← Back to all news
arXiv Machine Learning By Haopeng Deng, Thomas He, Dayuan Wang

FLAT: Smoothing the Rugged Landscape for Learnable, Sample-Efficient Traffic Calibration

Read the original on arXiv Machine Learning →

FLAT is a method for calibrating microscopic traffic models that smooths the rugged parameter-error landscape using an eight‑dimensional behavioral fingerprint, enabling learnable objectives from only a few dozen simulation runs. It couples this smoothed objective with an annealed lower‑confidence‑bound acquisition strategy to allocate remaining simulation budget efficiently. Across six real‑world scenes, FLAT‑GP outperforms several baselines, achieving the lowest behavioral error with a matched simulation budget.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv AI
Sep 10

Closed-Loop Evaluation of Bird's-Eye-View Maps from Cross-View Transformers as Inputs to Behavior-Cloning Policies

The paper studies how Bird's‑Eye‑View (BEV) maps predicted by Cross‑View Transformers (CVT) can be used directly as inputs to a Behavior‑Cloning (BC) driving policy in the CARLA simulator. It introduces a six‑channel BEV representation and a Kernel Density Estimation (KDE) weighting scheme to focus learning on underrepresented maneuvers. Closed‑loop tests show that the KDE‑weighted model is the only predicted‑BEV agent to finish an episode without infractions, highlighting that global segmentation scores are poor proxies for driving performance and that prediction quality at critical geometries, especially the route channel, is key to reliable navigation.

By Felipe Carlos dos Santos, Eric Antonelo, Gustavo Claudio Karl Couto
llmsagentscomputer-visionrobotics
More like this →
arXiv AI
Jul 15

TerraZero: Procedural Driving Simulation for Zero-Demonstration Self-Play at Scale

arXiv:2607. 13028v1 Announce Type: cross Abstract: Training robust autonomous driving agents requires a simulator that is fast enough for reinforcement learning at scale, realistic enough to ground behavior in real-world map structure, and diverse enough to cover the safety-critical long tail that logged data rarely contains.

By Zhouchonghao Wu, Akshay Rangesh, Weixin Li, Wei-Jer Chang, Zachary Lee, Tim Wang, Wei Zhan
agentsreinforcement-learningroboticsbenchmarkssafety
More like this →
arXiv AI
Sep 2

CoLT-Drive: Counterfactual Long-Tail Benchmarking and Knowledge-Preserving Adaptation for Driving Affordance Prediction

The paper introduces CoLT-Drive, a 3,536-sample counterfactual long‑tail benchmark for evaluating decision‑level driving affordance prediction, which tests whether models can infer how rare objects affect an ego vehicle’s high‑level actions. It also proposes KPA, a knowledge‑preserving adaptation framework that combines structured prompting, expert merging, and a regime‑aware LoRA mixture‑of‑experts module to improve small VLMs on driving tasks. Experiments show KPA achieves 60.8% pair accuracy on CoLT‑Drive, outperforming the Qwen3‑VL‑2B baseline and LoRA SFT while keeping competitive in‑domain performance.

By Zhengxu Tang, Guofeng Cui, Ziyu Gong, Xiaozhou Zhang, Ruifeng Deng, Chengzhi Qi, Ke Chen, Sachin Patil, Tianjun Xiao, Langechuan Liu, Pichao Wang
llmsagentsroboticsfine-tuningbenchmarks
More like this →