LSTM Cache Mechanism for Information-Centric Networking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Information-centric networking (ICN) lacks an efficient cache mechanism, which is crucial for managing network congestion and data transmission efficiency, particularly due to the challenges in accurately predicting long-term environmental states using reinforcement learning (RL) models.
Innovation Solution
The method involves using a memory layer, specifically a Long Short-Term Memory (LSTM) layer, to perform forward and backward processing on ICN states, determining hidden states, and training the machine learning model with future information to learn a more accurate model for an efficient cache mechanism.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If reinforcement learning models are used for cache mechanism in ICN, then cache strategy optimization is improved, but prediction accuracy of long-term environmental states deteriorates
Solution Approach 1:
The patent applies preliminary action by performing backward processing from future time steps to obtain future hidden states before making cache decisions. The LSTM network processes sequences in reverse order (from T to 1), allowing the model to incorporate future environmental state information into current cache strategies, thereby improving prediction accuracy without compromising optimization performance
Solution Approach 2:
The patent introduces a new dimension of time processing by implementing bidirectional LSTM that processes sequences both forward (from 1 to T) and backward (from T to 1). This dual-directional approach adds temporal depth to the model, enabling it to capture both past and future contextual information, thus resolving the contradiction between optimization speed and prediction accuracy
2Device complexity
If traditional cache mechanisms are used in ICN, then system complexity is reduced, but data access efficiency and traffic pressure management deteriorate
Solution Approach 1:
The patent replaces traditional mechanical cache management systems with an intelligent LSTM-based predictive system. Instead of using fixed rules or simple algorithms for cache decisions, the system employs neural network processing to dynamically predict future states and optimize cache strategies, significantly improving data access efficiency while the modular LSTM architecture keeps implementation complexity manageable
Data Source
AI summary
Embodiments of the present disclosure provide a method, an electronic device, and a computer program product for information-centric networking. In the method, a memory layer in a machine learning model is used to obtain, on the basis of an environmental state obtained from information-centric networking at a future moment, future information associated with a memory layer corresponding to the future moment, and the machine learning model is trained using the future information. By means of the solution, a model trained using future information can be obtained. By use of the model, information-centric networking based on reinforcement learning achieves a more efficient cache mechanism.