Smart Proactive Caching: Empower the Video Delivery for Autonomous Vehicles in ICN-Based Networks
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
The recent advances in vehicular communications and networking are bringing self-driving vehicles closer to reality. Once full automation (i.e., levels 4 and 5 as defined in the five levels of autonomous driving) becomes a reality, entertainment services for both drivers and passengers will shift from listening to radios (or music) to watching videos. Hence, how to improve the quality of experience (QoE) for autonomous vehicle (AV) users, and to reduce the network load will become a crucial problem. Information-centric networking (ICN) is seen as one of the potential paradigms for next-generation networks, and could potentially be used for content distribution in vehicular networks. Caching, an important feature in ICN-based networks, is an efficient way to reduce network load, and to improve QoE for users. However, traditional reactive caching approaches are inefficient for AV users due to the high delay caused by their high speed. In this paper, we propose a novel hierarchical proactive caching approach that considers both users' future demands and AV user mobility. The proposed approach uses the non-negative matrix factorization (NMF) technique to predict user's preferences which are then used to predict users' future demands by considering the historical popularity of videos. A user mobility prediction model is used to predict the AV users' next location based on the current location and the planned route information which can be retrieved from the self-driving system. Based on the predicted users' future demands and locations of AVs, the proposed caching approach can proactively cache videos that are likely to get requested at the next base station (BS) or roadside unit (RSU) that the users are moving to. The proposed approach has been evaluated under two scenarios: a highway scenario and a grid street scenario. Results show that the proposed approach can significantly improve the efficiency of caching in terms of hit ratio and the average number of hops.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 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 it