arXiv subjects

Jianguo Yu

Publications and source records attributed to Jianguo Yu.

4 recordsLinked to original sources

arXiv2026

S-squared-VLA: Decoupling Semantic and Spatial Streams in Vision-Language-Action Models for Autonomous Driving

Vision-Language Models (VLMs) have demonstrated remarkable potential for high-level reasoning in autonomous driving, yet they fundamentally struggle to generate precise, low-level control actions. This limitation is rooted in a semantic-physical gap caused by the inherent mismatch between discrete language tokens and continuous trajectory planning. While Vision-Language-Action (VLA) architectures attempt to bridge this gap by unifying perception and control into a single policy, this entanglement creates a new bottleneck. Standard VLAs experience a severe spatial representation collapse, which irreversibly degrades the fine-grained spatial and geometric priors essential for safe, boundary-aware navigation. To address this limitation, we propose the S-squared-VLA, which explicitly decouples the semantic and spatial streams in Vision-Language-Action models. The semantic stream leverages hierarchical bridging to extract multi-scale VLM features for robust intent reasoning. In parallel, an independent spatial stream bypasses the autoregressive language bottleneck, directly preserving uncompressed spatial features from the visual encoder. By integrating auxiliary perception supervision, this stream explicitly equips the model with rich spatial and geometric priors. Finally, a dual-stream planning adapter fuses high-level semantic intent with precise spatial constraints via cascaded attention mechanisms. Evaluations on the NAVSIM closed-loop benchmark show that S-squared-VLA achieves a Predictive Driver Model Score (PDMS) of 87.1, establishing a new state-of-the-art for VLA models under a purely supervised fine-tuning (SFT) setting. By mitigating the spatial representation collapse of traditional VLMs, our framework significantly outperforms baselines, achieving the highest No Collision (NC) rate of 98.4 among all evaluated methods.

cs.RO↗

arXiv2026

D$^3$-MoE:Dual Disentangled Diffusion Mixture-of-Experts for Style-Controllable End-to-End Autonomous Driving

Traditional end-to-end autonomous driving frameworks frequently suffer from the "style-averaging" dilemma when trained on high-variance human demonstrations, yielding homogenized, style-uncontrollable, and even kinematically unsafe policies. To overcome this limitation, we present D$^3$-MoE (Dual Disentangled Diffusion Mixture-of-Experts), which disentangles trajectory modeling along two complementary axes. On the behavioral axis, generation is decoupled from selection: a style-conditioned diffusion process synthesizes multi-style candidate trajectories in parallel within a single scene, allowing a downstream module to select the optimal trajectory based on user preference or an evaluation score. On the physical axis, decoupled longitudinal and lateral routers activate their respective experts during inference time, trained without manual labels using self-supervised targets from orthogonal ground-truth kinematics. These activated experts, architected as Diffusion Transformers (DiT) and equipped with style-conditioned AdaLN and asymmetric lateral-fusion cross-attention, independently predict their corresponding physical state before being reassembled into a unified, kinematically coherent trajectory. Extensive evaluations on the challenging NAVSIM benchmark demonstrate that D$^3$-MoE achieves state-of-the-art planning performance, reaching 88.2 PDMS and 84.3 EPDMS by default. Moreover, our Best-of-Three ensemble strategy effectively broadens the multi-modal solution space, raising performance to 91.3 PDMS and 87.5 EPDMS. Both quantitative and qualitative analyses jointly confirm the framework's advantages in planning quality and style controllability.

cs.RO↗

arXiv2021

Analyzing cases of significant nondynamic correlation with DFT using the Atomic Populations of Effectively Localized Electrons

Multireference effects are associated with degeneracies and near-degeneracies of the ground state and are critical to a variety of systems. Most approximate functionals of density functional theory (DFT) fail to properly describe these effects. A number of diagnostics have been proposed that allow to estimate the reliability of a given single-reference solution in this respect. Some of these diagnostics however lack size-consistency, while remaining computationally expensive. In this work we use the DFT method of determining atomic populations of effectively localized electrons (APELE) as a novel diagnostic in this vein. It is compared with several existing diagnostics of nondynamic correlation on select exemplary systems. We show that the APELE method is on average in good agreement with the existing diagnostics, while being both size-consistent and less costly. It becomes particularly informative in cases involving bond stretching or bond breaking. The APELE method is applied next to organic diradicals like the bis-acridine dimer and the p-quinodimethane molecule which possess unusually high nonlinear optical response, and to the reaction of ethylene addition to Ni dithiolene, where our results shed some more light on how the oxidation state of the Ni center may change when going from the initial reactant to the product.

quant-ph↗

arXiv2019

Speech Recognition with no speech or with noisy speech

The performance of automatic speech recognition systems(ASR) degrades in the presence of noisy speech. This paper demonstrates that using electroencephalography (EEG) can help automatic speech recognition systems overcome performance loss in the presence of noise. The paper also shows that distillation training of automatic speech recognition systems using EEG features will increase their performance. Finally, we demonstrate the ability to recognize words from EEG with no speech signal on a limited English vocabulary with high accuracy.

cs.LG↗