arXiv subjects

Haonan Wei

Publications and source records attributed to Haonan Wei.

3 recordsLinked to original sources

arXiv2026

A one-step generation model with a Single-Layer Transformer: Layer number re-distillation of FreeFlow

Currently, Flow matching methods aim to compress the iterative generation process of diffusion models into a few or even a single step, with MeanFlow and FreeFlow being representative achievements of one-step generation based on Ordinary Differential Equations (ODEs). We observe that the 28-layer Transformer architecture of FreeFlow can be characterized as an Euler discretization scheme for an ODE along the depth axis, where the layer index serves as the discrete time step. Therefore, we distill the number of layers of the FreeFlow model, following the same derivation logic as FreeFlow, and propose SLT (Single-Layer Transformer), which uses a single shared DiT block to approximate the depth-wise feature evolution of the 28-layer teacher. During training, it matches the teacher's intermediate features at several depth patches, fuses those patch-level representations, and simultaneously aligns the teacher's final velocity prediction. Through distillation training, we compress the 28 independent Transformer Blocks of the teacher model DiT-XL/2 into a single Transformer Block, reducing the parameter count from 675M to 4.3M. Furthermore, leveraging its minimal parameters and rapid sampling speed, SLT can screen more candidate points in the noise space within the same timeframe, thereby selecting higher-quality initial points for the teacher model FreeFlow and ultimately enhancing the quality of generated images. Experimental results demonstrate that within a time budget comparable to two random samplings of the teacher model, our method performs over 100 noise screenings and produces a high-quality sample through the teacher model using the selected points. Quality fluctuations caused by low-quality initial noise under a limited number of FreeFlow sampling calls are effectively avoided, substantially improving the stability and average generation quality of one-step generation.

cs.CV↗

arXiv2025

MFI-ResNet: Efficient ResNet Architecture Optimization via MeanFlow Compression and Selective Incubation

ResNet has achieved tremendous success in computer vision through its residual connection mechanism. ResNet can be viewed as a discretized form of ordinary differential equations (ODEs). From this perspective, the multiple residual blocks within a single ResNet stage essentially perform multi-step discrete iterations of the feature transformation for that stage. The recently proposed flow matching model, MeanFlow, enables one-step generative modeling by learning the mean velocity field to transform distributions. Inspired by this, we propose MeanFlow-Incubated ResNet (MFI-ResNet), which employs a compression-expansion strategy to jointly improve parameter efficiency and discriminative performance. In the compression phase, we simplify the multi-layer structure within each ResNet stage to one or two MeanFlow modules to construct a lightweight meta model. In the expansion phase, we apply a selective incubation strategy to the first three stages, expanding them to match the residual block configuration of the baseline ResNet model, while keeping the last stage in MeanFlow form, and fine-tune the incubated model. Experimental results show that on CIFAR-10 and CIFAR-100 datasets, MFI-ResNet achieves remarkable parameter efficiency, reducing parameters by 46.28% and 45.59% compared to ResNet-50, while still improving accuracy by 0.23% and 0.17%, respectively. This demonstrates that generative flow-fields can effectively characterize the feature transformation process in ResNet, providing a new perspective for understanding the relationship between generative modeling and discriminative learning.

cs.CV↗

arXiv2024

Vertical Emission of Blue Light from a Symmetry Breaking Plasmonic Nanocavity-Emitter System Supporting Bound States in the Continuum

The concept of photonic bound states in the continuum (BICs), introduced in structured metallic surface cavities, provides a crucial mechanism for designing plasmonic open-resonant cavities with high quality (high-Q) factors, making significant advances in plasmonic nanophotonics. However, the two major bottlenecks for plasmonic nanocavities: enhancing emission and big beam divergence for quantum emitters, due to the strong intrinsic Ohmic losses of metals. Here, we propose and realize a {\sigma}h symmetry-breaking plasmonic honeycomb nanocavities (PHC) that support quasi-BIC resonance modes with high-Q factors. Our anodic oxidation-engineered strategy breaks out-of-plane symmetry while preserving in-plane symmetry, enabling the PHC to exhibit collective plasmonic lattice resonances (PLR) couplings and achieve Q-factors exceeding 106. Experimentally, we couple perovskite quantum dots (PQDs) to the PHC, demonstrating effective tuning of their emission properties and beam quality in the blue spectral region, achieving a 32-fold emission enhancement by suppress Ohmic loss and the life time of quantum emitters, simultaneously realize vertical emission in the 2.556 - 2.638 eV region, with a far-field hexagonal beam shape and a full width at half maximum of 12.6 degree under optimal coupling conditions. Furthermore, we demonstrate topological band inversion characterized by Zak phase transitions by continuously tuning the system parameters, confirming that the PHC supports topologically non-trivial q-BIC due to PLR coupling. The PHC presents itself as a promising next-generation, high-brightness nanoscale light source matrix, which can be directly scaled up to cover a wide wavelength range from UV to IR.

physics.optics↗