Music generation aims to create music segments that align with human aesthetics based on diverse conditions. Despite advancements in generating music from specific textual descriptions (e.g., style, genre, instruments), the practical application is still hindered by ordinary users' limited expertise to write accurate prompts. To bridge this application gap, this paper introduces MusFlow, a novel multimodal music generation model using Conditional Flow Matching (CFM). We employ multiple Multi-Layer Perceptrons to align multimodal conditions into the audio's CLAP embedding space. CFM is trained to reconstruct the compressed Mel-spectrogram in the VAE latent space guided by aligned feature embedding. MusFlow can generate music from images, story texts, and music captions. To collect data for model training, inspired by multi-agent collaboration, we construct an intelligent annotation workflow centered around a fine-tuned Qwen2-VL model. Using this workflow, we build a new multimodal music dataset, MMusSet, with each sample containing a quadruple of image, story text, music caption, and music piece. We conduct four sets of experiments: image-to-music, story-to-music, caption-to-music, and multimodal music generation. Experimental results demonstrate that MusFlow can generate high-quality music pieces from multimoal conditions. We hope this work can advance the application of music generation in multimedia field, making music creation more accessible. Our generated samples are available at https://anonymous22356.github.io/musflow.github.io/

Since the advent of cloud computing, Elastic Computing (EC) has become the standard architecture for resource allocation and scheduling. EC typically allocates computing resources based on predefined specifications, such as virtual machine or container flavors. However, these flavors are often constrained by fixed CPU-to-memory ratios, which frequently fail to match the actual resource needs of applications. As a result, cloud providers experience high resource allocation rates nearing saturation ($> $80%) but with low utilization ($< $25%). This study introduces Flexible Computing (FC), a novel approach to resource allocation and scheduling. Unlike EC, FC allocates resources based on an application resource usage profile, derived from the historical resource consumption of workloads, rather than relying on fixed specifications. Additionally, FC incorporates a real-time performance degradation detection mechanism to address performance issues caused by the noisy-neighbor effect when colocated workloads interfere with each other. FC dynamically adjusts resource allocation according to actual usage, ensuring that application performance meets Service Level Agreements (SLAs), while preventing resource waste and performance degradation from improper resource over-commitment. Large-scale experimental validations conducted on the FC architecture within Huawei Cloud data centers demonstrate that, compared to EC, FC can reduce computing resource consumption by over 33% while managing the same workloads. Furthermore, FC's real-time performance degradation detection model achieves a prediction error of less than 5% across various testing environments, highlighting its commercial viability.

Nowadays, the management of resource contention in shared cloud remains a pending problem. The evolution and deployment of new application paradigms (e.g., deep learning training and microservices) and custom hardware (e.g., graphics processing unit (GPU) and tensor processing unit (TPU)) have posed new challenges in resource management system design. Current solutions tend to trade cluster efficiency for guaranteed application performance, e.g., resource over-allocation, leaving a lot of resources underutilized. Overcoming this dilemma is not easy, because different components across the software stack are involved. Nevertheless, massive efforts have been devoted to seeking effective performance isolation and highly efficient resource scheduling. The goal of this paper is to systematically cover related aspects to deliver the techniques from the coordination perspective, and to identify the corresponding trends they indicate. Briefly, four topics are involved. First, isolation mechanisms deployed at different levels (micro-architecture, system, and virtualization levels) are reviewed, including GPU multitasking methods. Second, resource scheduling techniques within an individual machine and at the cluster level are investigated, respectively. Particularly, GPU scheduling for deep learning applications is described in detail. Third, adaptive resource management including the latest microservice-related research is thoroughly explored. Finally, future research directions are discussed in the light of advanced work. We hope that this review paper will help researchers establish a global view of the landscape of resource management techniques in shared cloud, and see technology trends more clearly.

Deep learning often requires large amounts of data from different institutions. Federated learning, as a distributed training framework, enables multiple participants to collaboratively train models without collecting data together and hence protecting data privacy, but the datasets from different institutions usually bring the problem of domain shift, which affects the performance of the model. When addressing domain shift, previous works often use a single global model to share parameters. Therefore, we propose a novel method to train multiple public models with different structures under the federated framework to improve the reliability and robustness of the public models. And each participant keeps its own domain-tuned private model, the private model does not share parameters with other participants. We use Mixture of Experts (MoE) domain adaptation to dynamically combine different public models and private model, which utilizes the similarity between different datasets to update the parameters of the public models. We apply the proposed method to the multi-site Magnetic resonance imaging (MRI) end-to-end classification, and the experiments demonstrate its effectiveness.

Modern data analytics typically run tasks on statically reserved resources (e.g., CPU and memory), which is prone to over-provision to guarantee the Quality of Service (QoS), leading to a large amount of resource time fragments. As a result, the resource utilization of a data analytics cluster is severely under-utilized. Workload co-location on shared resources has been substantially studied, but they are unaware the sizes of resource time fragments, making them hard to improve the resource utilization and guarantee QoS at the same time. In this paper, we propose Para, an event-driven scheduling mechanism, to harvest the CPU time fragments in co-located big data analytic workloads. Para innovates three techniques: 1) identifying the Idle CPU Time Window (ICTW) associated with each CPU core by capturing the task-switch event; 2) designing a runtime communication mechanism between each task execution of a workload and the underlying resource management system; 3) designing a pull-based scheduler to schedule a workload to run in the ICTW of another workload. We implement Para based on Apache Mesos and Spark. And the experimental results show that Para improves the CPU utilization by 44% and 30% on average relative to the original Mesos and enhanced Mesos under Spark's dynamic mode (MSDM), respectively. Moreover, Para increases the averaged task throughput of Mesos and MSDM by 4.8x and 1.7x, respectively, while guaranteeing the execution time of the primary applications.

Bone image landmark detection plays a vital role in orthopedic diseases diagnosis and analysis. As a promising approach, the data-driven convolutional neural network has been widely applied in landmarks detection. However, the model’s performance relies heavily on numerous high-quality labeled data, which are difficult to access. A practical solution for this issue in bone images landmark detection has not been provided, so we propose a framework of semi-supervised bone images landmark detection encouraged by the recent success of the self-ensembling method. The student and teacher networks share the same structure in our framework, where unlabeled data are utilized by enforcing the prediction consistency of two networks under different perturbations to enhance the network’s performance. The two networks use a mutual learning approach to improve the efficiency of using unlabeled data. In addition, we also propose a transformative perturbation scheme to prevent model overfitting during training and enhance the generalization capability of the model on unseen data. Experimental results on an in-house hip dataset collected from the actual process of clinical diagnosis and a publicly available dataset demonstrate that our framework has excellent performance compared with other semi-supervised methods.

With the development of cloud server, reversible data hiding in encrypted domain has received widespread attention for management of encrypted media. Most of existing reversible data hiding methods are based on stream ciphers, which are mainly concerned with data storage security. While homomorphic ciphers focus on data processing security, which is popular in cloud and other third-party platforms. Reversible data hiding with homomorphic cryptography has been an active topic in the research filed. In this paper, we propose a new lossless data hiding method based on homomorphic cryptosystem. After a content owner generates encrypted media with homomorphic ciphers, a data hider embeds secret data by establishing a mapping between secret bits and losslessly modifying encrypted media. At the receiver side, the embedded data is extracted from encrypted domain with a little auxiliary message. And original media is recovered by directly decrypting marked encrypted media. During the embedding process, the modification of encrypted media does not change the corresponding original media owing to homomorphic and probabilistic properties. Thus, no distortion is introduced in the whole procedures of data embedding. And directly decrypted media containing embedded data is the same as the original media. Meanwhile, the proposed method achieves a high embedding rate through efficient mapping and skillful use of expanded pixel values. The experimental results show that compared with state-of-the-art methods, the proposed method has higher embedding rate without distortion.