H_i denotes the number of selected semantic triples.These quantities describe the partial semantic information and its recovered output.

D. Semantic Similarity Model

The metric of semantic similarity (MSS) evaluates recovered text using semantic accuracy and semantic completeness, combining both through a weighted formulation with a short-text penalty.

  • MSS jointly captures whether recovered information is correct and whether the recovered text contains the original information.
  • Semantic accuracy is the ratio of correct token occurrences in recovered text to all token occurrences in recovered text.
  • Semantic completeness is the ratio of correct token occurrences in recovered text to token occurrences in the original text.
  • The MSS combines semantic accuracy and completeness using weight parameter ϕ.The parameter ϕ adjusts the contributions of the two components.
  • An additional penalty ξ_i accounts for short recovered text, while token matching determines correct occurrences.If a recovered token appears more often than in the original, only the matching occurrences count as correct.
  • Changing ξ_i allows users to use received partial semantic information for text recovery while controlling the accuracy–completeness tradeoff.

E. Problem Formulation

The paper formulates maximizing the total MSS of users’ recovered texts under transmission-delay and resource-block constraints, jointly selecting semantic information and allocating resource blocks.

  • The optimization objective is to maximize total MSS across all users while satisfying the transmission delay requirement.
  • The problem jointly optimizes each user’s resource-block allocation and the partial semantic information transmitted.
  • Each user can occupy only one resource block, and each resource block can be allocated to only one user.
  • The total MSS depends on the selected semantic information, resource-block allocation, and a neural-network text-generation model.
  • The optimization is difficult because its non-convex objective relationship cannot be accurately characterized and semantic selection is coupled with resource-block allocation.

III. ATTENTION RL FOR SEMANTIC INFORMATION SELECTION AND RESOURCE ALLOCATION

The proposed APPO algorithm integrates PPO-based reinforcement learning with an attention network to evaluate semantic-triple importance and optimize resource allocation and semantic information selection.

  • APPO combines a proximal policy optimization-based reinforcement-learning algorithm with an attention network.
  • The attention network calculates an importance value for each semantic triple.
  • APPO relates the importance distribution of semantic triples to total MSS when optimizing resource-block allocation and semantic information selection.
  • The trained text-recovery model is shared by users and the base station, enabling the base station to obtain total MSS after allocation and partial selection are determined.

A. Attention Network for Importance Evaluation

The attention network converts semantic-triple and text-token representations into normalized importance distributions, allowing the base station to evaluate triple relevance for optimization.

  • A softmax layer normalizes triple importance values into an importance vector whose length equals the number of semantic triples.
  • The attention network receives token-vector sequences representing a semantic triple and the original text.
  • Its hidden layer finds correlations between tokens in each semantic triple and tokens in the original text.
  • The network uses parameter matrices W_tri and W_tex, each with dimensions D_a × D_x, to compute token representations.
  • The output layer produces an importance value for each semantic triple.
  • Each normalized importance value y_i lies in [0, 1] and represents the importance of its corresponding semantic triple.
  • Because MSS depends on received semantic information and resource-block allocation, the importance distribution alone cannot directly yield the optimal allocation and selection.

B. Components of the APPO Algorithm

The APPO agent is the base station, whose actions allocate resource blocks and implicitly select the most important semantic triples within the delay constraint. Its state represents semantic-information importance distributions, while its policy maps states to actions that maximize total MSS.

  • The BS acts as the agent, determining resource-block allocation and the partial semantic information transmitted to each user.
  • Each action is a resource-block allocation vector satisfying the optimization constraints; allocated blocks determine which important triples can be transmitted before the delay threshold.
  • For each user, the transmitted partial semantic information contains the most important triples in its semantic information subject to the delay constraint.
  • The state is the importance distribution of all users' semantic information, represented as s = [f(G1), . . . , f(GU)] in a continuous state space.
  • A DNN parameterized by θ maps each state to an action probability πθ(s, a) = P(a|s), with the objective of maximizing total MSS.
  • The reward is the total MSS resulting from an action at a state, so maximizing the APPO reward solves the total-MSS maximization problem.

C. APPO for Total MSS Maximization

APPO is trained offline from historical semantic-communication data using trajectories sampled from a stored policy. It optimizes total MSS while penalizing policy divergence, then adjusts the learning rate through the penalty coefficient to support local convergence.

  • APPO is trained offline using historical semantic-communication data for resource-block allocation and semantic-information selection.
  • The stored policy samples K resource-block allocation actions, which form the training set for the policy that maximizes total MSS.
  • The policy objective combines expected total-MSS reward with a penalty for the Kullback–Leibler divergence from the stored policy.
  • Policy parameters are updated by gradient ascent using a learning rate δ to improve total MSS.
  • The penalty coefficient λ is adjusted according to the divergence between stored and updated policies, controlling learning-rate changes.
  • Iterative policy updates learn the relationship between semantic-information importance distributions and total MSS, yielding a policy for resource allocation and semantic-information selection.

D. Complexity and Convergence of APPO

APPO complexity depends on the users, resource blocks, semantic triples, and policy-network structure. Under stated conditions, the algorithm is reported to converge to a locally optimal solution of the total-MSS problem.

  • Complexity: APPO complexity includes attention-network computation for semantic-triple importance distributions and trained-policy computation for resource allocation.
  • Complexity: The attention-network parameter matrices have size Da × Dx, with the stated assumption Da ≪ Dx, including Da = 64 and Dx = 500 as an example.
  • Complexity: The policy action space contains all possible resource allocations, with size |A| = U! (U−Q)!, while the state depends on semantic-information importance distributions.
  • Complexity: Given the APPO structure, computational complexity depends on the number of users, resource blocks, and semantic triples in each extracted semantic information.
  • Convergence: APPO convergence requires a lower-bound objective, evaluable actions sampled from a stored policy, and monotonic increases in the objective during updates.
  • Convergence: The KL-divergence penalty controls policy differences, while learning-rate adjustment supports convergence to a locally optimal solution.

IV. SIMULATION RESULTS AND ANALYSIS

The simulations show that attention-guided semantic extraction and APPO-based partial transmission reduce transmitted data while improving MSS, convergence, stability, and robustness across evaluated settings.

  • Semantic information extraction: The attention network estimates semantic-triple importance from context-dependent token representations, enabling the BS to identify which triples matter most.Correlations between triples and text tokens are computed from contextual representations rather than word-frequency statistics.
  • Data reduction: In one example, transmitting 86 semantic-information tokens instead of 178 original-text tokens reduces the data size by 51.7%.The semantic triples are assigned an importance distribution summing to 1; the example contrasts a triple importance of 0.1923 with another of 0.0385.
  • Partial semantic transmission: The APPO algorithm selects important triples and allocates resource blocks, improving total MSS by 36.7% over baseline a) and 2-fold over baseline b) in the illustrated example.The recovered text based on selected important triples contains the original text’s main meaning.
  • Data reduction: Compared with baselines c) and e), the framework reduces transmitted data size by up to 41.3% and 84%, respectively.The reduction becomes larger for longer texts because multiple sentences can be represented by a single semantic triple.
  • Algorithm performance: APPO converges after 100 iterations, versus about 300 for APG and 250 for ADQN, while improving MSS by 11.2% and 18.5%, respectively.The reported convergence comparison uses a semantic communication network with 10 RBs and averages results over 10 independent runs.
  • Algorithm performance: On DocRED, APPO improves total MSS by 17.8%, 26.8%, and 2-fold over baselines a), b), and c), respectively; on AGENDA, the gains are 16.9%, 22.3%, and 1-fold.The paper attributes these gains to dynamic learning-rate adjustment and transmitting important semantic triples under limited wireless resources.

V. CONCLUSION

The proposed semantic communication framework models textual meaning with a knowledge graph and optimizes resource allocation and partial semantic transmission to maximize total MSS. APPO uses attention-based triple importance to improve total MSS while reducing transmitted data.

  • The framework models textual semantic information with a knowledge graph and introduces MSS to capture semantic accuracy and completeness.
  • Total MSS is evaluated as the number of resource blocks, users, and masked tokens varies.
  • The optimization jointly selects resource blocks and partial semantic information under wireless resource and transmission-delay constraints.
  • APPO estimates triple importance with attention and links importance distributions to total MSS for resource allocation and semantic-information selection.
  • The proposed framework significantly reduces the data size required for transmission while increasing total MSS.
Loading 2208.08239v1…