Similarity Search Research Articles

Graph Contrastive Multi-view Learning: A Pre-training Framework for Graph Classification

Recent advancements in node and graph classification tasks can be attributed to the implementation of contrastive learning and similarity search. Despite considerable progress, these approaches present challenges. The integration of similarity search introduces an additional layer of complexity to the model. At the same time, applying contrastive learning to non-transferable domains or out-of-domain datasets results in less competitive outcomes. In this work, we propose maintaining domain specificity for these tasks, which has demonstrated the potential to improve performance by eliminating the need for additional similarity searches. We adopt a fraction of domain-specific datasets for pre-training purposes, generating augmented pairs that retain structural similarity to the original graph, thereby broadening the number of views. This strategy involves a comprehensive exploration of optimal augmentations to devise multi-view embeddings. An evaluation protocol, which focuses on error minimization, accuracy enhancement, and overfitting prevention, guides this process to learn inherent, transferable structural representations that span diverse datasets. We combine pre-trained embeddings and the source graph as a beneficial input, leveraging local and global graph information to enrich downstream tasks. Furthermore, to maximize the utility of negative samples in contrastive learning, we extend the training mechanism during the pre-training stage. Our method consistently outperforms comparative baseline approaches in comprehensive experiments conducted on benchmark graph datasets of varying sizes and characteristics, establishing new state-of-the-art results.

Identifying genes within pathways in unannotated genomes with PaGeSearch.

In biological research, the identification and comparison of genes within specific pathways across the genomes of various species are invaluable. However, annotating the entire genome is resource intensive, and sequence similarity searches often yield results that are not actually genes. To address these limitations, we introduce Pathway Gene Search (PaGeSearch), a tool designed to identify genes from predefined lists, especially those in specific pathways, within genomes. The tool uses an initial sequence similarity search to identify relevant genomic regions, followed by targeted gene prediction and neural network-based result filtering. PaGeSearch suggests the regions that are most likely the orthologs of the genes in the query and is designed to be applicable for species within five classes: mammals, fish, birds, eudicotyledons, and Liliopsida. Compared with GeMoMa and miniprot, PaGeSearch generally outperforms in terms of sensitivity and positive predictive value, as well as negative predictive value. Also, the exon coverage of gene models from PaGeSearch is higher compared with those in GeMoMa and miniprot. Although its performance shows increased variability when applied to actual biological pathways, it nonetheless maintains an acceptable level of accuracy. Evaluating PaGeSearch across different assembly levels, chromosome, scaffold, and contig shows minimal variation in outcomes, indicating that PaGeSearch is resilient to variations in assembly quality.

Systematic analysis of jellyfish galaxy candidates in Fornax, Antlia, and Hydra from the S-PLUS survey: a self-supervised visual identification aid

ABSTRACT We study 51 jellyfish galaxy candidates in the Fornax, Antlia, and Hydra clusters. These candidates are identified using the JClass scheme based on the visual classification of wide-field, twelve-band optical images obtained from the Southern Photometric Local Universe Survey. A comprehensive astrophysical analysis of the jellyfish (JClass > 0), non-jellyfish (JClass = 0), and independently organized control samples is undertaken. We develop a semi-automated pipeline using self-supervised learning and similarity search to detect jellyfish galaxies. The proposed framework is designed to assist visual classifiers by providing more reliable JClasses for galaxies. We find that jellyfish candidates exhibit a lower Gini coefficient, higher entropy, and a lower 2D Sérsic index as the jellyfish features in these galaxies become more pronounced. Jellyfish candidates show elevated star formation rates (including contributions from the main body and tails) by $\sim$1.75 dex, suggesting a significant increase in the SFR caused by the ram-pressure stripping phenomenon. Galaxies in the Antlia and Fornax clusters preferentially fall towards the cluster’s centre, whereas only a mild preference is observed for Hydra galaxies. Our self-supervised pipeline, applied in visually challenging cases, offers two main advantages: it reduces human visual biases and scales effectively for large data sets. This versatile framework promises substantial enhancements in morphology studies for future galaxy image surveys.

Detecting galaxy tidal features using self-supervised representation learning

ABSTRACT Low surface brightness substructures around galaxies, known as tidal features, are a valuable tool in the detection of past or ongoing galaxy mergers, and their properties can answer questions about the progenitor galaxies involved in the interactions. The assembly of current tidal feature samples is primarily achieved using visual classification, making it difficult to construct large samples and draw accurate and statistically robust conclusions about the galaxy evolution process. With upcoming large optical imaging surveys such as the Vera C. Rubin Observatory’s Legacy Survey of Space and Time, predicted to observe billions of galaxies, it is imperative that we refine our methods of detecting and classifying samples of merging galaxies. This paper presents promising results from a self-supervised machine learning model, trained on data from the Ultradeep layer of the Hyper Suprime-Cam Subaru Strategic Program optical imaging survey, designed to automate the detection of tidal features. We find that self-supervised models are capable of detecting tidal features, and that our model outperforms previous automated tidal feature detection methods, including a fully supervised model. An earlier method applied to real galaxy images achieved 76 per cent completeness for 22 per cent contamination, while our model achieves considerably higher (96 per cent) completeness for the same level of contamination. We emphasize a number of advantages of self-supervised models over fully supervised models including maintaining excellent performance when using only 50 labelled examples for training, and the ability to perform similarity searches using a single example of a galaxy with tidal features.

Poaceascoma zborayi sp. nov. and Agrorhizomyces patris gen. et spec. nov. – Two novel dark septate endophytes colonizing wheat (Triticum aestivum) roots from a cropland in Hungary

In this study, we investigated two distinct new phylogenetic lineages of root-colonizing dark septate endophytic fungi colonizing wheat (Triticum aestivum) roots from a long-term agricultural experimental site in Hungary. According to four-locus (internal transcribed spacer, partial large and small subunit regions of nuclear ribosomal DNA, and partial translation elongation factor 1-alpha) phylogenetic analyses, the isolates belong to the Lentitheciaceae and Didymosphaeriaceae of the Pleosporales (Dothideomycetes). We studied the morphology and culture characteristics of the strains. We carried out in vitro resynthesis pot experiments with their original hosts and found no overall negative effect of the inoculation with different isolates of the new taxa. One of the lineages belonged to the genus Poaceascoma (Lentitheciaceae) and represented a novel species described here as Poaceascoma zborayi. We could describe conidia-like structures from this species. Isolates of the other lineage represented a monotypic novel genus in the Didymosphaeriaceae. Accordingly, the new genus, Agrorhizomyces, represented by the species A. patris, is introduced. Sterile, globose structures resembling immature sporocarps were detected. Sequence similarity searches indicated that P. zborayi might be widely distributed, while no sequence similar to A. patris was found outside the sampling area.

Radio Galaxy Zoo: Leveraging latent space representations from variational autoencoder

We propose to learn latent space representations of radio galaxies, and train a very deep variational autoencoder (VDVAE) on RGZ DR1, an unlabeled dataset, to this end. We show that the encoded features can be leveraged for downstream tasks such as classifying galaxies in labeled datasets, and similarity search. Results show that the model is able to reconstruct its given inputs, capturing the salient features of the latter. We use the latent codes of galaxy images, from MiraBest Confident and FR-DEEP NVSS datasets, to train various non-neural network classifiers. It is found that the latter can differentiate FRI from FRII galaxies achieving accuracy ≥ 76%, roc-auc ≥ 0.86, specificity ≥ 0.73 and recall ≥ 0.78 on MiraBest Confident dataset, comparable to results obtained in previous studies. The performance of simple classifiers trained on FR-DEEP NVSS data representations is on par with that of a deep learning classifier (CNN based) trained on images in previous work, highlighting how powerful the compressed information is. We successfully exploit the learned representations to search for galaxies in a dataset that are semantically similar to a query image belonging to a different dataset. Although generating new galaxy images (e.g. for data augmentation) is not our primary objective, we find that the VDVAE model is a relatively good emulator. Finally, as a step toward detecting anomaly/novelty, a density estimator — Masked Autoregressive Flow (MAF) — is trained on the latent codes, such that the log-likelihood of data can be estimated. The downstream tasks conducted in this work demonstrate the meaningfulness of the latent codes.

P‐59: Deep Learning‐based Defect Map Classification and Similarity Search in Display Manufacturing

Defect map in glass substrates provides valuable information for process management, but statistical mapping classification using the number of defects by location within the glass has limited utility due to difficulties in classifying various types and low accuracy. A deep learning method is proposed to classify defect maps in glass substrates for display manufacturing. The method converts the defect maps into colored images and applies argumentation based on location‐based symmetry information. The method achieves 96.1% accuracy and 88.6% F1 score for minority defect types, and enables similarity search for quality management.

GTS: GPU-based Tree Index for Fast Similarity Search

Similarity search, the task of identifying objects most similar to a given query object under a specific metric, has gathered significant attention due to its practical applications. However, the absence of coordinate information to accelerate similarity search and the high computational cost of measuring object similarity hinder the efficiency of existing CPU-based methods. Additionally, these methods struggle to meet the demand for high throughput data management. To address these challenges, we propose GTS, a GPU-based tree index designed for the parallel processing of similarity search in general metric spaces, where only the distance metric for measuring object similarity is known. The GTS index utilizes a pivot-based tree structure to efficiently prune objects and employs list tables to facilitate GPU computing. To efficiently manage concurrent similarity queries with limited GPU memory, we have developed a two-stage search method that combines batch processing and sequential strategies to optimize memory usage. The paper also introduces an effective update strategy for the proposed GPU-based index, encompassing streaming data updates and batch data updates. Additionally, we present a cost model to evaluate search performance. Extensive experiments on five real-life datasets demonstrate that GTS achieves efficiency gains of up to two orders of magnitude over existing CPU baselines and up to 20x efficiency improvements compared to state-of-the-art GPU-based methods.

Multi‐label logo recognition and retrieval based on weighted fusion of neural features

AbstractClassifying logo images is a challenging task as they contain elements such as text or shapes that can represent anything from known objects to abstract shapes. While the current state of the art for logo classification addresses the problem as a multi‐class task focusing on a single characteristic, logos can have several simultaneous labels, such as different colours. This work proposes a method that allows visually similar logos to be classified and searched from a set of data according to their shape, colour, commercial sector, semantics, general characteristics, or a combination of features selected by the user. Unlike previous approaches, the proposal employs a series of multi‐label deep neural networks specialized in specific attributes and combines the obtained features to perform the similarity search. To delve into the classification system, different existing logo topologies are compared and some of their problems are analysed, such as the incomplete labelling that trademark registration databases usually contain. The proposal is evaluated considering 76,000 logos (seven times more than previous approaches) from the European Union Trademarks dataset, which is organized hierarchically using the Vienna ontology. Overall, experimentation attains reliable quantitative and qualitative results, reducing the normalized average rank error of the state‐of‐the‐art from 0.040 to 0.018 for the Trademark Image Retrieval task. Finally, given that the semantics of logos can often be subjective, graphic design students and professionals were surveyed. Results show that the proposed methodology provides better labelling than a human expert operator, improving the label ranking average precision from 0.53 to 0.68.

CASTpFold: Computed Atlas of Surface Topography of the universe of protein Folds.

Geometric and topological properties of protein structures, including surface pockets, interior cavitiesand cross channels, are of fundamental importance for proteins to carry out their functions. Computed Atlas of Surface Topography of proteins(CASTp) is a widely used web server for locating, delineating, and measuring these geometric and topological properties of protein structures. Recent developments in AI-based protein structure prediction such as AlphaFold2 (AF2) have significantly expanded our knowledge on protein structures. Here we present CASTpFold, a continuation of CASTp that provides accurate and comprehensive identifications and quantifications of protein topography. It now provides (i) results on an expanded database of proteins, including the Protein Data Bank(PDB) and non-singleton representative structures of AlphaFold2 structures, covering 183 million AF2 structures; (ii) functional pockets prediction with corresponding Gene Ontology (GO) terms or Enzyme Commission (EC) numbers for AF2-predicted structuresand (iii) pocket similarity search function for surface and protein-protein interface pockets. The CASTpFold web server is freely accessible at https://cfold.bme.uic.edu/castpfold/.

Anti-inflammatory action of new hybrid N-acyl-[1,2]dithiolo-[3,4-c]quinoline-1-thione

ABSTRACT Most of pharmaceutical agents display a number of biological activities. It is obvious that testing even one compound for thousands of biological activities is not practically possible. A computer-aided prediction is therefore the method of choice in this case to select the most promising bioassays for particular compounds. Using the PASS Online software, we determined the probable anti-inflammatory action of the 12 new hybrid dithioloquinolinethiones derivatives. Chemical similarity search in the World-Wide Approved Drugs (WWAD) and DrugBank databases did not reveal close structural analogues with the anti-inflammatory action. Experimental testing of anti-inflammatory activity of the synthesized compounds in the carrageenan-induced inflammation mouse model confirmed the computational predictions. The anti-inflammatory activity of the studied compounds (2a, 3a-3k except for 3j) varied between 52.97% and 68.74%, being higher than the reference drug indomethacin (47%). The most active compounds appeared to be 3h (68.74%), 3k (66.91%) and 3b (63.74%) followed by 3e (61.50%). Thus, based on the in silico predictions a novel class of anti-inflammatory agents was discovered.

Screening of natural epigenetic modifiers for managing glycemic memory and diabetic nephropathy

Short hyperglycaemic episodes trigger metabolic memory (MM) in which managing hyperglycaemia alone is not enough to tackle the progression of Diabetic nephropathy on the epigenetic axis. We used a structural similarity search approach to identify phytochemicals similar to natural epigenetic modifiers and docked with SIRT1 protein and did ADME studies. We found that UMB was 84.3% similar to esculetin. Upon docking, we found that UMB had a binding energy of −9.2 kcal/mol while the standard ligand had −11.8 kcal/mol. ADME showed UMB to be a good lead. 2,2-diphenyl-1-picrylhydrazyl (DPPH) assay showed it to be a good antioxidant with IC50 of 107 µg/mL and MTT stands for 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) showed that it does not promote cell death. Oxidative biomarkers in vitro showed UMB was able to ameliorate glycemic memory induced by high glucose. Western blot revealed decreased histone acetylation under hyperglycaemic conditions and upon treatment with UMB along with DR, its levels increased. This led us to check our hypothesis of whether concomitant diet reversal (DR) together with UMB can alleviate high-fat diet-induced metabolic memory and diabetic nephropathy (DN) in SD rats. UMB was able to decrease blood glucose, lipid, renal, and liver profile concluding UMB was able to ameliorate DN and MM by increasing the histone acetylation level.

Music Information Retrieval using Deep Learning Techniques

Music Information Retrieval (MIR) is gaining attention due to the surge in digital music and the need for efficient search and recommendation systems. Traditional MIR methods rely on hand-crafted features and rule-based systems, limiting their adaptability. Deep Learning (DL) shows promise in automatically extracting complex patterns from raw data. This paper offers an extensive overview of MIR tasks like classification, genre recognition, similarity search, and recommendation, along with DL models like Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), including LSTM and Transformer architectures, tailored for MIR. Challenges such as data scarcity, computational complexity, and interpretability persist, with proposed solutions like data augmentation, transfer learning, and attention mechanisms. Experimental results on benchmark datasets demonstrate DL's superiority in accuracy, scalability, and robustness over traditional methods. Practical examples highlight DL's potential to revolutionize music search, recommendation, and analysis. Emphasizing the importance of large annotated datasets for training high-quality DL models, strategies for data collection, labeling, and preprocessing are outlined. DL offers promising prospects for advancing MIR by addressing its inherent challenges and establishing new performance benchmarks. Further DL development is expected to drive innovation and enhance digital music consumption experiences through MIR systems. Keywords: Deep Learning, Convolution Neural Network (CNNs), Recurrent Neural Network (RNNs), Digital Music Consumption, Genre recognition.

Crossing Linguistic Barriers: Authorship Attribution in Sinhala Texts

Authorship attribution involves determining the original author of an anonymous text from a pool of potential authors. The author attribution task has applications in several domains, such as plagiarism detection, digital text forensics, and information retrieval. While these applications extend beyond any single language, existing research has predominantly centered on English, posing challenges for application in languages such as Sinhala due to linguistic disparities and a lack of language processing tools. We present the first comprehensive study on cross-topic authorship attribution for Sinhala texts and propose a solution that can effectively perform the authorship attribution task even if the topics within the test and training samples differ. Our solution consists of three main parts: (i) extraction of topic-independent stylometric features, (ii) generation of a small candidate author set with the help of similarity search, and (iii) identification of the true author. Several experimental studies were carried out to demonstrate that the proposed solution can effectively handle real-world scenarios involving a large number of candidate authors and a limited number of text samples for each candidate author.

Prison Nurseries: A Review of Maternal and Infant Rooming in Outcomes for Incarcerated Mothers

Women are the fastest growing prison population in Canada, and at incarceration, 4-10 percent of women are pregnant. These women, their correctional facilities, and Canadian health care services are increasingly forced to address the issues of motherhood and reproductive health care during incarceration. Most incarcerated women are separated from their infants soon after birth. The authors claim that prison nurseries, as a harm reduction strategy, are a positive alternative to this separation. Midwives could play a valuable role in these health care units. Methods: This paper is a literature review examining the outcomes of mother-infant dyads who have access to prison nurseries. The search strategy included 15 health research databases, applying similar search terms to all databases. This article has been peer reviewed.

DIDS: Double Indices and Double Summarizations for Fast Similarity Search

Data series has been one of the significant data forms in various applications. It becomes imperative to devise a data series index that supports both approximate and exact similarity searches for large data series collections in high-dimensional metric spaces. The state-of-the-art works employ summarizations and indices to reduce the accesses to the data series. However, we discover two significant flaws that severely limit performance enhancement. Firstly, the state-of-the-art works often employ segment-based summarizations, whose lower bound distances decrease significantly when representing a data series collection, resulting in numerous invalid accesses. Secondly, the disk-based indices for the exact search mainly rely on tree-based indices, which results in low-quality approximate answers, consequently impacting the exact search. To address these problems, we propose a novel solution, Double Indices and Double Summarizations (DIDS). Besides segment-based summarizations, DIDS introduces reference-point-based summarizations to improve the pruning rate by the sorted-based representation strategy. Moreover, DIDS employs reference points and a cost model to cluster similar data series, and uses a graph-based approach to interconnect various regions, enhancing approximate search capabilities. We conduct experiments on extensive datasets, validating the superior search performance of DIDS.

Safety evaluation of the food enzyme cellobiose phosphorylase from the genetically modified Escherichia coli strain LE1B109-pPB130.

The food enzyme cellobiose phosphorylase (cellobiose: phosphate α-d-glucosyltransferase; EC 2.4.1.20) is produced with the genetically modified Escherichia coli strain LE1B109-pPB130 by c-LEcta GmbH. The genetic modifications do not give rise to safety concerns. The food enzyme is considered free from viable cells of the production organism and its DNA. It is intended to be used in combination with a sucrose phosphorylase in the production of the specialty carbohydrate cellobiose. Since residual amounts of total organic solids are removed by downstream purification steps, the Panel considered that toxicological studies other than assessment of allergenicity were unnecessary and a dietary exposure was not estimated. A search for similarity of the amino acid sequence of the food enzyme to known allergens was made and no match was found. The Panel considered that, under the intended conditions of use, the risk of allergic reactions upon dietary exposure cannot be excluded, but the likelihood is low. Based on the data provided, the Panel concluded that this food enzyme does not give rise to safety concerns under the intended conditions of use.

Few-shot learning for similarity search in 12-lead ECG with deep Siamese networks

Few-shot learning for similarity search in 12-lead ECG with deep Siamese networks

From Text to Recommendations: How Vector Databases are Revolutionizing Personalized Content Delivery

Abstract: Effective similarity search and retrieval are now possible thanks to vector databases, which have become a potent tool for organizing and searching high-dimensional data. This article provides a thorough examination of vector databases, their underlying theories, and their applications across a range of industries. In handling complex data types, we address the significance of vector representation and emphasize the benefits of vector databases over traditional databases [1]. The article explores the process of creating a vector database, highlighting the critical function of indexing strategies such as IVF (Inverted File Indexing) and HNSW (Hierarchical Navigable Small World) in guaranteeing the effectiveness and precision of searches [2]. In addition, we discuss the problems caused by the curse of dimensionality and offer solutions to lessen its effects on nearest neighbor searches [3]

Повышение эффективности методов подбора персонала на основе глубоких нейронных сетей

The recruitment industry is at an inflection point: the integration of artificial intelligence has already made its impact on traditional recruitment processes and has the potential to revolutionize it. This article presents an approach to classify resumes into job categories, using semantic similarity search to improve the candidate selection mechanism in recruiting. Our method differs from traditional keyword-based systems and is a deep learning framework that understands and processes the complex semantics of work-related documents. The purpose of the study is to develop a method for classifying resume texts with a complex organizational structure. This study solves several problems at once: increasing the accuracy of resume classification and finding the most stable model for solving the problem of resume classification. We compared standard machine learning methods with neural network ones and showed the effectiveness of the latter. The results indicate an improvement over traditional ML models, suggesting an approach that can be used for pre-screening artificial intelligence recruiting that selects suitable candidates from other applicants. Further, we discovered problems with instability of results when retraining large language models, when the model, even with the same values of the hyperparameters, gives different results. To better understand this phenomenon, we conducted a series of experiments with the main BERT models, varying two parameters – learning rate and seed. As a result, we find a significant increase in performance at a certain threshold parameter, and we quantify which of the found models perform better.