Hao Zhou

Yihong Sun, Hao Zhou, Liangzhe Yuan et al. · deepmind

We explore a novel video creation experience, namely Video Creation by Demonstration. Given a demonstration video and a context image from a different scene, we generate a physically plausible video that continues naturally from the context image and carries out the action concepts from the demonstration. To enable this capability, we present $δ$-Diffusion, a self-supervised training approach that learns from unlabeled videos by conditional future frame prediction. Unlike most existing video generation controls that are based on explicit signals, we adopts the form of implicit latent control for maximal flexibility and expressiveness required by general videos. By leveraging a video foundation model with an appearance bottleneck design on top, we extract action latents from demonstration videos for conditioning the generation process with minimal appearance leakage. Empirically, $δ$-Diffusion outperforms related baselines in terms of both human preference and large-scale machine evaluations, and demonstrates potentials towards interactive world simulation. Sampled video generation results are available at https://delta-diffusion.github.io/.

Hao Zhou, Xiaobao Guo, Yuzhe Zhu et al.

Propelled by the breakthrough in deep generative models, audio-to-image generation has emerged as a pivotal cross-modal task that converts complex auditory signals into rich visual representations. However, previous works only focus on single-source audio inputs for image generation, ignoring the multi-source characteristic in natural auditory scenes, thus limiting the performance in generating comprehensive visual content. To bridge this gap, we propose a method called MACS to conduct multi-source audio-to-image generation. To our best knowledge, this is the first work that explicitly separates multi-source audio to capture the rich audio components before image generation. MACS is a two-stage method. In the first stage, multi-source audio inputs are separated by a weakly supervised method, where the audio and text labels are semantically aligned by casting into a common space using the large pre-trained CLAP model. We introduce a ranking loss to consider the contextual significance of the separated audio signals. In the second stage, effective image generation is achieved by mapping the separated audio signals to the generation condition using only a trainable adapter and a MLP layer. We preprocess the LLP dataset as the first full multi-source audio-to-image generation benchmark. The experiments are conducted on multi-source, mixed-source, and single-source audio-to-image generation tasks. The proposed MACS outperforms the current state-of-the-art methods in 17 out of the 21 evaluation indexes on all tasks and delivers superior visual quality.

Yuqi Yu, Hanbing Yan, Hongchao Guan et al.

In the Internet age, cyber-attacks occur frequently with complex types. Traffic generated by access activities can record website status and user request information, which brings a great opportunity for network attack detection. Among diverse network protocols, Hypertext Transfer Protocol (HTTP) is widely used in government, organizations and enterprises. In this work, we propose DeepHTTP, a semantics structure integration model utilizing Bidirectional Long Short-Term Memory (Bi-LSTM) with attention mechanism to model HTTP traffic as a natural language sequence. In addition to extracting traffic content information, we integrate structural information to enhance the generalization capabilities of the model. Moreover, the application of attention mechanism can assist in discovering critical parts of anomalous traffic and further mining attack patterns. Additionally, we demonstrate how to incrementally update the data set and retrain model so that it can be adapted to new anomalous traffic. Extensive experimental evaluations over large traffic data have illustrated that DeepHTTP has outstanding performance in traffic detection and pattern discovery.