Live intelligence

Latest AI updates

3 updates
Research

Triplet Incremental Training Adapts General VLMs for Autonomous Driving

Researchers introduce a Triplet Incremental Training framework that instills 3D geometric awareness into general vision-language model backbones for autonomous driving without architectural overhauls. The method uses automatically constructed image-depth-QA triplets in a tri-stream pipeline: an unfrozen primary VLM for driving tasks, an auxiliary depth stream for implicit geometry learning, and a frozen twin VLM with logit-level KL distillation to prevent semantic drift. The depth branch is discarded at deployment, so the added geometric awareness incurs no inference overhead. Across four benchmarks, the approach reportedly outperforms state-of-the-art driving VLMs, rivals domain-specific experts in 3D perception, and retains general multimodal reasoning.

Vision-Language ModelsAutonomous Driving3D Perception
Research

Rethinking Radar 3D Perception: Representation Learning from Raw Spectra

The paper argues that raw radar spectra can support direct representation learning if the pretraining objective separates structured signal from noise. It proposes Leakage-aware Spectral MAE, a masked autoencoder for raw radar cubes with a corruption strategy inspired by cross-bin energy spreading from spectral leakage. It also treats Doppler as a primary spatial dimension in an x-y-D representation, encoding height as features. The authors report that this approach outperforms state-of-the-art radar-based detectors and point-cloud self-supervised learning counterparts without handcrafted reductions.

radar perceptionself-supervised learning3D perception
Research Editor selected

LiDAR-SHIELD Unifies Adversarial Testing for LiDAR Perception Models

Researchers introduce LiDAR-SHIELD, a Dockerized framework for assessing adversarial robustness in LiDAR-based 3D semantic segmentation and object detection. Models connect through three HTTP endpoints, reducing the need for model-specific environment reconstruction and integration. The framework brings multiple attack methods into a shared testing protocol, and its codebase is publicly available.

Why it matters: The framework offers autonomous-driving and robotics researchers a unified way to stress-test different LiDAR perception models for adversarial vulnerabilities.

LiDAR-SHIELDAdversarial robustness3D perception