Search

Large language models have demonstrated exceptional performance across various natural language processing tasks. However, their reliance on unstructured text corpora for pre-training limits their effectiveness in tasks requiring structured reasoning such as multi-hop question-answering. Knowledge Graphs provide a rich, structured source of relational data, offering an opportunity to enhance the reasoning capabilities of Large language models. In this paper, we propose a novel framework, Knowledge Graph as Pre-training Corpus (KGPC), which transforms knowledge graphs into text using a multi-hop linearization process. Unlike existing approaches that linearize singular triples, our method captures the interconnected nature of knowledge graphs by linking multiple triples across multiple hops, preserving their relational structure during the pre-training phase. This structured knowledge injection improves language models to perform complex reasoning tasks. We evaluate our approach on multi-hop reasoning benchmarks, demonstrating significant performance gains over existing models, particularly in question-answering tasks. Our results highlight the potential of multi-hop linearization in enhancing the structural reasoning capacity of language models, reducing error propagation, and improving the integration of structured knowledge into language models.

Abstract High-throughput screening techniques are pivotal to unlocking the mysteries of biology. Yet, the promise of droplet microfluidics in enabling single-cell resolution, ultra-high-throughput screening remains largely unfulfilled. Droplet sorting errors caused by polydisperse droplet sizes that are often inevitable in multi-step assays have severely limited the effectiveness and utility of this technique, especially when screening large libraries. Even a relatively low 1% sorting error results in 10,000 false calls in a 1,000,000 droplet screen, imposing an unreasonably large burden on downstream validation. Here, we present NOVAsort (Next-generation Opto-Volume-based Accurate droplet sorter), a device capable of discerning droplets based on both size and fluorescence intensity. With a 1000- and 10,000-fold reduction in false positives and false negatives, respectively. NOVAsort addresses the challenges of conventional droplet sorting approaches and sets standards for accuracy and throughput in droplet microfluidic assays.

Abstract Credit scoring models are critical for financial institutions to assess borrower risk and maintain profitability. Although machine learning models have improved credit scoring accuracy, imbalanced class distributions remain a major challenge. The widely used Synthetic Minority Oversampling TEchnique (SMOTE) struggles with high-dimensional, non-linear data and may introduce noise through class overlap. Generative Adversarial Networks (GANs) have emerged as an alternative, offering the ability to model complex data distributions. Conditional Wasserstein GANs (cWGANs) have shown promise in handling both numerical and categorical features in credit scoring datasets. However, research on extracting latent features from non-linear data and improving model explainability remains limited. To address these challenges, this paper introduces the Non-parametric Oversampling Technique for Explainable credit scoring (NOTE). The NOTE offers a unified approach that integrates a Non-parametric Stacked Autoencoder (NSA) for capturing non-linear latent features, cWGAN for oversampling the minority class, and a classification process designed to enhance explainability. The experimental results demonstrate that NOTE surpasses state-of-the-art oversampling techniques by improving classification accuracy and model stability, particularly in non-linear and imbalanced credit scoring datasets, while also enhancing the explainability of the results.

Credit scoring models play a crucial role for financial institutions in evaluating borrower risk and sustaining profitability. Logistic regression is widely used in credit scoring due to its robustness, interpretability, and computational efficiency; however, its predictive power decreases when applied to complex or non-linear datasets, resulting in reduced accuracy. In contrast, tree-based machine learning models often provide enhanced predictive performance but struggle with interpretability. Furthermore, imbalanced class distributions, which are prevalent in credit scoring, can adversely impact model accuracy and robustness, as the majority class tends to dominate. Despite these challenges, research that comprehensively addresses both the predictive performance and explainability aspects within the credit scoring domain remains limited. This paper introduces the Non-pArameTric oversampling approach for Explainable credit scoring (NATE), a framework designed to address these challenges by combining oversampling techniques with tree-based classifiers to enhance model performance and interpretability. NATE incorporates class balancing methods to mitigate the impact of imbalanced data distributions and integrates interpretability features to elucidate the model's decision-making process. Experimental results show that NATE substantially outperforms traditional logistic regression in credit risk classification, with improvements of 19.33% in AUC, 71.56% in MCC, and 85.33% in F1 Score. Oversampling approaches, particularly when used with gradient boosting, demonstrated superior effectiveness compared to undersampling, achieving optimal metrics of AUC: 0.9649, MCC: 0.8104, and F1 Score: 0.9072. Moreover, NATE enhances interpretability by providing detailed insights into feature contributions, aiding in understanding individual predictions. These findings highlight NATE's capability in managing class imbalance, improving predictive performance, and enhancing model interpretability, demonstrating its potential as a reliable and transparent tool for credit scoring applications.

Recommender systems play a crucial role in personalizing online user experiences by creating user profiles based on user–item interactions and preferences. Knowledge graphs (KGs) are intricate data structures that encapsulate semantic information, expressing users and items in a meaningful way. Although recent deep learning-based recommendation algorithms that embed KGs have demonstrated impressive performance, the richness of semantics and explainability embedded in the KGs are often lost due to the opaque nature of vector representations in deep neural networks. To address this issue, we propose a novel user profiling method for recommender systems that can encapsulate user preferences while preserving the original semantics of the KGs, using frequent subgraph mining. Our approach involves creating user profile vectors from a set of frequent subgraphs that contain information about user preferences and the strength of those preferences, measured by frequency. Subsequently, we trained a deep neural network model to learn the relationship between users and items, thereby facilitating effective recommendations using the neural network’s approximation ability. We evaluated our user profiling methodology on movie data and found that it demonstrated competitive performance, indicating that our approach can accurately represent user preferences while maintaining the semantics of the KGs. This work, therefore, presents a significant step towards creating more transparent and effective recommender systems that can be beneficial for a wide range of applications and readers interested in this field.

Optogenetics has the potential to transform the study of the peripheral nervous system (PNS), but the complex anatomy of the PNS poses unique challenges for the focused delivery of light to specific tissues. This protocol describes the fabrication of a wireless telemetry system for studying peripheral sensory pathways. Unlike existing wireless approaches, the low-power wireless telemetry offers organ specificity via a sandwiched pre-curved tether, and enables high-throughput analysis of behavioral experiments with a channel isolation strategy. We describe the technical procedures for the construction of these devices, the wireless power transmission (TX) system with antenna coils, and their implementation for in vivo experimental applications. In total, the timeline of the procedure, including device fabrication, implantation, and preparation to begin in vivo experimentation can be completed in ~2-4 weeks. Implementation of these devices allows for chronic (>1 month) wireless optogenetic manipulation of peripheral neural pathways in freely behaving animals navigating homecage environments (up to 8).

Infrared thermal imaging has been widely used to show the correlation between thermal characteristics of the body and muscle activation. This study aims to investigate a method using thermal imaging to visualize and differentiate target muscles during resistance training. Thermal images were acquired to monitor three target muscles (i.e., biceps brachii, triceps brachii, and deltoid muscle) in the brachium while varying the training weight, duration, and order of training. The acquired thermal images were segmented and converted to heat maps. By generating difference heat maps from pairs of heat maps during training, the target muscles were clearly visualized, with an average temperature difference of 0.86 °C. It was observed that training order had no significant effect on skin surface temperature. The difference heat maps were also used to train a convolutional neural network (CNN) to show the feasibility of target muscle classification, with an accuracy of 92.3%. This study demonstrated that infrared thermal imaging could be effectively utilized to locate and differentiate target muscle activation during resistance training.

In many areas, vast amounts of information are rapidly accumulating in the form of ontology-based knowledge graphs, and the use of information in these forms of knowledge graphs is becoming increasingly important. This study proposes a novel method for efficiently learning frequent subgraphs (i.e., knowledge) from ontology-based graph data. An ontology-based large-scale graph is decomposed into small unit subgraphs, which are used as the unit to calculate the frequency of the subgraph. The frequent subgraphs are extracted through candidate generation and chunking processes. To verify the usefulness of the extracted frequent subgraphs, the methodology was applied to movie rating prediction. Using the frequent subgraphs as user profiles, the graph similarity between the rating graph and new item graph was calculated to predict the rating. The MovieLens dataset was used for the experiment, and a comparison showed that the proposed method outperformed other widely used recommendation methods. This study is meaningful in that it proposed an efficient method for extracting frequent subgraphs while maintaining semantic information and considering scalability in large-scale graphs. Furthermore, the proposed method can provide results that include semantic information to serve as a logical basis for rating prediction or recommendation, which existing methods are unable to provide.