This paper presents TRACK (Telemetry-Based Racing Analysis and Coaching Kit), which is a framework for analyzing driving performance in sim racing and profiling how individual drivers behave behind th...
arXiv:2606. 12616v1 Announce Type: new Abstract: Closed-loop driving simulators typically populate their environments with non-ego traffic agents that behave largely the same way, produced either by rule-based traffic managers or by learned models trained toward a single behavioral mode.
arXiv:2607. 06495v1 Announce Type: cross Abstract: Live sports commentary is grounded generation under a deadline: statements concern real, named athletes, the grounding state changes every few seconds, and no reference text exists at generation time.
arXiv:2607. 26061v1 Announce Type: new Abstract: Pre-match tactical decision-making in professional football relies heavily on subjective expert analysis and identity-based scouting systems that cannot generalize to unseen teams.
The paper introduces Context-driven Personalized ACC (CoP-ACC), a data‑driven framework that learns from drivers’ throttle overrides to tailor Adaptive Cruise Control behavior. It uses unsupervised clustering to identify representative acceleration profiles, a context classifier to select the appropriate profile based on pre‑maneuver conditions, and a residual regressor to smooth the final profile. Evaluations on real‑world public‑road data show that CoP-ACC reconstructs driver‑expected acceleration patterns more accurately than a standard forced‑ACC baseline, suggesting it can reduce manual interventions and improve ride comfort.
The paper proposes a method for verifying whether two gameplay replays in Counter‑Strike 2 belong to the same player by extracting a behavioral fingerprint that captures crosshair control, movement, economy, combat, and rhythm. Using a six‑fold evaluation on two datasets (Perfect and Professional), the pairwise model achieves ROC AUCs of 0.926 and 0.956, with aiming and low‑level mechanics providing the strongest identity signals. Aggregating evidence across multiple historical demos further improves account‑history AUC, reaching 0.982 on Perfect and 0.975 on Professional.
SpeedrunBench is a new benchmark that tests large language model agents on their ability to develop and refine strategies for completing nine different video games as quickly as possible. The benchmark requires agents to repeatedly improve, reflect, and reason over long action horizons to beat both themselves and other players. Experiments show that while agents can approach human records on simple platformers, they lag behind humans on longer, more complex games within realistic resource limits.
arXiv:2602. 23056v2 Announce Type: replace Abstract: In Formula 1, race strategies are adapted according to evolving race conditions and competitors' actions.
arXiv:2607. 00190v1 Announce Type: cross Abstract: Recent advances in reinforcement learning have produced superhuman agents across a wide range of competitive games.
arXiv:2604. 01608v4 Announce Type: replace Abstract: Multi-agent systems (MAS) tackle complex tasks by distributing expertise, though this often comes at the cost of heavy coordination overhead, context fragmentation, and brittle phase ordering.
DrivingBench is the first benchmark that tests general‑purpose vision‑language models on the task of driving a real Toyota Corolla around a parking‑lot cone course. The models receive live camera frames and issue steering and velocity commands, with inference latency counted as part of the challenge. In tests, only GPT‑6 Astra completed the course, while other models showed limited progress or failed to pass half the course.
The paper introduces an edge‑assisted crowdsensing framework that uses handlebar‑mounted cameras to capture street‑level video, a vision‑language model to extract structured traffic attributes, and an explainable boosting machine to combine these with OpenStreetMap data and on‑bike kinematics for a 1‑9 safety score. On 1,362 annotated frames from two London routes, the model achieves 72.2 % binary accuracy and a mean absolute error of 1.011, with feature attribution revealing the impact of lane separation, painted lanes, and construction sites on safety ratings. The system operates efficiently at the network edge, requiring only 2.49 s per frame for extraction and 0.4 ms for scoring, and supports continuous monitoring at a cost of 1.4 accuracy points for 1,000 riders over a 20‑second interval.