Searcharxiv⌕ Search

arXiv subjects

Yizhi Chen

Publications and source records attributed to Yizhi Chen.

14 recordsLinked to original sources
arXiv2026

Global-Local Attention Decomposition for Terrain Encoding in Humanoid Perceptive Locomotion

Although reinforcement learning has significantly advanced humanoid locomotion, perceptive policies still struggle on sparse-foothold terrain and constrained environments. Success in these scenarios requires both broad terrain awareness and precise foothold selection, two perceptual roles that conventional encoders often entangle. To address this challenge, we propose Global-Local Attention Decomposition (GLAD) for terrain encoding in humanoid locomotion. Realized by a coarse-to-fine encoder over a robot-centric elevation map, GLAD explicitly separates these objectives: a global attention branch uses attention pooling to summarize the surrounding terrain context, while a local attention branch sparsifies the local features by terrain saliency and applies state-conditioned attention to encode precise foothold-relevant geometry. This explicit attention decomposition prevents the dilution of fine-grained spatial cues while reducing training overhead. Experiments demonstrate that GLAD enables reliable locomotion over challenging gaps, stepping stones, and stairs. Furthermore, the learned policy exhibits emergent terrain-responsive behaviors, autonomously following narrow paths and avoiding obstacles under forward-velocity commands alone, without explicit navigation planners. In real-world deployment on a Unitree G1 humanoid robot using onboard LiDAR, the proposed method achieves robust zero-shot sim-to-real transfer across diverse sparse-foothold and obstacle-rich domains.

cs.RO↗
arXiv2026

P3: Probabilistic Policy Propagation for Stable VAE-Based Robot Learning

Variational Autoencoders are widely used to encode high-dimensional and noisy observations in robotics. However, their stochastic latent creates a mismatch with Proximal Policy Optimization (PPO): an effective policy marginalizes over the latent distribution, whereas former implementations estimate its probability ratio and KL divergence using only one latent sample. We identify a fundamental but overlooked theoretical cause: naive single-sample approximations in stochastic latent space induce significant variance and bias in the surrogate loss. To address this, we introduce P^3 (Probabilistic Policy Propagation), a distribution-aware optimization framework for VAE-based policies. $P^3$ couples moment-based probabilistic method for stable and efficient learning with sampling-based calibration for robust policy behavior under latent uncertainty. In our experiments, P^3 boosts data efficiency from 64.6% to >96%, reduces convergence steps by >20%. Furthermore, P^3 is evaluated on challenging humanoid parkour tasks and shows an effective foundation for VAE-based PPO. Code is available at https://github.com/ylyem9x/P3_Open.

cs.RO↗
arXiv2026

Localized crystallization of Ce:YIG thin films on Si using CO2 laser annealing for integrated nonreciprocal photonic device applications

Laser annealing (LA) technique has emerged as an effective method for localized crystallization of magneto-optical (MO) garnet thin films on semiconductor substrates. However, no studies have explored the crystallization and magneto-optical (MO) properties of cerium-substituted yttrium iron garnet (Ce:YIG, Ce1Y2Fe5O12) thin films for integrated photonic device applications using LA technique. In this study, we provide a comprehensive investigation into the laser annealing of Ce:YIG films deposited on SiO2 substrates and silicon nitride photonic waveguides for integrated nonreciprocal photonic device applications. Garnet phase was successfully observed in films grown on SiO2 substrates, and SiN waveguides with laser annealing of sputtered Ce:YIG films on top of a laser annealed Y3Fe5O12 seed layer. The magneto-optical (MO) properties of Ce:YIG films on oxidized Si substrates were found to be comparable to those prepared by rapid thermal annealing (RTA). A Mach-Zehnder Interferometer (MZI) type optical isolator based on Ce:YIG film on SiN was fabricated, exhibiting a saturation Faraday rotation of -2317.7 deg/cm and propagation loss of 188.2 dB/cm. Isolation ratio of 27.1 dB and insertion loss of 10.1 dB were achieved at 1552.7 nm wavelength.

cond-mat.mtrl-sci↗
arXiv2026

UniLab: A Heterogeneous Architecture for Robot RL Beyond GPU-Dominant Paradigms

Simulation-based RL for contemporary robot control is increasingly organized around GPU-resident simulation: physics, rollout collection, and learning are placed on a single GPU-centric execution path. This paradigm has greatly improved training speed, but it has also encouraged a default assumption that efficient training requires physics to reside on the GPU. We revisit this assumption. Our view is that, in simulation-dominated robot control, the essential question is not which processor runs physics, but whether simulation throughput, policy learning, and runtime synchronization form an efficient end-to-end loop. We present UniLab, a heterogeneous CPU-simulation / GPU-learning architecture that decouples CPU-parallel simulation from GPU policy updates through a unified runtime for data movement, buffering, and synchronization. UniLab is implemented as a complete and extensible training system using MuJoCoUni and MotrixSim CPU-batched physics backends, supporting PPO, FastSAC, FlashSAC, and APPO. On representative simulation-based robot control tasks, UniLab improves end-to-end training efficiency by 3--10$\times$ under the same hardware configuration, while reducing dependence on the NVIDIA CUDA-based software stack and supporting cross-platform execution on the Apple macOS platform and the AMD ROCm and Intel XPU accelerator backends. These results show that GPU simulation is an effective path to efficient training, but not a necessary one, broadening the practical system choices available for robot RL training. Project page: https://unilabsim.github.io.

cs.RO↗
arXiv2026

GeoAlign: Beyond Semantics with State-Guided Spatial Alignment in VLA Models

Current Vision--Language--Action (VLA) models often optimize for semantic grounding, whereas executable manipulation requires geometry-aware spatial alignment and dynamic affordance selection. We introduce GeoAlign, a state-guided spatial alignment architecture for VLA policy learning. GeoAlign post-trains an RGB geometry branch with robot-domain RGB-D supervision, yielding RGB-derived Geometry-Enhanced Post-Trained (GEP) features for policy rollout. The robot's proprioceptive state queries the GEP feature grid, producing compact, phase-dependent geometry tokens for action prediction. GeoAlign achieves 99.0% on LIBERO, 85.3% across three SimplerEnv-Fractal tasks, and 78.8% on eight geometry-critical real-world ALOHA tasks, with ablations confirming the value of geometry post-training and proprioceptive-state-guided querying.

cs.RO↗
arXiv2026

HierKick: Hierarchical Reinforcement Learning for Vision-Guided Soccer Robot Control

Controlling soccer robots involves multi-time-scale decision-making, which requires balancing long-term tactical planning and short-term motion execution. Traditional end-to-end reinforcement learning (RL) methods face challenges in complex dynamic environments. This paper proposes HierKick, a vision-guided soccer robot control framework based on dual-frequency hierarchical RL. The framework adopts a hierarchical control architecture featuring a 5 Hz high-level policy that integrates YOLOv8 for real-time detection and selects tasks via a coach model, and a pre-trained 50 Hz low-level controller for precise joint control. Through this architecture, the framework achieves the four steps of approaching, aligning, dribbling, and kicking. Experimental results show that the success rates of this framework are 95.2\% in IsaacGym, 89.8\% in Mujoco, and 80\% in the real world. HierKick provides an effective hierarchical paradigm for robot control in complex environments, extendable to multi-time-scale tasks, with its modular design and skill reuse offering a new path for intelligent robot control.

cs.RO↗
arXiv2026

'1'-bit Count-based Sorting Unit to Reduce Link Power in DNN Accelerators

Interconnect power consumption remains a bottleneck in Deep Neural Network (DNN) accelerators. While ordering data based on '1'-bit counts can mitigate this via reduced switching activity, practical hardware sorting implementations remain underexplored. This work proposes the hardware implementation of a comparison-free sorting unit optimized for Convolutional Neural Networks (CNN). By leveraging approximate computing to group population counts into coarse-grained buckets, our design achieves hardware area reductions while preserving the link power benefits of data reordering. Our approximate sorting unit achieves up to 35.4% area reduction while maintaining 19.50\% BT reduction compared to 20.42% of precise implementation.

cs.AR↗
arXiv2026

Late Breaking Results: Quamba-SE: Soft-edge Quantizer for Activations in State Space Models

We propose Quamba-SE, a soft-edge quantizer for State Space Model (SSM) activation quantization. Unlike existing methods, using standard INT8 operation, Quamba-SE employs three adaptive scales: high-precision for small values, standard scale for normal values, and low-precision for outliers. This preserves outlier information instead of hard clipping, while maintaining precision for other values. We evaluate on Mamba- 130M across 6 zero-shot benchmarks. Results show that Quamba- SE consistently outperforms Quamba, achieving up to +2.68% on individual benchmarks and up to +0.83% improvement in the average accuracy of 6 datasets.

cs.LG↗
arXiv2025

Bit Transition Reduction by Data Transmission Ordering in NoC-based DNN Accelerator

As Deep Neural Networks (DNN) are becoming essential, Network-on-Chip (NoC)-based DNN accelerators gained increasing popularity. To save link power in NoC, many researchers focus on reducing the Bit Transition (BT). We propose '1'-bit count-based ordering method to reduce BT for DNN workloads. We provide a mathematical proof of the efficacy of proposed ordering. We evaluate our method through experiments without NoC and with NoC. Without NoC, our proposed ordering method achieves up to 20.38% BT reduction for floating-point-32 data and 55.71% for fixed-point-8 data, respectively. We propose two data ordering methods, affiliated-ordering and separated-ordering to process weight and input jointly or individually and apply them to run full DNNs in NoC-based DNN accelerator. We evaluate our approaches under various configurations, including different DNN models such as LeNet and DarkNet, various NoC sizes with different numbers of memory controllers, random weights and trained weights, and different data precision. Our approach efficiently reduces the link power by achieving up to 32.01% BT reduction for floating-point-32 data and 40.85% BT reduction for fixed-point-8 data.

cs.AR↗
arXiv2025

LLM4GNAS: A Large Language Model Based Toolkit for Graph Neural Architecture Search

Graph Neural Architecture Search (GNAS) facilitates the automatic design of Graph Neural Networks (GNNs) tailored to specific downstream graph learning tasks. However, existing GNAS approaches often require manual adaptation to new graph search spaces, necessitating substantial code optimization and domain-specific knowledge. To address this challenge, we present LLM4GNAS, a toolkit for GNAS that leverages the generative capabilities of Large Language Models (LLMs). LLM4GNAS includes an algorithm library for graph neural architecture search algorithms based on LLMs, enabling the adaptation of GNAS methods to new search spaces through the modification of LLM prompts. This approach reduces the need for manual intervention in algorithm adaptation and code modification. The LLM4GNAS toolkit is extensible and robust, incorporating LLM-enhanced graph feature engineering, LLM-enhanced graph neural architecture search, and LLM-enhanced hyperparameter optimization. Experimental results indicate that LLM4GNAS outperforms existing GNAS methods on tasks involving both homogeneous and heterogeneous graphs.

cs.LG↗
arXiv2024

Travel Time Based Task Mapping for NoC-Based DNN Accelerator

Network-on-Chip (NoC) based architectures are recently proposed to accelerate deep neural networks in specialized hardware. Given that the hardware configuration is fixed post-manufacture, proper task mapping attracts researchers' interest. We propose a travel time-based task mapping method that allocates uneven counts of tasks across different Processing Elements (PEs). This approach utilizes the travel time recorded in the sampling window and implicitly makes use of static NoC architecture information and dynamic NoC congestion status. Furthermore, we examine the effectiveness of our method under various configurations, including different mapping iterations, flit sizes, and NoC architecture. Our method achieves up to 12.1% improvement compared with even mapping and static distance mapping for one layer. For a complete NN example, our method achieves 10.37% and 13.75% overall improvements to row-major mapping and distance-based mapping, respectively. While ideal travel time-based mapping (post-run) achieves 10.37% overall improvements to row-major mapping, we adopt a sampling window to efficiently map tasks during the running, achieving 8.17% (sampling window 10) improvement.

cs.AR↗
arXiv2024

Probing the shape of the brown dwarf desert around main-sequence A-F-G-type stars using post-common-envelope WD$-$BD binaries

Brown dwarfs (BDs) possessing masses within the range $40-60 M_{\rm Jup}$ are rare around solar-type main-sequence (MS) stars, which gives rise to the brown dwarf desert (BDD). One caveat associated with previous studies of BDD is the relatively limited sample size of MS$-$BD binaries with accurately determined BD masses. We aim to produce a large sample of brown dwarf companions with precisely determined mass around main-sequence A-F-G type stars using observations of post common-envelope white dwarf (WD)$-$BD binaries. We employ the rapid binary evolution code COMPAS to deduce the properties of MS$-$BD binary progenitors from post common-envelope WD$-$BD binaries. This method supplements the directly observed MS$-$BD binary sample, enriching the data available for analyzing BDD around main-sequence A-F-G type stars. Our study opens a new window for studying the shape of BDD around A-F-G type main-sequence stars in the short period regime. We find tentative evidence that the `driest' part of BDD around A-F-G type stars may extend into an orbital period of several hundred days, albeit with a small sample size. More post common-envelope WD$-$BD binaries detected in the future will advance our understanding of the BDD around A-F-G type stars.

astro-ph.SR↗
arXiv2024

An Overview of the Development of Stereotactic Body Radiation Therapy

Stereotactic body radiation therapy (SBRT) refers to focusing high-energy rays in three-dimensional space on the tumor lesion area, reducing the dose received by surrounding normal tissues, which can effectively improve the local control rate of the tumor and reduce the probability of complications. With the comprehensive development of medical imaging, radiation biology and other disciplines, this less-fractional, high-dose radiotherapy method has been increasingly developed and applied in clinical practice. The background, radio-biological basis, key technologies and main equipment of SBRT are discussed, and its future development direction is prospected.

physics.med-ph↗