arXiv subjects

Chenye Yang

Publications and source records attributed to Chenye Yang.

5 recordsLinked to original sources

arXiv2026

Native Audio-Visual Alignment for Generation

Joint audio-video generation aims to synthesize temporally synchronized and semantically coherent visual-acoustic content. However, existing open-source methods mainly rely on either dual-tower designs with posterior alignment or fully unified tri-modal designs that mix textual context, audio and video in one shared space. The former weakens fine-grained audio-video co-evolution, while the latter couples semantic conditioning with low-level synchronization. To address these limitations, we propose NAVA, a Native Audio-Visual Alignment framework for joint audio-video generation. NAVA is built upon context-conditioned native audio-visual alignment: it first establishes audio-video correspondence in a dedicated interaction space, and then uses external context to condition the joint denoising process. Specifically, NAVA is instantiated with an Align-then-Fuse MMDiT architecture, which transitions from modality-aware audio-video alignment to modality-shared joint denoising. Furthermore, we introduce Timbre-in-Context Conditioning to associate reference timbre cues with corresponding speech spans to achieve controllable speech timbre. Experiments on Verse-Bench and Seed-TTS, together with a user study, demonstrate that NAVA achieves superior video quality, precise audio-visual synchronization, competitive audio quality, and stronger reference-timbre controllability using only 6.3B parameters.

cs.CV↗

arXiv2026

Efficient Preference Poisoning Attack on Offline RLHF

Offline Reinforcement Learning from Human Feedback (RLHF) pipelines such as Direct Preference Optimization (DPO) train on a pre-collected preference dataset, which makes them vulnerable to preference poisoning attack. We study label flip attacks against log-linear DPO. We first illustrate that flipping one preference label induces a parameter-independent shift in the DPO gradient. Using this key property, we can then convert the targeted poisoning problem into a structured binary sparse approximation problem. To solve this problem, we develop two attack methods: Binary-Aware Lattice Attack (BAL-A) and Binary Matching Pursuit Attack (BMP-A). BAL-A embeds the binary flip selection problem into a binary-aware lattice and applies Lenstra-Lenstra-Lov\'asz reduction and Babai's nearest plane algorithm; we provide sufficient conditions that enforce binary coefficients and recover the minimum-flip objective. BMP-A adapts binary matching pursuit to our non-normalized gradient dictionary and yields coherence-based recovery guarantees and robustness (impossibility) certificates for $K$-flip budgets. Experiments on synthetic dictionaries and the Stanford Human Preferences dataset validate the theory and highlight how dictionary geometry governs attack success.

cs.LG↗

arXiv2025

ColaVLA: Leveraging Cognitive Latent Reasoning for Hierarchical Parallel Trajectory Planning in Autonomous Driving

Autonomous driving requires generating safe and reliable trajectories from complex multimodal inputs. Traditional modular pipelines separate perception, prediction, and planning, while recent end-to-end (E2E) systems learn them jointly. Vision-language models (VLMs) further enrich this paradigm by introducing cross-modal priors and commonsense reasoning, yet current VLM-based planners face three key challenges: (i) a mismatch between discrete text reasoning and continuous control, (ii) high latency from autoregressive chain-of-thought decoding, and (iii) inefficient or non-causal planners that limit real-time deployment. We propose ColaVLA, a unified vision-language-action framework that transfers reasoning from text to a unified latent space and couples it with a hierarchical, parallel trajectory decoder. The Cognitive Latent Reasoner compresses scene understanding into compact, decision-oriented meta-action embeddings through ego-adaptive selection and only two VLM forward passes. The Hierarchical Parallel Planner then generates multi-scale, causality-consistent trajectories in a single forward pass. Together, these components preserve the generalization and interpretability of VLMs while enabling efficient, accurate and safe trajectory generation. Experiments on the nuScenes benchmark show that ColaVLA achieves state-of-the-art performance in both open-loop and closed-loop settings with favorable efficiency and robustness.

cs.CV↗

arXiv2017

A Review of Smart Materials in Tactile Actuators for Information Delivery

As the largest organ in the human body, the skin provides the important sensory channel for humans to receive external stimulations based on touch. By the information perceived through touch, people can feel and guess the properties of objects, like weight, temperature, textures, and motion, etc. In fact, those properties are nerve stimuli to our brain received by different kinds of receptors in the skin. Mechanical, electrical, and thermal stimuli can stimulate these receptors and cause different information to be conveyed through the nerves. Technologies for actuators to provide mechanical, electrical or thermal stimuli have been developed. These include static or vibrational actuation, electrostatic stimulation, focused ultrasound, and more. Smart materials, such as piezoelectric materials, carbon nanotubes, and shape memory alloys, play important roles in providing actuation for tactile sensation. This paper aims to review the background biological knowledge of human tactile sensing, to give an understanding of how we sense and interact with the world through the sense of touch, as well as the conventional and state-of-the-art technologies of tactile actuators for tactile feedback delivery.

physics.app-ph↗

arXiv2017

Smart materials and structures for energy harvesters

Vibrational energy harvesters capture mechanical energy from ambient vibrations and convert the mechanical energy into electrical energy to power wireless electronic systems. Challenges exist in the process of capturing mechanical energy from ambient vibrations. For example, resonant harvesters may be used to improve power output near their resonance, but their narrow bandwidth makes them less suitable for applications with varying vibrational frequencies. Higher operating frequencies can increase harvesters power output, but many vibrational sources are characterized by lower frequencies, such as human motions. This paper provides a thorough review of state of the art energy harvesters based on various energy sources such as solar, thermal, electromagnetic and mechanical energy, as well as smart materials including piezoelectric materials and carbon nanotubes. The paper will then focus on vibrational energy harvesters to review harvesters using typical transduction mechanisms and various techniques to address the challenges in capturing mechanical energy and delivering it to the transducers.

physics.app-ph↗