Modular Information Bottleneck Encoding for Scalable Task Inference in IoT Networks
Bibliographic record
Abstract
The proliferation of intelligent IoT devices and the growing demand for low-latency, task-aware inference have placed significant strain on communication and computation resources in next-generation (6G) wireless networks. Nevertheless, it is still an open problem to enable a semantic communication that alleviates the memory management in memory-constrained IoT networks. In this paper, we propose a modular and memory-efficient semantic communication framework for distributed task inference in IoT networks. Leveraging the information bottleneck (IB) principle, we introduce a novel sequential multi-step feature encoding scheme, wherein each encoder block is trained independently with earlier blocks held constant. This design supports flexible and reusable deployment of pre-trained encoders across heterogeneous devices with memory constraints. Building on this foundation, we formulate a multi-gateway edge inference architecture, where gateways selectively offload compressed task-relevant features to a core edge server. To handle resource limitations at the gateway level, we model a nonlinear mixed-integer optimization problem that selects encoder blocks for tasks in a manner that maximizes inference quality while satisfying delay and memory constraints. A greedy heuristic algorithm is proposed to efficiently solve the resource allocation problem. Simulation results demonstrate that our framework achieves up to 40% reduction in normalized delay compared to baseline methods, while tripling the number of concurrently supported tasks in the network.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".