A multiagent social interaction model for autonomous vehicle testing
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
Social interaction capability (SIC) is essential for autonomous vehicles (AVs) when they interact with surrounding vehicles, as the ability of understanding and reacting to the behaviors of other road users can significantly enhance AVs’ rapid deployment. Virtual simulation testing is a core approach for evaluating AVs, including their SIC, on the basis of traffic simulation models. However, existing simulation models focus mainly on generating accurate vehicle trajectories and do not explicitly model the high-level sociality nature of interaction decisions that guide specific movements. This study aims to address this gap by developing a multiagent simulation model for the social interaction of human driving behavior on the basis of the multiagent imitation learning (MAIL) approach, which is referred to as the Social-MAIL model. Specifically, to quantify the sociality of decisions, we introduce social value orientation into the reward function to quantify cooperation or competition intent and guide the generation of social driving behaviors. Furthermore, to fully depict the complex interaction environment, we develop a heterogeneous policy network with temporal‒spatial attention mechanisms to describe the impact of multiple interactive objects and historical states on driving behavior. Through training and validation on the SinD dataset, we demonstrate that, compared with a set of baseline models, the proposed Social-MAIL model can accurately capture complex and time-varying social intent and reproduce the most realistic vehicle trajectories and macroscopic traffic flow characteristics at intersections. Moreover, we apply the Social-MAIL model for evaluating the SIC of AVs via comparison experiments.
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 | 0.001 | 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 | 0.000 | 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