Retrieval-augmented generation (RAG) enhances the reasoning capabilities of large language models (LLMs) by incorporating external knowledge. Among available sources, knowledge graphs (KGs) offer a structured and reliable foundation for factual information, making them increasingly popular in efforts to improve reasoning faithfulness in RAG. Most existing KG-based RAG methods rely on LLMs to extract knowledge from KGs. However, these approaches often require costly fine-tuning and struggle to navigate deep graph structures, limiting their effectiveness in multi-hop reasoning tasks. To address these challenges, we propose Stepwise Contrastive Reasoning (SCR), a lightweight framework that integrates graph structure and textual context for efficient and interpretable RAG over KGs. SCR combines relational message passing layers to encode KG entities with a Transformer encoder for processing question text. It decomposes reasoning into a series of alignment steps. At each step, SCR compares the current topic entity and its neighbors with the question representation, selecting the most relevant entity as the next topic entity. The question is then updated with this entity's textual description. This process continues until the selected entity no longer changes, indicating that the answer entity has been reached. Through stepwise alignment, SCR enables compact models to perform faithful and interpretable reasoning over large-scale KGs. Extensive experiments on several widely used KGQA benchmarks demonstrate that SCR not only achieves state-of-the-art performance but also effectively boosts the capabilities of smaller language models to match those of LLMs.

Machine learning has been successfully applied to drug combination prediction in recent years. However, in some situations, the class imbalance problem still shows highly negative impacts on the modeling process, which cannot be directly handled by traditional methods. In addition, the interpretability of models is another key point for biological and medical experts. In this study, a clustering-based oblique decision tree (MOCT) algorithm is proposed to extract interpretable knowledge for the multi-class datasets. It firstly clusters samples of different classes, and then a proper feature subspace is generated to split data and forms a nonleaf node. Unlike traditional decision trees, our MOCT only grows one none-leaf node in each layer to generate a concise tree structure. Datasets of drug combinations were collected from three cell lines with three classes (Additive, Antagonism, and Synergy) in experiments, and the results show that our MOCT algorithm is superior to other methods with better interpretability.

Recent advances in AI-driven storytelling have enhanced video generation and story visualization. However, translating dialogue-centric scripts into coherent storyboards remains a significant challenge due to limited script detail, inadequate physical context understanding, and the complexity of integrating cinematic principles. To address these challenges, we propose Dialogue Visualization, a novel task that transforms dialogue scripts into dynamic, multi-view storyboards. We introduce Dialogue Director, a training-free multimodal framework comprising three agents–Script Director, Cinematographer, and Storyboard Maker. This framework leverages large multimodal models and diffusion-based architectures, employing techniques such as Chain-of-Thought reasoning, Retrieval-Augmented Generation, and multi-view synthesis to improve script understanding, physical context comprehension, and cinematic knowledge integration. Experimental results demonstrate that Dialogue Director outperforms state-of-the-art methods in script interpretation, physical world understanding, and cinematic principle application, significantly advancing the quality and controllability of dialogue-based story visualization.

For the interaction prediction task in the biomedical field, most machine learning algorithms overlook the relationships between entities within a pair by treating their features independently. To address this issue, this paper proposes a novel Cluster-based Paired Decision Tree model (CPDT), which pairs synonymous features of entity pairs to form paired feature spaces for simultaneous processing. It employs an adaptive grid-based clustering algorithm to partition these spaces in an axis-parallel manner, constructing interpretable decision boundaries. Moreover, the clustering algorithm leverages the probability density function to accommodate various data distributions in paired feature spaces, enhancing the effectiveness of sample partitioning. Experimental results demonstrate that CPDT performs well in two interaction prediction tasks: Drug Combination and Synthetic Lethality predictions. Furthermore, CPDT yields simple and interpretable decision rules that uncover potential patterns in biomedical interaction prediction. It also identifies molecules with medical significance, suggesting promising applications in the biomedical domain.

The existing methods for audio-driven talking head video editing have the limitations of poor visual effects. This paper tries to tackle this problem through editing talking face images seamless with different emotions based on two modules: (1) an audio-to-landmark module, consisting of the CrossReconstructed Emotion Disentanglement and an alignment network module. It bridges the gap between speech and facial motions by predicting corresponding emotional landmarks from speech; (2) a landmark-based editing module edits face videos via StyleGAN. It aims to generate the seamless edited video consisting of the emotion and content components from the input audio. Extensive experiments confirm that compared with state-of-the-arts methods, our method provides high-resolution videos with high visual quality.

Detecting pneumonia, especially coronavirus disease 2019 (COVID-19), from chest X-ray (CXR) images is one of the most effective ways for disease diagnosis and patient triage. The application of deep neural networks (DNNs) for CXR image classification is limited due to the small sample size of the well-curated data. To tackle this problem, this article proposes a distance transformation-based deep forest framework with hybrid-feature fusion (DTDF-HFF) for accurate CXR image classification. In our proposed method, hybrid features of CXR images are extracted in two ways: hand-crafted feature extraction and multigrained scanning. Different types of features are fed into different classifiers in the same layer of the deep forest (DF), and the prediction vector obtained at each layer is transformed to form distance vector based on a self-adaptive scheme. The distance vectors obtained by different classifiers are fused and concatenated with the original features, then input into the corresponding classifier at the next layer. The cascade grows until DTDF-HFF can no longer gain benefits from the new layer. We compare the proposed method with other methods on the public CXR datasets, and the experimental results show that the proposed method can achieve state-of-the art (SOTA) performance. The code will be made publicly available at https://github.com/hongqq/DTDF-HFF.

The interpretable machine learning method is important in drug discovery. Unlike traditional ensemble learning methods, this paper proposes an interpretable algorithm based on Bayesian rule extraction to obtain reliable and explainable results for hepatotoxicity prediction. To extract information from different types of omics data, our algorithm employs a multi-view learning strategy to enhance performance. Specifically, a random forest is trained in each view, and then the Bayesian rule extraction algorithm is designed to select an optimal rule subset, controlling the size and accuracy of the ruleset through probabilities. These rule sets are integrated through multi-view voting to get the final decisions. The performance of our algorithm is tested on the hepatotoxicity dataset, demonstrating that compared to traditional machine learning algorithms and rule-based algorithms, our approach maintains excellent performance while achieving high interpretability in most cases. Our python source code and the related Supplementary Materials are available at: github.com/MLDMXM2017/MV-BRS.

Speech emotion recognition (SER) plays a vital role in improving the interactions between humans and machines by inferring human emotion and affective states from speech signals. Whereas recent works primarily focus on mining spatiotemporal information from hand-crafted features, we explore how to model the temporal patterns of speech emotions from dynamic temporal scales. Towards that goal, we introduce a novel temporal emotional modeling approach for SER, termed Temporal-aware bI-direction Multi-scale Network (TIM-Net), which learns multi-scale contextual affective representations from various time scales. Specifically, TIM-Net first employs temporal-aware blocks to learn temporal affective representation, then integrates complementary information from the past and the future to enrich contextual representations, and finally fuses multiple time scale features for better adaptation to the emotional variation. Extensive experimental results on six benchmark SER datasets demonstrate the superior performance of TIM-Net, gaining 2.34% and 2.61% improvements of the average UAR and WAR over the second-best on each corpus. The source code is available at https://github.com/Jiaxin-Ye/TIM-Net_SER.

The texture recognition can provide clues for robots to interact with the external environment. The traditional tactile material recognition task is studied under the close-set assumption, which means that all types of materials are included in the training set. However, the open-set materials recognition for robots is of much greater significance because in the real-world applications, there is usually something that doesn't belong to any known class. Up to now, there is no researcher to further the discussion of this problem. To cope with unknown classes, this study proposes the Open set Material Recognition (OpenMR) based on General Convolutional Prototype Learning (GCPL). To handle the open space risk for GCPL caused by the lack of unknown samples in the training stage, we use Generative Adversarial Networks (GAN) to synthesize open-set samples as unknowns. The proposed framework is implemented and tested on two batches of tactile data collected in different exploratory motions on 8 material textures using the electronic skin. Compared with other open-set classifiers, experiments reveal that the proposed framework achieves competitive performance in both known classification and unknown detection.

Cross-corpus speech emotion recognition (SER) seeks to generalize the ability of inferring speech emotion from a well-labeled corpus to an unlabeled one, which is a rather challenging task due to the significant discrepancy between two corpora. Existing methods, typically based on unsupervised domain adaptation (UDA), struggle to learn corpus-invariant features by global distribution alignment, but unfortunately, the resulting features are mixed with corpus-specific features or not class-discriminative. To tackle these challenges, we propose a novel Emotion Decoupling aNd Alignment learning framework (EMO-DNA) for cross-corpus SER, a novel UDA method to learn emotion-relevant corpus-invariant features. The novelties of EMO-DNA are two-fold: contrastive emotion decoupling and dual-level emotion alignment. On one hand, our contrastive emotion decoupling achieves decoupling learning via a contrastive decoupling loss to strengthen the separability of emotion-relevant features from corpus-specific ones. On the other hand, our dual-level emotion alignment introduces an adaptive threshold pseudo-labeling to select confident target samples for class-level alignment, and performs corpus-level alignment to jointly guide model for learning class-discriminative corpus-invariant features across corpora. Extensive experimental results demonstrate the superior performance of EMO-DNA over the state-of-the-art methods in several cross-corpus scenarios. Source code is available at https://github.com/Jiaxin-Ye/Emo-DNA.

Knowledge graph completion plays a crucial role in downstream applications.However, existing methods tend to only rely on the structure or textual information, resulting in suboptimal model performance.Moreover, recent attempts to leverage pre-trained language models to complete knowledge graphs have proved unsatisfactory.To overcome these limitations, we propose a novel model that combines structural embedding and semantic information of the knowledge graph.Compared with previous works based on pre-trained language models, our model can better use the implicit knowledge of pre-trained language models by using relation templates, entity definitions, and learnable tokens.Furthermore, our model employs a multihead attention mechanism to transform the embedding semantic space of entities and relations obtained from the knowledge graph embedding model, thereby enhancing their expressiveness and unifying the semantic space of both types of information.Finally, we utilize convolutional neural networks to extract features from the matrices created by combining these two types of information for link prediction and triplet classification tasks.Empirical evaluations on two knowledge graph completion datasets demonstrate that our model is effective for both tasks.

INTRODUCTION: There are countless possibilities for drug combinations, which makes it expensive and time-consuming to rely solely on clinical trials to determine the effects of each possible drug combination. In order to screen out the most effective drug combinations more quickly, scholars began to apply machine learning to drug combination prediction. However, most of them are of low interpretability. Consequently, even though they can sometimes produce high prediction accuracy, experts in the medical and biological fields can still not fully rely on their judgments because of the lack of knowledge about the decision-making process. RELATED WORK: Decision trees and their ensemble algorithms are considered to be suitable methods for pharmaceutical applications due to their excellent performance and good interpretability. We review existing decision trees or decision tree ensemble algorithms in the medical field and point out their shortcomings. METHOD: This study proposes a decision stump (DS)-based solution to extract interpretable knowledge from data sets. In this method, a set of DSs is first generated to selectively form a decision tree (DST). Different from the traditional decision tree, our algorithm not only enables a partial exchange of information between base classifiers by introducing a stump exchange method but also uses a modified Gini index to evaluate stump performance so that the generation of each node is evaluated by a global view to maintain high generalization ability. Furthermore, these trees are combined to construct an ensemble of DST (EDST). EXPERIMENT: The two-drug combination data sets are collected from two cell lines with three classes (additive, antagonistic and synergistic effects) to test our method. Experimental results show that both our DST and EDST perform better than other methods. Besides, the rules generated by our methods are more compact and more accurate than other rule-based algorithms. Finally, we also analyze the extracted knowledge by the model in the field of bioinformatics. CONCLUSION: The novel decision tree ensemble model can effectively predict the effect of drug combination datasets and easily obtain the decision-making process.

Despite the development of computer vision techniques, the micro-expression (ME) recognition task still remains a great challenge because MEs have very low intensity and short duration. However, the ME recognition is of great significance since it provides important clues for real affective states detection. This paper proposes a novel Block Division Convolutional Network (BDCNN) with the implicit deep features augmentation. In detail, BDCNN learns from four optical flow features computed by the onset and apex frames of each video. It innovatively divides each image into a set of small blocks in the deep learning model, then the convolution and pooling operations are performed on these small blocks in sequence. To handle the small sample size problem in the micro-expression data, this study uses the improved implicit semantic data augmentation algorithm in the deep features space. Experiments are conducted on three publicly available databases, viz, CASME II, SMIC, and SAMM. Experimental results show that our model outperforms the state-of-the-art methods by attaining the accuracy of 84.32% and F1-score of 82.13% on the 3-class datasets, and the accuracy of 81.82% and F1-score of 75.46% on the 5-class datasets, respectively. Our source code is publicly available for non-commercial or research use athttps://github.com/MLDMXM2017/BDCNN.

Hepatotoxicity prediction is key to diseases with the high mortality rate. However, most of the algorithms used by now are black box in nature and lack of clear interpretability. This paper proposes a genetic algorithm-based interpretable algorithm based on rules extracted from a random forest. To take advantages from different types of omics data and molecular representations gathered from various datasets, our algorithm utilizes multiple distinct features to form a multi-view learning strategy. In detail, the genetic algorithm is designed to select optimal rules from each view, which are then used to form the ensemble of multi-view rule sets. The experiments are carried out to verify the performance of our algorithm on the hepatotoxicity data. The results confirm that our algorithm can gain high accuracy in most cases with more compact and shorter rules, compared with the original random forest or other rule-based algorithms. Our python source code and the related Supplementary Materials are available at: github.com/MLDMXM2017/MVR-GA.

Speech Emotion Recognition (SER) has become a growing focus of research in human-computer interaction. An essential challenge in SER is to extract common attributes from different speakers or languages, especially when a specific source corpus has to be trained to recognize the unknown data coming from another speech corpus. To address this challenge, a Capsule Network (CapsNet) and Transfer Learning based Mixed Task Net (CTL-MTNet) are proposed to deal with both the single-corpus and cross-corpus SER tasks simultaneously in this paper. For the single-corpus task, the combination of Convolution-Pooling and Attention CapsNet module (CPAC) is designed by embedding the self-attention mechanism to the CapsNet, guiding the module to focus on the important features that can be fed into different capsules. The extracted high-level features by CPAC provide sufficient discriminative ability. Furthermore, to handle the cross-corpus task, CTL-MTNet employs a Corpus Adaptation Adversarial Module (CAAM) by combining CPAC with Margin Disparity Discrepancy (MDD), which can learn the domain-invariant emotion representations through extracting the strong emotion commonness. Experiments including ablation studies and visualizations on both single- and cross-corpus tasks using four well-known SER datasets in different languages are conducted for performance evaluation and comparison. The results indicate that in both tasks the CTL-MTNet showed better performance in all cases compared to a number of state-of-the-art methods. The source code and the supplementary materials are available at: https://github.com/MLDMXM2017/CTLMTNet.

Combination therapy has shown an obvious curative effect on complex diseases, whereas the search space of drug combinations is too large to be validated experimentally even with high-throughput screens. With the increase of the number of drugs, artificial intelligence techniques, especially machine learning methods, have become applicable for the discovery of synergistic drug combinations to significantly reduce the experimental workload. In this study, in order to predict novel synergistic drug combinations in various cancer cell lines, the cell line-specific drug-induced gene expression profile (GP) is added as a new feature type to capture the cellular response of drugs and reveal the biological mechanism of synergistic effect. Then, an enhanced cascade-based deep forest regressor (EC-DFR) is innovatively presented to apply the new small-scale drug combination dataset involving chemical, physical and biological (GP) properties of drugs and cells. Verified by the dataset, EC-DFR outperforms two state-of-the-art deep neural network-based methods and several advanced classical machine learning algorithms. Biological experimental validation performed subsequently on a set of previously untested drug combinations further confirms the performance of EC-DFR. What is more prominent is that EC-DFR can distinguish the most important features, making it more interpretable. By evaluating the contribution of each feature type, GP feature contributes 82.40%, showing the cellular responses of drugs may play crucial roles in synergism prediction. The analysis based on the top contributing genes in GP further demonstrates some potential relationships between the transcriptomic levels of key genes under drug regulation and the synergism of drug combinations.

The diagnosis and analysis of cancer are usually roughly judged through the accumulation of professional knowledge, which is difficult to deal with a large number of patient samples and a variety of causes and symptoms. Moreover, most of the existing machine learning methods are black-box, and can not give reasonable diagnosis basis. Therefore, an accurate and interpretable method is urgently required for cancer diagnosis. In this paper, a total of 700 serum samples consisting of three groups of patients and one group of healthy individuals were collected and subjected to SERS measurements. We rank the Raman spectra of 700 human SERA according to the feature importance, and construct the feature importance vector reflecting the spectral feature importance. We further construct candidate feature sets based on importance selection, so as to construct a random forest model based on feature ensemble selection. On the one hand, we compare the proposed method with the popular machine learning methods to verify the effectiveness in the task of cancer screening. On the other hand, we conduct qualitative and quantitative analysis of cancer characteristics, and give model basis and biomedical explanation for the impact of different important cancer characteristics on the final classification and diagnosis. Some more experimental results and discussions are included in the appendix. Our source code and appendix are available under https://github.com/liujiatong429/BIBM2021

Cardiac medical image segmentation plays an important role in the diagnosis and clinical treatment of cardiovascular diseases. However, due to the variability of cardiac anatomy and the ambiguity between cardiac substructures, it is still difficult to quickly segment the entire heart from medical images. Most of the current researches utilize neural network structure to perform whole heart segmentation. Although good segmentation accuracy has been achieved, it usually requires a long training time. This paper aims to build a new whole heart segmentation model based on Deep Forest, called Multi-Resolution Deep Forest Framework(MRDFF), which performs segmentation through two stages. In the first stage, the heart region is located by rough binary classification, and similarity screening is used to reduce redundancy. The second stage subdivides the heart substructures based on the results of the first stage and uses multi-scale fusion to achieve high segmentation accuracy. The experimental results conducted on the public data set MM-WHS show that under the same training data and configuration, our model can be trained in only 4.5 hours, which is about 1/2 of the training time of neural network models, and can reach the accuracy not lower than neural network models, which shows the feasibility and efficiency of our model. The code will be made publicly available at https://github.com/xufeixf/MRDFF.

Speech emotion recognition (SER) is an important and challenging task. It requires that a machine learning model process a person's speech signals and to judge his emotional state accurately. Due to the high dimensionality of the audio data, the extracted features are always noisy. Besides, the abstraction of audio features makes it impossible to fully use the inherent relationship among audio features. This paper proposes a model that combines a convolutional neural network (CNN) and a capsule network (CapsNet), named as CapCNN. The advantage of CapCNN lies in that it provides a solution for time sensitivity, and gives the overall characteristics. In this study, it is found that CapCNN can well handle the SER task. Compared with other state-of-the-art methods, our algorithm shows high performances on the CASIA and EMODB datasets. The detailed analysis confirms that our method provides balanced results on various classes.

The classification of multi-class microarray datasets is a hard task because of the small samples size in each class and the heavy overlaps among classes. To effectively solve these problems, we propose a novel Error Correcting Output Code (ECOC) algorithm by Enhance Class Separability related Data Complexity measures during encoding process, named as ECOCECS. In this algorithm, two nearest neighbor related DC measures are deployed to extract the intrinsic overlapping information from microarray data. Our ECOC algorithm aims to search an optimal class split scheme by minimizing these measures. The class splitting process ends when each class is separated from others, and then the class assignment scheme is mapped as a coding matrix. Experiments are carried out on seven microarray datasets, and results demonstrate the effectiveness and robustness of our method in comparison with four state-of-art ECOC methods. In short, our work shows that it is promising to apply the DC theory to ECOC framework.

Microarray technology has been widely applied to search for biomarkers of diseases, diagnose diseases and analyze gene regulatory network. Abundance of expression data from microarray experiments are processed by informatics tools, such as supporting vector machines (SVM), artificial neural network (ANN), and so on. These methods achieve good results in single dataset. Nevertheless, most analyses of microarray data are only focused on a series of data obtained from the same lab or gene chip. Then the discoveries may only be suitable for data they experimented on but lack of general sense. In this paper, we propose a genetic programming (GP) based approach to analyze microarray datasets. The GP implements classification and feature selection at the same time. To validate the significance of the selected genes and generated classification rules, the results are tested on different datasets obtained from different experimental conditions. The results confirm the efficiency of GP in the classification of different samples.

MOTIVATION: Feature selection approaches have been widely applied to deal with the small sample size problem in the analysis of micro-array datasets. For the multiclass problem, the proposed methods are based on the idea of selecting a gene subset to distinguish all classes. However, it will be more effective to solve a multiclass problem by splitting it into a set of two-class problems and solving each problem with a respective classification system. RESULTS: We propose a genetic programming (GP)-based approach to analyze multiclass microarray datasets. Unlike the traditional GP, the individual proposed in this article consists of a set of small-scale ensembles, named as sub-ensemble (denoted by SE). Each SE consists of a set of trees. In application, a multiclass problem is divided into a set of two-class problems, each of which is tackled by a SE first. The SEs tackling the respective two-class problems are combined to construct a GP individual, so each individual can deal with a multiclass problem directly. Effective methods are proposed to solve the problems arising in the fusion of SEs, and a greedy algorithm is designed to keep high diversity in SEs. This GP is tested in five datasets. The results show that the proposed method effectively implements the feature selection and classification tasks.

In this paper, an efficient feature extraction method named as Constrained Maximum Variance Mapping (CMVM) is developed for dimensionality reduction. The proposed algorithm can be viewed as a linear approximation of multi-manifolds based learning approach, which takes the local geometry and manifold labels into account. After the local scatters have been characterized, the proposed method focuses on developing a linear transformation that can maximize the distances matrix between all the manifolds under the constraint of locality preserving. Then, YALE face database, ORL face database are all taken to examine the effectiveness and efficiency of the proposed method. Experimental results validate that the proposed approach is superior to other widely used feature extraction methods.

It is a challenging task to diagnose tumor type precisely based on microarray data because the number of variables p (genes) is far larger than that of samples, n. Many independent component analysis (ICA) based models had been proposed to tackle the microarray data classification problem with great success. Although it was pointed out that different independent components (ICs) are of different biological significance, up to now, it is still far from well explored for the problem that how to select proper IC subsets to predict new samples best. We try to improve the performance of ICA based classification models by using proper IC subsets instead of all the ICs. A genetic algorithms (GA) based selection process is proposed in this paper, and the selected IC subset is evaluated by the leave-one-out cross validation (LOOCV) technique. The experimental results demonstrate that our GA based IC selection method can further improve the classification accuracy of the ICA based prediction models.

Microarray data prediction is a hard task due to the small sample and high dimension property. This paper proposes a classifier fusion approch to solve this problem based on genetic algorithm (GA). In this fusion strategy, GA is applied to select proper feature subsets and weight value for the fusion of classifiers. The experimental results show that the proposed scheme can improve the prediction accuracy.

This paper presents a novel multi-sub-swarm Particle Swarm Optimization (PSO) algorithm. The proposed algorithm can effectively imitate a natural ecosystem, in which the different sub-populations can compete with each other. After competing, the winner will continue to explore the original district, while the loser will be obliged to explore another district. Four benchmark multimodal functions of varying difficulty are used as test functions. The experimental results show that the proposed method has a stronger adaptive ability and a better performance for complicated multimodal functions with respect to other methods.

This paper presents a new approach based on HMM/ANN hybrid for online signature verification. A group of ANNs are used as local probability estimators for an HMM. The Viterbi algorithm is employed to work out the global posterior probability of a model. The proposed HMM/ANN hybrid has a strong discriminant ability, i.e, from a local sense, the ANN can be regarded as an efficient classifier, and from a global sense, the posterior probability is consistent with that of a Bayes classifier. Finally, the experimental results show that this approach is promising and competing.

In this paper, self-organizing map (SOM) is used to visualize and cluster the data set of aerosol single particle mass spectrum, which was collected by aerosol time-of-flight mass spectrometry (ATOFMS). In view of the characteristic feature of aerosol particle data, the TF-IDF scheme used widely in document clustering is employed to preprocess. Subsequently for data clustering analysis, a two-level clustering framework is proposed, wherein SOM is firstly used to cluster input data and get the primary results, and then the results are again clustered by semiautomatic k-means algorithm. In order to demonstrate the validity of clustering, the chemical significance for cluster centroid is also investigated, wherein inorganic salts, "calcium-containing" particles, biogenic soot particles, and carbonaceous particles etc. are identified.