paper-with-me

Papers

Delving into Latent Spectral Biasing of Video VAEs for Superior Diffusability

2025-12-05 · Shizhan Liu, Xinran Deng, Zhuoyi Yang, Jiayan Teng, Xiaotao Gu, Jie Tang arxiv

Latent diffusion models pair VAEs with diffusion backbones, and the structure of VAE latents strongly influences the difficulty of diffusion training. However, existing video VAEs typically focus on reconstruction fidelity, overlooking latent structure. We present a statistical analysis of video VAE latent spaces and identify two spectral properties essential for diffusion training: a spatio-temporal frequency spectrum biased toward low frequencies, and a channel-wise eigenspectrum dominated by a few modes. To induce these properties, we propose two lightweight, backbone-agnostic regularizers: Local Correlation Regularization and Latent Masked Reconstruction. Experiments show that our Spectral-Structured VAE (SSVAE) achieves a $3\times$ speedup in text-to-video generation convergence and a 10\% gain in video reward, outperforming strong open-source VAEs. The code is available at https://github.com/zai-org/SSVAE.

📄 PDF Abstract BibTeX arXiv:2512.05394

Code (0)

등록된 구현이 없습니다.

Tasks

Text-to-Video Generation

Similar Papers 제목 키워드 기반

Delving into CLIP latent space for Video Anomaly Recognition

2023-10-04 · Luca Zanella, Benedetta Liberatori, Willi Menapace, Fabio Poiesi 외

We tackle the complex problem of detecting and recognising anomalies in surveillance videos at the frame level, utilising only video-level supervision. We introduce the novel method AnomalyCLIP, the first to combine Larg…

Anomaly DetectionMultiple Instance LearningVideo Anomaly Detection

Neutralizing Gender Bias in Word Embeddings with Latent Disentanglement and Counterfactual Generation

2020-11-01 · Findings of the Association for Computational Linguistics 2020 · Seungjae Shin, Kyungwoo Song, JoonHo Jang, Hyemi Kim 외

Recent research demonstrates that word embeddings, trained on the human-generated corpus, have strong gender biases in embedding spaces, and these biases can result in the discriminative results from the various downstre…

counterfactualDisentanglementWord Embeddings

Neutralizing Gender Bias in Word Embedding with Latent Disentanglement and Counterfactual Generation

2020-04-07 · Seungjae Shin, Kyungwoo Song, JoonHo Jang, Hyemi Kim 외

Recent research demonstrates that word embeddings, trained on the human-generated corpus, have strong gender biases in embedding spaces, and these biases can result in the discriminative results from the various downstre…

counterfactualDisentanglementSentiment AnalysisWord Embeddings

The Malignant Tail: Spectral Segregation of Label Noise in Over-Parameterized Networks

2026-03-02 · Zice Wang arxiv

While implicit regularization facilitates benign overfitting in low-noise regimes, recent theoretical work predicts a sharp phase transition to harmful overfitting as the noise-to-signal ratio increases. We experimentall…

Self-supervised debiasing using low rank regularization

2022-10-11 · CVPR 2024 1 · Geon Yeong Park, Chanyong Jung, Sangmin Lee, Jong Chul Ye 외

Spurious correlations can cause strong biases in deep neural networks, impairing generalization ability. While most existing debiasing methods require full supervision on either spurious attributes or target labels, trai…

Self-Supervised Learning