Similarity Inference Research Articles

Open-Vocabulary Predictive World Models from Sensor Observations.

Cognitive scientists believe that adaptable intelligent agents like humans perform spatial reasoning tasks by learned causal mental simulation. The problem of learning these simulations is called predictive world modeling. We present the first framework for a learning open-vocabulary predictive world model (OV-PWM) from sensor observations. The model is implemented through a hierarchical variational autoencoder (HVAE) capable of predicting diverse and accurate fully observed environments from accumulated partial observations. We show that the OV-PWM can model high-dimensional embedding maps of latent compositional embeddings representing sets of overlapping semantics inferable by sufficient similarity inference. The OV-PWM simplifies the prior two-stage closed-set PWM approach to the single-stage end-to-end learning method. CARLA simulator experiments show that the OV-PWM can learn compact latent representations and generate diverse and accurate worlds with fine details like road markings, achieving 69 mIoU over six query semantics on an urban evaluation sequence. We propose the OV-PWM as a versatile continual learning paradigm for providing spatio-semantic memory and learned internal simulation capabilities to future general-purpose mobile robots.

Human-inspired similarity control system: Enhancing line-following robot perception

Human-Inspired Control (HIC) holds promise for endowing machines with human-like cognition, decision-making, and adaptability. In this study, we employ a fusion of cognitive modeling, machine learning, and control theory as the foundational architecture and empirically validate its suitability for robot control within the realm of HIC. Fuzzy logic stands out as a viable approach for HIC control, wherein control rules can be devised drawing from human intuitive inspiration. Specifically, this study explores similarity inference control systems in robotics, with the objective of enhancing line-following control as an alternative to fuzzy systems. The experimental optimization results provide insights into the advantages and limitations of the similarity inference control system. Despite achieving performance comparable to that of traditional two-stage fuzzy control systems, careful consideration of noise sensitivity is paramount. While the similarity inference approach streamlines implementation and obviates the necessity for expert-designed fuzzy rules, its susceptibility to noise can compromise performance, particularly in noisy environments. These considerations are pivotal for the development of control systems aimed at mitigating noise sensitivity, enhancing task-specific performance, and ensuring the adaptability and robustness of line-following robots. To tackle this challenge, we both discuss and experimentally evaluate potential solutions and their applicability in this paper.

Multimodal Social Media Fake News Detection Based on Similarity Inference and Adversarial Networks

Multimodal Social Media Fake News Detection Based on Similarity Inference and Adversarial Networks

Genealogical inference and more flexible sequence clustering using iterative-PopPUNK.

Bacterial genome data are accumulating at an unprecedented speed due to the routine use of sequencing in clinical diagnoses, public health surveillance, and population genetics studies. Genealogical reconstruction is fundamental to many of these uses; however, inferring genealogy from large-scale genome data sets quickly, accurately, and flexibly is still a challenge. Here, we extend an alignment- and annotation-free method, PopPUNK, to increase its flexibility and interpretability across data sets. Our method, iterative-PopPUNK, rapidly produces multiple consistent cluster assignments across a range of sequence identities. By constructing a partially resolved genealogical tree with respect to these clusters, users can select a resolution most appropriate for their needs. We showed the accuracy of clusters at all levels of similarity and genealogical inference of iterative-PopPUNK based on simulated data and obtained phylogenetically concordant results in real data sets from seven bacterial species. Using two example sets of Escherichia/Shigella and Vibrio parahaemolyticus genomes, we show that iterative-PopPUNK can achieve cluster resolutions ranging from phylogroup down to sequence typing (ST). The iterative-PopPUNK algorithm is implemented in the "PopPUNK_iterate" program, available as part of the PopPUNK package.

Hierarchical Deep Multitask Learning With the Attention Mechanism for Similarity Learning

Similarity learning is often adopted as an auxiliary task of deep multitask learning methods to learn discriminant features. Most existing approaches only use the single-layer features extracted by the last fully connected layer, which ignores the abundant information of feature channels in lower layers. Besides, small cliques are the most commonly used methods in similarity learning tasks to model the correlation of data, which can lead to the limited relation learning. In this article, we present an end-to-end hierarchical deep multitask learning framework for similarity learning which can learn more discriminant features by sharing information from different layers of network and dealing with complex correlation. Its main task is graph similarity inference. We build focus graphs for each sample. Then, an attention mechanism and a node feature enhancing model are introduced into backbone network to extract the abundant and important channel information from multiple layers of network. In the similarity inference task, a relation enhancing mechanism is applied to graph convolutional network to leverage the crucial relation in channels, which can effectively facilitate the learning ability of the whole framework. The extensive experiments have been conducted to demonstrate the effectiveness of the proposed method on person reidentification and face clustering applications.

RFE-SRN: Image-text similarity reasoning network based on regional feature enhancement

Image-text matching aims to make a connection between visual and natural information. Some of the current methods have made great progress by using global alignment between images and sentences and local alignment between the image region and its corresponding word. However, the importance of the correlation between global alignment and local alignment is ignored to some extends. Therefore, in this paper, we propose a new region feature enhancement-based image text similarity inference network. Firstly, the image region feature is enhanced by graph convolutional neural network, which is used to find the correlation among different image region features with the production of features’ sematic relationship. Secondly, we propose a vector-based similarity representation to describe local and global alignment with a more comprehensive way. Finally, a graph convolution neural network is introduced to construct a similarity graph for propagating correlation between local alignment and global alignment to every part. By testing on the MSCOCO and Flickr30k dataset, our proposed method shows great accuracy performance and competitiveness.

Dual-granularity feature alignment for cross-modality person re-identification

RGB-infrared cross-modality person re-identification is a challenging image retrieval task due to the intra-class variations and cross-modality discrepancy. However, it is insufficient for existing methods to identify discriminative patterns with different granularities and align them across granularity and modalities in a synergistic manner. In this paper, we propose Dual-granularity Feature Alignment (DFA) approach for cross-modality re-ID, which accommodates dual-granularity feature extraction and cross-granularity feature alignment. More specifically, a feature extractor is introduced to generate discriminative patterns by dividing feature map into different granularities. Based on the above features, prototype learning is developed to assign different granularity features of the same identity to the same prototype and adaptively build the alignment across different granularities. To mine contextual relationships between training samples, a similarity inference is presented to enforce the similarity matrix closing to identity matrix. This allows model to find feature that conserves as much modality-shared information as possible while being least possible modality-specific information. Both modules work in a collaborative way to obtain the optimal intra-class compactness and inter-class separability for cross-modality image matching. Extensive experiments on several benchmarks demonstrate that our DAF surpasses the state-of-the-art methods. The code is available at https://github.com/PRIS-CV/DFA.

FacTeR-Check: Semi-automated fact-checking through semantic similarity and natural language inference

Our society produces and shares overwhelming amounts of information through Online Social Networks (OSNs). Within this environment, misinformation and disinformation have proliferated, becoming a public safety concern in most countries. Allowing the public and professionals to efficiently find reliable evidence about the factual veracity of a claim is a crucial step to mitigate this harmful spread. To this end, we propose FacTeR-Check, a multilingual architecture for semi-automated fact-checking and hoaxes propagation analysis that can be used to implement applications designed for both the general public and for fact-checking organisations. FacTeR-Check implements three different modules relying on the XLM-RoBERTa Transformer architecture to evaluate semantic similarity, to calculate natural language inference and to build search queries through automatic keywords extraction and Named-Entity Recognition. The three modules have been validated using state-of-the-art benchmark datasets, exhibiting good performance in all of them. Besides, FacTeR-Check is employed to collect and label a dataset, called NLI19-SP, composed of more than 40,000 tweets supporting or denying 60 hoaxes related to COVID-19, released publicly. Finally, an analysis of the data collected in this dataset is provided, which allows to obtain a deep insight of how disinformation operated during the COVID-19 pandemic in Spanish-speaking countries.

Animal consciousness and phenomenal concepts

ABSTRACT A phenomenal concept is a concept that one possesses only if one has the relevant experience. In this essay, I argue that phenomenal concept theorists, namely, those who believe that we acquire phenomenal concepts through being acquainted with the relevant experience, can never succeed in determining which species of non-human animals are phenomenally conscious because they prohibit any a priori correlation between phenomenal and non-phenomenal concepts. I make my argument by first discussing several ways in which a phenomenal concept theorist may explain which animal species are conscious, namely, with neuro-identity theory, behaviorism, and functionalism, and the problem that these approaches entail. I then illustrate how the alternative approaches of scientific inference to the best explanation, analogical inference, similarity inference, and inference by prototype face the same problem.

Optimal transport improves cell-cell similarity inference in single-cell omics data.

MotivationHigh‐throughput single-cell molecular profiling is revolutionizing biology and medicine by unveiling the diversity of cell types and states contributing to development and disease. The identification and characterization of cellular heterogeneity are typically achieved through unsupervised clustering, which crucially relies on a similarity metric.ResultsWe here propose the use of Optimal Transport (OT) as a cell–cell similarity metric for single-cell omics data. OT defines distances to compare high-dimensional data represented as probability distributions. To speed up computations and cope with the high dimensionality of single-cell data, we consider the entropic regularization of the classical OT distance. We then extensively benchmark OT against state-of-the-art metrics over 13 independent datasets, including simulated, scRNA-seq, scATAC-seq and single-cell DNA methylation data. First, we test the ability of the metrics to detect the similarity between cells belonging to the same groups (e.g. cell types, cell lines of origin). Then, we apply unsupervised clustering and test the quality of the resulting clusters. OT is found to improve cell–cell similarity inference and cell clustering in all simulated and real scRNA-seq data, as well as in scATAC-seq and single-cell DNA methylation data.Availability and implementationAll our analyses are reproducible through the OT-scOmics Jupyter notebook available at https://github.com/ComputationalSystemsBiology/OT-scOmics.Supplementary information Supplementary data are available at Bioinformatics online.

Mining the capacity of human-associated microorganisms to trigger rheumatoid arthritis—A systematic immunoinformatics analysis of T cell epitopes

Autoimmune diseases, often triggered by infection, affect ~5% of the worldwide population. Rheumatoid Arthritis (RA)–a painful condition characterized by the chronic inflammation of joints—comprises up to 20% of known autoimmune pathologies, with the tendency of increasing prevalence. Molecular mimicry is recognized as the leading mechanism underlying infection-mediated autoimmunity, which assumes sequence similarity between microbial and self-peptides driving the activation of autoreactive lymphocytes. T lymphocytes are leading immune cells in the RA-development. Therefore, deeper understanding of the capacity of microorganisms (both pathogens and commensals) to trigger autoreactive T cells is needed, calling for more systematic approaches. In the present study, we address this problem through a comprehensive immunoinformatics analysis of experimentally determined RA-related T cell epitopes against the proteomes of Bacteria, Fungi, and Viruses, to identify the scope of organisms providing homologous antigenic peptide determinants. By this, initial homology screening was complemented with de novo T cell epitope prediction and another round of homology search, to enable: i) the confirmation of homologous microbial peptides as T cell epitopes based on the predicted binding affinity to RA-related HLA polymorphisms; ii) sequence similarity inference for top de novo T cell epitope predictions to the RA-related autoantigens to reveal the robustness of RA-triggering capacity for identified (micro/myco)organisms. Our study reveals a much larger repertoire of candidate RA-triggering organisms, than previously recognized, providing insights into the underestimated role of Fungi in autoimmunity and the possibility of a more direct involvement of bacterial commensals in RA-pathology. Finally, our study pinpoints Endoplasmic reticulum chaperone BiP as the most potent (most likely mimicked) RA-related autoantigen, opening an avenue for identifying the most potent autoantigens in a variety of different autoimmune pathologies, with possible implications in the design of next-generation therapeutics aiming to induce self-tolerance by affecting highly reactive autoantigens.

Word n-gram attention models for sentence similarity and inference

Semantic Textual Similarity and Natural Language Inference are two popular natural language understanding tasks used to benchmark sentence representation models where two sentences are paired. In such tasks sentences are represented as bag of words, sequences, trees or convolutions, but the attention model is based on word pairs. In this article we introduce the use of word n-grams in the attention model. Our results on five datasets show an error reduction of up to 41% with respect to the word-based attention model. The improvements are especially relevant with low data regimes and, in the case of natural language inference, on the recently released hard subset of Natural Language Inference datasets.

Tripartite Network-Based Repurposing Method Using Deep Learning to Compute Similarities for Drug-Target Prediction.

The drug discovery process is conventionally regarded as resource intensive and complex. Therefore, research effort has been put into a process called drug repositioning with the use of computational methods. Similarity-based methods are common in predicting drug-target association or the interaction between drugs and targets based on various features the drugs and targets have. Heterogeneous network topology involving many biomedical entities interactions has yet to be used in drug-target association. Deep learning can disclose features of vertices in a large network, which can be incorporated with heterogeneous network topology in order to assist similarity-based solutions to provide more flexibility for drug-target prediction. Here we describe a similarity-based drug-target prediction method that utilizes a topology-based similarity measure and two inference methods based on the similarities. We used DeepWalk, a deep learning method, to calculate the vertex similarities based on Linked Tripartite Network (LTN), which is a heterogeneous network created from different biomedical-linked datasets. The similarities are further used to feed to the inference methods, drug-based similarity inference (DBSI) and target-based similarity inference (TBSI), to obtain the predicted drug-target associations. Our previous experiments have shown that by utilizing deep learning and heterogeneous network topology, the proposed method can provide more promising results than current topology-based similarity computation methods.

A shopping behavior prediction system: Considering moving patterns and product characteristics

In recent years, the development of location determination technologies such as GPS, Wi-Fi, and RFID have made it possible to collect locational data for moving customers, and many behavior prediction and recommendation systems have been proposed based on moving path and purchase transactions. However, these systems do not take item-specific profit margins into consideration. In addition, few such proposals accounted for location similarity and item similarity. To address these issues, a shopping behavior prediction (SBP) system is proposed consisting of a behavior mining module, a similarity inference module, and a behavior prediction module. The behavior mining module can discover high-utility mobile sequential patterns (UMSPs) using the UMSPL algorithm. In the similarity inference module, store-to-store (StoS) similarities and item-to-item (ItoI) similarities are derived by a proposed similarity inference algorithm. When evaluating StoS and ItoI similarities, the quantities of items purchased are considered. Finally, based on UMSPs, StoS similarities, and ItoI similarities, the behavior prediction module generates a list of shopping suggestions for the target user.

Knowledge modeling based on interval-valued fuzzy rough set and similarity inference: prediction of welding distortion

Knowledge-based modeling is a trend in complex system modeling technology. To extract the process knowledge from an information system, an approach of knowledge modeling based on interval-valued fuzzy rough set is presented in this paper, in which attribute reduction is a key to obtain the simplified knowledge model. Through defining dependency and inclusion functions, algorithms for attribute reduction and rule extraction are obtained. The approximation inference plays an important role in the development of the fuzzy system. To improve the inference mechanism, we provide a method of similarity-based inference in an interval-valued fuzzy environment. Combining the conventional compositional rule of inference with similarity based approximate reasoning, an inference result is deduced via rule translation, similarity matching, relation modification, and projection operation. This approach is applied to the problem of predicting welding distortion in marine structures, and the experimental results validate the effectiveness of the proposed methods of knowledge modeling and similarity-based inference.

Tensor-product kernel-based representation encoding joint MRI view similarity.

To support 3D magnetic resonance image (MRI) analysis, a marginal image similarity (MIS) matrix holding MR inter-slice relationship along every axis view (Axial, Coronal, and Sagittal) can be estimated. However, mutual inference from MIS view information poses a difficult task since relationships between axes are nonlinear. To overcome this issue, we introduce a Tensor-Product Kernel-based Representation (TKR) that allows encoding brain structure patterns due to patient differences, gathering all MIS matrices into a single joint image similarity framework. The TKR training strategy is carried out into a low dimensional projected space to get less influence of voxel-derived noise. Obtained results for classifying the considered patient categories (gender and age) on real MRI database shows that the proposed TKR training approach outperforms the conventional voxel-wise sum of squared differences. The proposed approach may be useful to support MRI clustering and similarity inference tasks, which are required on template-based image segmentation and atlas construction.

Composer Similarities through “The Classical Music Navigator”: Similarity Inference from Composer Influences

Data from “The Classical Music Navigator” ( CMN) website are used to generate statistics bearing on the relative stylistic similarities of 500 classical music composers. When the CMN was compiled, notice was taken of which composers are thought to have influenced which others; it is reasoned here that composers with similar arrays of composer influences should usually themselves be adjudged as similar. Measures of association like the ones more commonly used in fields such as biogeography and biological systematics were applied to pairwise comparisons across the 500 by 500 matrix of composer influences lists, and then examined for the relevant characteristics. Samples of the results using three measures in particular are given.

Similarity-based methods for potential human microRNA-disease association prediction

BackgroundThe identification of microRNA-disease associations is critical for understanding the molecular mechanisms of diseases. However, experimental determination of associations between microRNAs and diseases remains challenging. Meanwhile, target diseases need to be revealed for some new microRNAs without any known target disease association information as new microRNAs are discovered each year. Therefore, computational methods for microRNA-disease association prediction have gained a lot of research interest.MethodsHerein, based on the assumption that functionally related microRNAs tend to be associated with phenotypically similar diseases, three inference methods were presented for microRNA-disease association prediction, namely MBSI (microRNA-based similarity inference), PBSI (phenotype-based similarity inference) and NetCBI (network-consistency-based inference). Global network similarity measure was used in the three methods to predict new microRNA-disease associations.ResultsWe tested the three methods on 242 known microRNA-disease associations by leave-one-out cross-validation for prediction evaluation, and achieved AUC values of 74.83%, 54.02% and 80.66%, respectively. The best-performed method NetCBI was then chosen for novel microRNA-disease association prediction. Some associations strongly predicted by NetCBI were confirmed by the publicly accessible databases, which indicated the usefulness of this method. The newly predicted associations were publicly released to facilitate future studies. Moreover, NetCBI was especially applicable to predicting target diseases for microRNAs whose target association information was not available.ConclusionsThe encouraging results suggest that our method NetCBI can not only provide help in identifying novel microRNA-disease associations but also guide biological experiments for scientific research.

Prediction of Polypharmacological Profiles of Drugs by the Integration of Chemical, Side Effect, and Therapeutic Space

Prediction of polypharmacological profiles of drugs enables us to investigate drug side effects and further find their new indications, i.e. drug repositioning, which could reduce the costs while increase the productivity of drug discovery. Here we describe a new computational framework to predict polypharmacological profiles of drugs by the integration of chemical, side effect, and therapeutic space. On the basis of our previous developed drug side effects database, named MetaADEDB, a drug side effect similarity inference (DSESI) method was developed for drug-target interaction (DTI) prediction on a known DTI network connecting 621 approved drugs and 893 target proteins. The area under the receiver operating characteristic curve was 0.882 ± 0.011 averaged from 100 simulated tests of 10-fold cross-validation for the DSESI method, which is comparative with drug structural similarity inference and drug therapeutic similarity inference methods. Seven new predicted candidate target proteins for seven approved drugs were confirmed by published experiments, with the successful hit rate more than 15.9%. Moreover, network visualization of drug-target interactions and off-target side effect associations provide new mechanism-of-action of three approved antipsychotic drugs in a case study. The results indicated that the proposed methods could be helpful for prediction of polypharmacological profiles of drugs.

The Paradox of Translation

Abstract For a good translation the “interpretant sign” or “translatant” must not only repeat the “interpreted sign”, the “translated sign”, the original text in translation, but also must respond to it with interpretants of “answering comprehension”. In interlingual translation, the relation between the “interpreted” and the “interpretant” is oriented by the logic of responsive creativity, dialogism and otherness, by iconic similarity and abductive inference. Interlingual translation involves the disposition to respond to the word of the other in the form of reported, indirect discourse masked as direct discourse. The paradox of a good translation, which is the same text as the original but other from it, is the paradox of the sign.