Using dynamic proxy agent replicate groups to improve fault-tolerance in multi-agent systems
Bibliographic record
Abstract
Despite the considerable effort spent researching and developing multi-agent systems the actual number of deployed systems is still surprisingly small. One of the reasons for the significant gap between developed and deployed systems is their brittleness. Multi-agent systems are susceptible to all of the same faults as any distributed system, they lack centralized control components, which makes it difficult to detect and treat failures of individual agents, and the agents making up the system are social, thus risking fault-propagation These faults can impact system performance and lead to overall system failure. Multi-agent systems must be made more dependable before they will be deployed on a large scale. Using redundancy by replication of individual agents within a multi-agent system is one possible approach for improving fault-tolerance, and hence improving reliability and availability-two key components of dependability. Having a group of agents, a replicate group, act like an individual agent leads to increased complexity and system load, and it introduces new challenges to system construction. Using a message proxy, to handle communication for the group, and passive replication strategies effectively deals with the complexity and overhead issues. This paper presents an architecture for implementing agent replicate groups using a message proxy and passive replicate group management. Experimentation and application testing using an implementation of the architecture is presented. The architecture is demonstrated to be a viable technique for increasing dependability in multi-agent systems.
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 | 0.005 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 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".