Pre-trained text-to-image (T2I) diffusion models produce high-quality images from random noise guided by textual prompts, but classifier-free guidance (CFG) often yields suboptimal samples. Retraining diffusion models can enhance performance, though the associated computational cost is often prohibitive. To address this, researchers have proposed modifying the CFG step through perturbation techniques for cost-effective sampling improvements. For instance, Smoothed Energy Guidance (SEG) enhances image quality by reducing the curvature of the energy landscape in self-attention layers. However, SEG's indiscriminate curvature reduction can suppress fine details and cause prompt misalignment. We introduce Orthogonal Smoothed Energy Guidance (OSEG), which selectively smooths the orthogonal components of token embeddings to preserve critical details while maintaining prompt alignment. Theoretical analysis justifies OSEG's effectiveness, supported by empirical comparisons with recent methods. By plugging into text-to-video (T2V) models, OSEG consistently improves visual and temporal coherence. Finally, we present extensive metric comparisons to demonstrate the efficiency of OSEG for both image and video domains. Project homepage: https://osegguidance.github.io/

Microservice architectures have become widely adopted due to their scalability and flexibility. These applications are commonly deployed in edge-cloud environments. However, in such environments, unexpected peak loads on microservices often lead to violations of service level objectives (SLOs) due to limited compute resources and bandwidth. Moreover, resource utilization of nodes and network states are constantly changing, making it essential to reschedule microservices to ensure SLO compliance. To minimize downtime during rescheduling, the microservice temporarily occupies additional resources until the process is complete. Therefore, frequent rescheduling can result in inefficient resource utilization. In this paper, we introduce SUPLEC, a new scheduling technique that optimally schedules container-based microservices and minimizes the number of reschedules to handle unexpected peak loads in Edge-Cloud environments. SUPLEC consists of two components: Placer and Rescheduler. The Placer identifies frequently interacting microservices, clusters them into groups, and places them on nodes using a worst-fit approach. The Rescheduler analyzes the response times of recent requests to determine whether rescheduling should be performed and, if so, it reschedules a microservice to a better node. Experiment results show that the proposed strategy reduces the average response time by up to 68.5 % under dynamic cluster conditions and loads and reduces the number of reschedules by up to 80 %.

The extensive capability of sensors let sensors autonomously collect various information on smart objects and store them to the spatial database through the wireless sensor networks. Based on spatial database, location-dependent information services (LDISs) can supply the resources according to the user locations. In LDISs, nearest-neighbour queries which return the closest object around the query location is recognised as the key component for searching the easily accessible services in smart spaces. However, existing works only consider the Euclidean distance. Thus, they have limitations to provide user-centric services that require the consideration for not only the distance but also the status of smart objects. Motivated by the issues of nearest-neighbour queries, this study proposes the new type of query called specified nearest-neighbour (SNN). SNN query considers the status and the locations of smart objects. For the SNN, the authors suggest a novel signature-based R-tree (SR-tree) index structure that handles non-spatial information of objects efficiently. Further, the authors propose an SNN query processing technique. Finally, they evaluate the performance of the proposed algorithm in various circumstances. Performance results indicate that SNN algorithm with SR-tree outperforms the existing works in terms of computational cost and disk input/output (I/O).

Fault tolerance is essential to the further development of desktop grid computing system in order to guarantee continuous and reliable execution of tasks in spite of failures. In a desktop grid computing environment, volunteers are often susceptible to volunteer autonomy failures such as volatility failure and interference failure in the middle of execution of tasks because a desktop grid computing maximally respects autonomy of volunteers. The failures result in an independent livelock problem (i.e. the delay and blocking of the entire execution of a job). Therefore, the failures should be considered in a scheduling mechanism. In This work, in order to tolerate volunteer autonomy failures, we propose a new fault tolerant scheduling mechanism. First, we specify a volunteer autonomy failures and an independent livelock problem. Then, we propose a volunteer availability which reflects the degree of volunteer autonomy failures. Finally, we propose a fault tolerant scheduling mechanism based on volunteer availability (which is called VAFTSM).

Recently, hierarchical architecture for location databases has been proposed in order to accommodate the growing number of PCSs user. With the 3-level hierarchical database architecture, which is compatible with the current cellular mobile systems, newly developed additional databases, including the regional location database (RLR), are positioned between the HLR and the VLRs. We propose an efficient cache scheme, the Double T-thresholds Location Cache Scheme, could reduce the network and database costs to lookup a portable using the 3-level architecture. The idea behind our proposed scheme is to use two pieces of cache information, VLR and RLR, which are serving the called portables. These two pieces are required in order to exploit not only the locality of a registration area (RA) but also the locality of a regional registration area (RRA), a wide area covered by the RLR. We also use two threshold values in order to determine whether the two pieces are obsolete. The cost analysis presented in this paper shows that the double T-thresholds location cache scheme significantly reduces network and database costs for most patterns of portables.

This paper proposes a restoration scheme based on mobility learning and prediction in the presence of the failure of mobility databases in personal communication systems (PCSs). In PCSs, mobility databases must maintain the current location information of users to provide a fast connection for them. However, the malfunction of mobility databases may cause some location information to be lost. As a result, without an explicit restoration procedure, incoming calls to users may be rejected. Therefore, an explicit restoration scheme against the malfunction of mobility databases is needed to guarantee continuous service availability to users. Introducing mobility learning and prediction into the restoration process allows systems to locate users after a failure of mobility databases. In failure-free operations, the movement patterns of users are learned by a neuro-fuzzy inference system (NFIS). After a failure, an inference process of the NFIS is initiated and the users' future location is predicted. This is used to locate lost users after a failure. This proposal differs from previous approaches using a checkpoint because it does not need a backup process nor additional storage space to store checkpoint information. In addition, simulations show that our proposal can reduce the cost needed to restore the location records of lost users after a failure when compared to the checkpointing scheme.

In a sender-initiated load balancing algorithm, the overloaded processor continues to send unnecessary request messages for load transfer until underloaded processor is found while the system load is heavy. Therefore, it yields inefficient inter-processor communications, low cpu utilization, and low system throughput. To solve these problems, we propose an improved genetic algorithm method for sender-initiated load balancing in distributed systems, and define a suitable fitness function. In this scheme, the processors that the request messages are transfered to are determined by genetic algorithm. The method also decreases unnecessary request messages. Compared with the conventional sender-initiated algorithms, we show that the proposed algorithm performs better.