The advancement of vision-language models, particularly the Contrastive Language-Image Pre-training (CLIP) model, has revolutionized the field of machine learning by enabling robust zero-shot learning capabilities. These capabilities allow models to understand and respond to previously unseen data without task-specific training. However, adapting CLIP to integrate specialized knowledge from various domains while retaining its zero-shot capabilities remains a significant challenge. To address this, we introduce a novel prompt ensemble learning approach called Group-wise Prompt Ensemble (GPE). This method aims to enhance CLIP's zero-shot capabilities by incorporating new domain knowledge while improving its adaptability and robustness against data distribution shifts. Our approach hinges on three main strategies: prompt grouping with masked attention to optimize CLIP's adaptability while safeguarding its zero-shot capabilities; the incorporation of auxiliary prompts for the seamless integration of new domain insights without disrupting the original model's representation; and an ensemble learning strategy that effectively merges original and new knowledge. Through rigorous experimentation, including more challenging cross-dataset transfer evaluations, our GPE method redefines the benchmarks for the adaptability and efficiency of vision-language models, surpassing existing models across various scenarios.

The study proposes a rotational multipyramid network (RoMP Net) with bounding-box transformation for object detection. The RoMP Net is a single-stage object detection neural network featuring three characteristics. First, the network uses a rotational bounding box to minimize the effect of background images when extracting features of objects. Bounding-box transformation was proposed to compensate for the limitation of the rotational bounding boxes, which have relatively low prediction accuracy for objects with a high aspect ratio. Second, the RoMP Net introduces a multi-scale and multi-level feature pyramid network to extract distinct and semantic features efficiently. This network architecture ensures high prediction accuracy and robustness regardless of the size and complexity of objects. Third, hyperparameters in the bounding boxes are automatically determined through an unsupervised clustering method. This optimization method is also critical in improving accuracy. The performance of the proposed network and preprocessing methods are validated through image-sets comprising critical components in power transmission facilities, which have a variety of sizes and aspect ratios. This case study demonstrates the effectiveness and robustness of the three key characteristics in the RoMP Net. Furthermore, the RoMP Net outperforms other state-of-the-art deep neural networks in prediction accuracy and robustness for object detection. Specifically, the mean average precision of the RoMP Net in the validation image-sets shows that it has the highest prediction accuracy, whereas its values in the test image-sets confirm the network's robustness. The fast yet accurate RoMP Net will expand the range of object detection through deep neural networks.

In uplink 3GPP LTE, SC-FDMA has been regarded as an attractive alternative to OFDMA due to lower Peak-to-Average Power Ratio (PAPR) characteristic. However, in order to get the benefit of PAPR, it requires a contiguous assignment of resource blocks to a single user. In this paper, we suggest heuristic algorithm for proportional fair (PF) scheduling which is known to be NP-hard problem. Under the contiguity constraint, conventional heuristic algorithms assign resource block to a user who has the highest metric and keep on doing assignment according to greedy selection but without any examination of prior selected choice. Our work differs from the conventional works in considering user selection based on PF metric difference and comparison of marginal scheduling metric for examination of selected choice. In simulation results, we demonstrated that the proposed scheme shows better fairness in terms of CDF of normalized throughput and Jain's fairness index.

In uplink 3GPP Long-Term Evolution (LTE), single-carrier frequency division multiple access (SC-FDMA) is considered partly because of the fact that Peak-to-Average Power Ratio (PAPR) for Multi-Carrier Orthogonal Frequency Division Multiplexing (MC-OFDM) is still a hindrance. But, an unique characteristic of SC-FDMA is that subcarriers have to be assigned by contiguous manner for a specific user, which induces a complicated scheduling algorithm to SC-FDMA because a user is not able to exploit frequency selectivity gain fully. To increase system throughput, we propose a new frequency scheduling method using each user's channel gain difference and a contiguity assigning method. Focused on maximizing the total system capacity of SC-FDMA uplink, we compare our algorithms with others in uplink 3GPP LTE. Simulation results show that our proposed scheme outperforms over the existing schemes.