variational auto-encoder
PulseAugur coverage of variational auto-encoder — every cluster mentioning variational auto-encoder across labs, papers, and developer communities, ranked by signal.
7 天有情绪数据
-
Splatent framework enhances 3D Gaussian Splatting with diffusion latents for novel view synthesis
Researchers have introduced Splatent, a novel framework that enhances 3D Gaussian Splatting within the latent space of VAEs for improved novel view synthesis. Unlike previous methods that struggled with multi-view consi…
-
New LLM techniques and benchmarks advance 3D indoor scene generation
Researchers have developed new methods for generating 3D indoor scenes using AI, addressing challenges like spatial errors and data scarcity. One approach, SpatialGrammar, introduces a domain-specific language to repres…
-
VAE-Inf 框架将生成学习与假设检验相结合，用于不平衡分类
研究人员推出 VAE-Inf，这是一个新颖的两阶段框架，旨在解决机器学习中长期存在的不平衡分类挑战。该方法将深度表示学习与统计上可解释的假设检验相结合。第一阶段在多数类数据上训练变分自编码器，以建立参考分布，然后用它来构建全局高斯参考模型。第二阶段使用有限的少数样本微调编码器，创建一个判别式分类器，该分类器在不要求严格参数假设的情况下，对第一类错误提供精确控制。
-
AI学习肌肉驱动控制，实现逼真钢琴演奏
研究人员开发了一种新颖的数据驱动方法，用于控制基于物理的、肌肉驱动的手来演奏钢琴，具有卓越的灵活性。他们的方法结合了高频肌肉控制和低频潜在空间协调，使手能够演奏新的乐曲。该系统利用强化学习进行肌肉激活跟踪，并使用变分自编码器来抽象肌肉动力学，从而实现特定乐曲的协调策略。该方法在基于物理的灵巧钢琴演奏控制方面取得了最先进的性能，并生成了生理上合理的肌肉激活模式。
-
Researchers propose novel VAE reparameterization for non-trivial latent space topologies
Researchers have developed a novel method to generalize the reparameterization trick used in Variational Autoencoders (VAEs). This new technique allows VAEs to handle latent spaces with complex, non-trivial topologies, …
-
VDLF-Net advances few-shot visual learning with variational feature fusion
Researchers have developed VDLF-Net, a novel architecture for adaptive and few-shot visual learning. This model integrates a Variational Autoencoder (VAE) with a multi-scale Convolutional Neural Network (CNN) backbone. …
-
Tuna-2 model ditches vision encoders for direct pixel embeddings, achieving SOTA
Researchers have developed Tuna-2, a novel unified multimodal model that bypasses traditional vision encoders for visual understanding and generation. By directly processing pixel embeddings, Tuna-2 simplifies architect…
-
Study systematically assesses dimensionality reduction impact on clustering performance
A new study systematically evaluates how five different dimensionality reduction techniques affect the performance of four common clustering algorithms. Researchers found that the choice of dimensionality reduction meth…
-
ORSIFlow框架改进光学遥感显著目标检测
研究人员推出了一种新颖的光学遥感图像显著目标检测框架ORSIFlow。该方法将问题重新表述为确定性潜在流生成任务，在一个由变分自编码器派生的紧凑潜在空间内运行。ORSIFlow旨在通过引入显著特征判别器以实现全局语义理解，以及引入显著特征校准器以进行详细边界细化，来提高效率和准确性。实验表明，ORSIFlow在各种基准测试中取得了最先进的结果，并提高了效率。
-
CLVAE model enhances long-term customer revenue forecasting with flexible VAE approach
Researchers have introduced CLVAE, a novel variational autoencoder model designed for forecasting long-term customer revenue from sparse transaction data. This approach combines the structural robustness of traditional …
-
AI model detects retinal abnormalities without expert annotations
Researchers have developed a novel unsupervised anomaly detection framework for Optical Coherence Tomography (OCT) imaging, aiming to overcome the reliance on expert annotations for diagnosing retinal disorders. This ne…
-
MISTY motion planner achieves state-of-the-art autonomous driving with single-step inference
Researchers have developed MISTY, a novel generative motion planner designed for autonomous driving that achieves high throughput with single-step inference. Unlike existing diffusion-based planners that require iterati…
-
OpenAI unveils VAEs for improved representation learning and density estimation
OpenAI has published research on a Variational Autoencoder (VAE) that combines VAEs with autoregressive models like RNNs and PixelCNNs. This new VAE architecture allows for control over what the latent code learns, enab…