CryoKRAQEN：面向冷冻电镜异质重建中量化嵌入网络的核正则化退火方法
CryoKRAQEN: Kernel-Regularized Annealing for Quantized Embedding Networks in Cryo-EM Heterogeneous Reconstruction
把三平面隐式表示与核引导的量化潜变量分配结合，做成无编码器的解码器框架，可迁移到其他需要从极噪、连续变化观测中恢复潜结构的逆问题（如动态场景重建、低剂量成像），其避免码本坍缩的退火式量化分配策略也值得借鉴。
原文依据 a decoder-only framework that integrates triplane implicit representations with kernel-guided latent assignment and quantized embeddings
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论冷冻电子显微镜（Cryo-EM）中的异质重建是理解大分子结构多样性的基础，但由于极端噪声、连续构象变化以及图像到结构映射的歧义性，该问题仍极具挑战。现有神经方法往往依赖编码器—解码器流程或固定码本，计算开销大或难以处理复杂的异质性。我们提出 CryoKRAQEN，一个仅含解码器的框架，将三平面隐式表示与核引导的潜变量分配及量化嵌入相结合，以提升稳定性与结构判别能力。该方法摆脱了对编码器的依赖，并缓解了训练过程中的坍缩，从而能够准确建模构象变化与组成变化。在多个 Cryo-EM 基准上，与最优的神经方法和经典方法相比，CryoKRAQEN 取得了有竞争力的性能、稳健的重建结果以及可解释的潜变量组织。
英文原文摘要 999 字符
Heterogeneous reconstruction in cryo-electron microscopy (Cryo-EM) is fundamental for understanding macromolecular structural diversity, yet remains challenging due to extreme noise, continuous conformational changes, and ambiguous image-to-structure mappings. Existing neural approaches often rely on encoder--decoder pipelines or fixed codebooks, which can be computationally demanding or struggle with complex heterogeneity. We propose CryoKRAQEN, a decoder-only framework that integrates triplane implicit representations with kernel-guided latent assignment and quantized embeddings to improve stability and structural discrimination. The method avoids encoder dependencies and mitigates collapse during training, enabling accurate modeling of both conformational and compositional variations. Across diverse Cryo-EM benchmarks, CryoKRAQEN delivers competitive performance, robust reconstructions, and interpretable latent organization compared to state-of-the-art neural and classical methods.