Large language models (LLMs) have demonstrated significant advancements in code generation, yet they still face challenges when tackling tasks that extend beyond their basic capabilities. Recently, the concept of self-debugging has been proposed as a way to enhance code generation performance by leveraging execution feedback from tests. However, the availability of high-quality tests in real-world scenarios is often limited. In this context, self-debugging with self-generated tests emerges as a promising solution, though its limitations and practical potential have not been fully explored. To address this gap, we investigate the efficacy of self-debugging in code generation tasks. We propose and analyze two distinct paradigms for the self-debugging process: post-execution and in-execution self-debugging. Our findings reveal that post-execution self-debugging struggles with the test bias introduced by self-generated tests, which can lead to misleading feedback. In contrast, in-execution self-debugging enables LLMs to mitigate this bias and leverage intermediate states during program execution. By focusing on runtime information rather than relying solely on potentially flawed self-generated tests, this approach demonstrates significant promise for improving the robustness and accuracy of LLMs in code generation tasks.

In the context of intensified market competition, consumer demand for aesthetically pleasing and functionally designed products has grown significantly. Color matching in product appearance plays a critical role in influencing consumer choice. However, designers often face challenges related to low color recognition accuracy, which hampers efficiency and design quality. This study explores the application of artificial intelligence (AI) technology to assist in the color matching process of product appearance and functional design. Experimental evaluation across various product types demonstrates that AI integration improves color accuracy by 2.09% and enhances the stability of image representation by 3.47%. Additionally, it reduces design analysis time, increases designer productivity, and boosts satisfaction scores by 5.4%. The findings confirm that AI technology effectively supports designers in achieving more accurate, efficient, and satisfactory color matching outcomes.

Geo-textual objects with both geographical location and textual description are gaining in prevalence. Over the past decades, substantial research has been conducted on spatial keyword queries, which integrate location into keyword-based querying of geo-textual content. However, existing proposals mostly focus on efficiency for processing spatial keyword queries, and little effort was made to address the effectiveness perspectives. In this work, using two datasets with ground truth query results, we evaluate the effectiveness of standard spatial keyword queries. Our evaluation results show that the TkQ query that ranks objects by a weighted combination of spatial proximity and text relevance is the most effective. Motivated by the finding, we propose a Deep relevance with Weight learning (DrW) model to further improve the effectiveness of the retrieval ranking. DrW is featured with two novel ideas: First, we propose a neural network architecture to learn the text relevance matching over the local interaction between the query and geo-textual objects. Second, we find that a query-dependent weight to balance text relevance and spatial proximity in ranking can improve effectiveness, and we develop a learning-based method to learn the query-dependent weight. Experimental results reveal that our model outperforms state-of-the-art methods on effectiveness, with improvements up to 32.15%, 32.34%, and 33.00% in terms of NDCG@3, NDCG@5, and MRR.

As a fundamental component in location-based services, inferring the relationship between points-of-interests (POIs) is very critical for service providers to offer good user experience to business owners and customers. Most of the existing methods for relationship inference are not targeted at POI, thus failing to capture unique spatial characteristics that have huge effects on POI relationships. In this work we propose PRIM to tackle POI relationship inference for multiple relation types. PRIM features four novel components, including a weighted relational graph neural network, category taxonomy integration, a self-attentive spatial context extractor, and a distance-specific scoring function. Extensive experiments on two real-world datasets show that PRIM achieves the best results compared to state-of-the-art baselines and it is robust against data sparsity and is applicable to unseen cases in practice.

To provide more accurate personalized product search (PPS) results, it is compulsory to go beyond modeling user-query-item interaction. Graph embedding techniques open the potential to integrate node information and topological structure information. Existing graph embedding enhanced PPS methods are mostly based on entity-relation-entity graph learning. In this work, we propose to consider structural relationship in users' product search scenario with graph embedding by latent representation learning. We argue that explicitly modeling the structural relationship in graph embedding is essential for more accurate PPS results. We propose a novel method, Graph embedding based Structural Relationship Representation Learning (GraphSRRL), which explicitly models the structural relationship in users-queries-products interaction. It combines three key conjunctive graph patterns to learn graph embedding for better PPS. In addition, GraphSRRL facilitates the learning of affinities between users (resp. queries or products) in the designed geometric operation in low-dimensional latent space. We conduct extensive experiments on four datasets to evaluate GraphSRRL for PPS. Experimental results show that GraphSRRL outperforms the state-of-the-art algorithm on real-world search datasets by at least 50.7% in term of [email protected] and 48.7% in terms of [email protected]

Eliminating the influence of cogging torque and model uncertainty on the tracking control of a dc motor when its speed varies nonperiodically is a challenge. In this paper, an adaptive robust triple-step control method is proposed for compensating cogging torque and model uncertainty. First, a new presentation of the cogging torque and a simplified model of the friction torque are presented to facilitate the online estimation of the unknown model parameters. The load torque, motor disturbance, and model errors are considered as model uncertainty. Based on these considerations, a control-oriented model that contains unknown parameters and model uncertainty is obtained. Second, benefitting from the new presentation, an adaptive algorithm is employed to identify the unknown parameters online. The model uncertainty is estimated by an extended state observer. Third, the model-based triple-step nonlinear method is extended to a system with both parameter uncertainty and model uncertainty, and an adaptive robust triple-step nonlinear controller is derived. The robust stability of the closed-loop system is proven in the framework of Lyapunov theory. Finally, the effectiveness and the satisfactory control performance of this controller are evaluated through comparative experiments on a J60LYS05 motor.