Personalizing text-to-image diffusion models is crucial for adapting the pre-trained models to specific target concepts, enabling diverse image generation. However, fine-tuning with few images introduces an inherent trade-off between aligning with the target distribution (e.g., subject fidelity) and preserving the broad knowledge of the original model (e.g., text editability). Existing sampling guidance methods, such as classifier-free guidance (CFG) and autoguidance (AG), fail to effectively guide the output toward well-balanced space: CFG restricts the adaptation to the target distribution, while AG compromises text alignment. To address these limitations, we propose personalization guidance, a simple yet effective method leveraging an unlearned weak model conditioned on a null text prompt. Moreover, our method dynamically controls the extent of unlearning in a weak model through weight interpolation between pre-trained and fine-tuned models during inference. Unlike existing guidance methods, which depend solely on guidance scales, our method explicitly steers the outputs toward a balanced latent space without additional computational overhead. Experimental results demonstrate that our proposed guidance can improve text alignment and target distribution fidelity, integrating seamlessly with various fine-tuning strategies.

This paper considers two types of imbalance problems commonly inherent in large-scale datasets: multiple domain and class imbalance. Class imbalance causes the algorithm to be biased toward the majority classes, and multiple-domain data results in significant performance disparities for different domains. To tackle these challenges, we propose a novel learning approach for multi-domain imbalanced datasets, featuring two techniques: (i) distribution-aware partial mask and (ii) domain-wise interprototype loss function. The distribution-aware partial mask selects negative class centers based on class-level distribution and domain labels, adjusting the ratio of positive and negative updates for prototype vectors and enhancing discriminative feature learning within each domain. Additionally, the domain-wise interprototype loss enforces orthogonality among prototype vectors within each domain, leading to increased discriminativeness. We demonstrate the superiority of our approach over baselines through experiments on publicly available speaker recognition datasets, including CN-Celeb and Mozilla Common Voice.

Deep learning models for verification systems often fail to generalize to new users and new environments, even though they learn highly discriminative features.To address this problem, we propose a few-shot domain generalization framework that learns to tackle distribution shift for new users and new domains.Our framework consists of domain-specific and domainaggregation networks, which are the experts on specific and combined domains, respectively.By using these networks, we generate episodes that mimic the presence of both novel users and novel domains in the training phase to eventually produce better generalization.To save memory, we reduce the number of domain-specific networks by clustering similar domains together.Upon extensive evaluation on artificially generated noise domains, we can explicitly show generalization ability of our framework.In addition, we apply our proposed methods to the existing competitive architecture on the standard benchmark, which shows further performance improvements.

Convolutional Neural Networks are widely used in various machine learning domains. In image processing, the features can be obtained by applying 2D convolution to all spatial dimensions of the input. However, in the audio case, frequency domain input like Mel-Spectrogram has different and unique characteristics in the frequency dimension. Thus, there is a need for a method that allows the 2D convolution layer to handle the frequency dimension differently. In this work, we introduce SubSpectral Normalization (SSN), which splits the input frequency dimension into several groups (sub-bands) and performs a different normalization for each group. SSN also includes an affine transformation that can be applied to each group. Our method removes the inter-frequency deflection while the network learns a frequency-aware characteristic. In the experiments with audio data, we observed that SSN can efficiently improve the network’s performance.

Summary Last‐mile bottleneck, fire‐wall bottleneck, and internal network congestion bottleneck are well‐known performance bottleneck problems of private sector in national R&E network since Internet is adopted for R&E network. The emergency of big data science makes these problems more severe because a famous scientist under such a circumstance is severely restricted in his/her big data R&D. Therefore, government cannot leave these stubs in the realm of a private organization any longer. We, KREONET (Korea Research Environment Open Network), tried to solve these problems by SDN‐IP (software‐defined network—Internet protocol) based on ONOS (Open Network Operation System); SDN‐IP is the leading technology for the softwarization of network, and it is developed by On Lab. KREONET began to deploy SDN‐IP as a tool for the softwarization of Korea R&E network even though others studied to business its' feature for the enhancement of resource management of data center first. The initial goal during the construction of SDN‐IP for KREONET is to solve well‐known problems in R&E network, which are related with last‐mile, fire‐wall, and internal network congestion. These 3 problems are not solved for a long time because it costs too much when we try to solve them by the way of the existed hardware networking. In this paper, we introduce our experience that can catch two rabbits at the same time. These experiences are about the provision of the initial popular service of SDN‐IP with the solution of those well‐known bottleneck problems and about the economic way of SDN‐IP construction by incremental hybrid networking of legacy Internet and SDN‐IP.