Discovery Tasks Research Articles

Effects of the association of registers of semiotic representation on the resolution of problems of direct proportionality

This research focuses on the two main registers of semiotic representations in direct proportionality. Based on the theory of registers of semiotic representation, we examine the effects of tabular and graphical registers in solving problems of direct proportionality. We hypothesize that mobilized registers will have an impact on the performance of our learners. Using a factorial design, our study of ninety-two learners in the second year of secondary school (13-14 years old) evaluates the implementation of four pedagogical scenarios characterized by an identical discovery task, but mobilizing different semiotic registers.

Error Rates and Uncertainty Reduction in Rule Discovery

Three new versions of Wason’s 2-4-6 rule discovery task incorporating error rates or feedback of uncertainty reduction, inspired by the error-statistical account in philosophy of science, were employed. In experiments 1 and 2, participants were instructed that some experimenter feedback would be erroneous (control was original 2-4-6 without error). The results showed that performance was impaired when there was probabilistic error. In experiment 3, participants were given uncertainty reduction feedback as they generated different number triples and the negative effects of probabilistic error were not observed. These findings are informative not only about rule discovery tasks in general but also about contexts of inference under uncertainty.

A Scalable Embedding Based Neural Network Method for Discovering Knowledge From Biomedical Literature

Nowadays, the amount of biomedical literatures is growing at an explosive speed, and much useful knowledge is yet undiscovered in the literature. Classical information retrieval techniques allow to access explicit information from a given collection of information, but are not able to recognize implicit connections. Literature-based discovery (LBD) is characterized by uncovering hidden associations in non-interacting literature. It could significantly support scientific research by identifying new connections between biomedical entities. However, most of the existing approaches to LBD are not scalable and may not be sufficient to detect complex associations in non-directly-connected literature. In this article, we present a model which incorporates biomedical knowledge graph, graph embedding, and deep learning methods for literature-based discovery. First, the relations between biomedical entities are extracted from biomedical abstracts and then a knowledge graph is constructed by using these obtained relations. Second, the graph embedding technologies are applied to convert the entities and relations in the knowledge graph into a low-dimensional vector space. Third, a bidirectional Long Short-Term Memory (BLSTM) network is trained based on the entity associations represented by the pre-trained graph embeddings. Finally, the learned model is used for open and closed literature-based discovery tasks. The experimental results show that our method could not only effectively discover hidden associations between entities, but also reveal the corresponding mechanism of interactions. It suggests that incorporating knowledge graph and deep learning methods is an effective way for capturing the underlying complex associations between entities hidden in the literature.

A framework for dependency estimation in heterogeneous data streams

Estimating dependencies from data is a fundamental task of Knowledge Discovery. Identifying the relevant variables leads to a better understanding of data and improves both the runtime and the outcomes of downstream Data Mining tasks. Dependency estimation from static numerical data has received much attention. However, real-world data often occurs as heterogeneous data streams: On the one hand, data is collected online and is virtually infinite. On the other hand, the various components of a stream may be of different types, e.g., numerical, ordinal or categorical. For this setting, we propose Monte Carlo Dependency Estimation (MCDE), a framework that quantifies multivariate dependency as the average statistical discrepancy between marginal and conditional distributions, via Monte Carlo simulations. MCDE handles heterogeneity by leveraging three statistical tests: the Mann–Whitney U, the Kolmogorov–Smirnov and the Chi-Squared test. We demonstrate that MCDE goes beyond the state of the art regarding dependency estimation by meeting a broad set of requirements. Finally, we show with a real-world use case that MCDE can discover useful patterns in heterogeneous data streams.

TransformerCPI: improving compound-protein interaction prediction by sequence-based deep learning with self-attention mechanism and label reversal experiments.

Identifying compound-protein interaction (CPI) is a crucial task in drug discovery and chemogenomics studies, and proteins without three-dimensional structure account for a large part of potential biological targets, which requires developing methods using only protein sequence information to predict CPI. However, sequence-based CPI models may face some specific pitfalls, including using inappropriate datasets, hidden ligand bias and splitting datasets inappropriately, resulting in overestimation of their prediction performance. To address these issues, we here constructed new datasets specific for CPI prediction, proposed a novel transformer neural network named TransformerCPI, and introduced a more rigorous label reversal experiment to test whether a model learns true interaction features. TransformerCPI achieved much improved performance on the new experiments, and it can be deconvolved to highlight important interacting regions of protein sequences and compound atoms, which may contribute chemical biology studies with useful guidance for further ligand structural optimization. https://github.com/lifanchen-simm/transformerCPI.

Extending the small-molecule similarity principle to all levels of biology with the Chemical Checker.

Small molecules are usually compared by their chemical structure, but there is no unified analytic framework for representing and comparing their biological activity. We present the Chemical Checker (CC), which provides processed, harmonized and integrated bioactivity data on ~800,000 small molecules. The CC divides data into five levels of increasing complexity, from the chemical properties of compounds to their clinical outcomes. In between, it includes targets, off-targets, networks and cell-level information, such as omics data, growth inhibition and morphology. Bioactivity data are expressed in a vector format, extending the concept of chemical similarity to similarity between bioactivity signatures. We show how CC signatures can aid drug discovery tasks, including target identification and library characterization. We also demonstrate the discovery of compounds that reverse and mimic biological signatures of disease models and genetic perturbations in cases that could not be addressed using chemical information alone. Overall, the CC signatures facilitate the conversion of bioactivity data to a format that is readily amenable to machine learning methods.

Fully-Unsupervised Embeddings-Based Hypernym Discovery

The hypernymy relation is the one occurring between an instance term and its general term (e.g., “lion” and “animal”, “Italy” and “country”). This paper we addresses Hypernym Discovery, the NLP task that aims at finding valid hypernyms from words in a given text, proposing HyperRank, an unsupervised approach that therefore does not require manually-labeled training sets as most approaches in the literature. The proposed algorithm exploits the cosine distance of points in the vector space of word embeddings, as already proposed by previous state of the art approaches, but the ranking is then corrected by also weighting word frequencies and the absolute level of similarity, which is expected to be similar when measuring co-hyponyms and their common hypernym. This brings us two major advantages over other approaches—(1) we correct the inadequacy of semantic similarity which is known to cause a significant performance drop and (2) we take into accounts multiple words if provided, allowing to find common hypernyms for a set of co-hyponyms—a task ignored in other systems but very useful when coupled with set expansion (that finds co-hyponyms automatically). We then evaluate HyperRank against the SemEval 2018 Hypernym Discovery task and show that, regardless of the language or domain, our algorithm significantly outperforms all the existing unsupervised algorithms and some supervised ones as well. We also evaluate the algorithm on a new dataset to measure the improvements when finding hypernyms for sets of words instead of singletons.

Bioactivity Prediction Based on Matched Molecular Pair and Matched Molecular Series Methods.

Enhancing a compound's biological activity is the central task for lead optimization in small molecules drug discovery. However, it is laborious to perform many iterative rounds of compound synthesis and bioactivity tests. To address the issue, it is highly demanding to develop high quality in silico bioactivity prediction approaches, to prioritize such more active compound derivatives and reduce the trial-and-error process. Two kinds of bioactivity prediction models based on a large-scale structure-activity relationship (SAR) database were constructed. The first one is based on the similarity of substituents and realized by matched molecular pair analysis, including SA, SA_BR, SR, and SR_BR. The second one is based on SAR transferability and realized by matched molecular series analysis, including Single MMS pair, Full MMS series, and Multi single MMS pairs. Moreover, we also defined the application domain of models by using the distance-based threshold. Among seven individual models, Multi single MMS pairs bioactivity prediction model showed the best performance (R2 = 0.828, MAE = 0.406, RMSE = 0.591), and the baseline model (SA) produced the most lower prediction accuracy (R2 = 0.798, MAE = 0.446, RMSE = 0.637). The predictive accuracy could further be improved by consensus modeling (R2 = 0.842, MAE = 0.397 and RMSE = 0.563). An accurate prediction model for bioactivity was built with a consensus method, which was superior to all individual models. Our model should be a valuable tool for lead optimization.

Chemically Interpretable Graph Interaction Network for Prediction of Pharmacokinetic Properties of Drug-Like Molecules

Solubility of drug molecules is related to pharmacokinetic properties such as absorption and distribution, which affects the amount of drug that is available in the body for its action. Computational or experimental evaluation of solvation free energies of drug-like molecules/solute that quantify solubilities is an arduous task and hence development of reliable computationally tractable models is sought after in drug discovery tasks in pharmaceutical industry. Here, we report a novel method based on graph neural network to predict solvation free energies. Previous studies considered only the solute for solvation free energy prediction and ignored the nature of the solvent, limiting their practical applicability. The proposed model is an end-to-end framework comprising three phases namely, message passing, interaction and prediction phases. In the first phase, message passing neural network was used to compute inter-atomic interaction within both solute and solvent molecules represented as molecular graphs. In the interaction phase, features from the preceding step is used to calculate a solute-solvent interaction map, since the solvation free energy depends on how (un)favorable the solute and solvent molecules interact with each other. The calculated interaction map that captures the solute-solvent interactions along with the features from the message passing phase is used to predict the solvation free energies in the final phase. The model predicts solvation free energies involving a large number of solvents with high accuracy. We also show that the interaction map captures the electronic and steric factors that govern the solubility of drug-like molecules and hence is chemically interpretable.

In Silico Prediction of the Dissociation Rate Constants of Small Chemical Ligands by 3D-Grid-Based VolSurf Method.

Accumulated evidence suggests that binding kinetic properties—especially dissociation rate constant or drug-target residence time—are crucial factors affecting drug potency. However, quantitative prediction of kinetic properties has always been a challenging task in drug discovery. In this study, the VolSurf method was successfully applied to quantitatively predict the koff values of the small ligands of heat shock protein 90α (HSP90α), adenosine receptor (AR) and p38 mitogen-activated protein kinase (p38 MAPK). The results showed that few VolSurf descriptors can efficiently capture the key ligand surface properties related to dissociation rate; the resulting models demonstrated to be extremely simple, robust and predictive in comparison with available prediction methods. Therefore, it can be concluded that the VolSurf-based prediction method can be widely applied in the ligand-receptor binding kinetics and de novo drug design researches.

A Bayesian perspective of statistical machine learning for big data

Statistical Machine Learning (SML) refers to a body of algorithms and methods by which computers are allowed to discover important features of input data sets which are often very large in size. The very task of feature discovery from data is essentially the meaning of the keyword ‘learning’ in SML. Theoretical justifications for the effectiveness of the SML algorithms are underpinned by sound principles from different disciplines, such as Computer Science and Statistics. The theoretical underpinnings particularly justified by statistical inference methods are together termed as statistical learning theory. This paper provides a review of SML from a Bayesian decision theoretic point of view—where we argue that many SML techniques are closely connected to making inference by using the so called Bayesian paradigm. We discuss many important SML techniques such as supervised and unsupervised learning, deep learning, online learning and Gaussian processes especially in the context of very large data sets where these are often employed. We present a dictionary which maps the key concepts of SML from Computer Science and Statistics. We illustrate the SML techniques with three moderately large data sets where we also discuss many practical implementation issues. Thus the review is especially targeted at statisticians and computer scientists who are aspiring to understand and apply SML for moderately large to big data sets.

Thresholding Comparison Display Using Graphical User Interface (GUI) with Mobile Notification

Thresholding is a sort of image division that separates protest apportioning an image into a closer view and foundation. The fundamental utilization of image division are medical imaging, object discovery and recognition tasks, content-based image recovery, video surveillance, and automatic traffic control systems. The objective of the thresholding method is to simplify an image into something that is easier to examine. By using MATLAB R2017b as its core programming software, the image will be separated by unused background with uncertainty. This project will describe a few selected thresholding methods such as Fuzzy C-Mean Algorithm’s method, Wolf’s method, Bradley’s method, Bernsen’s method, Triangle’s Method and Deghost’s Method. Each method will experiment with medical and document images. The thresholding method will undergo image quality analysis such as Sensitivity and Specificity. This process is using Graphical User Interface (GUI) and sending messages to phones with Global System for Mobile Notification (GSM) modems is implemented with the aim of allowing researchers to know the best methods by the ease in a convenient manner. This message can send the receiver any time, users only insert telephone number to send their message.

Constrained Truth Discovery

To aggregate useful information among diversified sources, a hotspot research topic called truth discovery has emerged in recent years. Existing truth discovery methods attempt to infer the true attribute values for the entities by identifying and trusting reliable data sources. However, all these methods neglect the relations among different entities, which play important roles in truth discovery task. When reliable data sources cannot provide sufficient information of entities, the true attribute values of these entities can still be inferred by propagating trustworthy information from related entities. Motivated by this, in this paper, we introduce the constrained truth discovery problem. We incorporate denial constraints, a universally quantified first-order logic formalism, into the process of truth discovery. We formulate it as a constrained optimization problem and analyze its hardness. To address the problem, we propose algorithms to partition the entities into disjoint groups, and generate arithmetic constraints for each disjoint group separately. Then, the true attribute values of the entities in each disjoint group are derived by minimizing the objective function under the corresponding arithmetic constraints. Experimental results on both real-world and synthetic datasets demonstrate that the proposed approach achieves good performance even with very few constraints and reliable sources.

An unsupervised approach to disjointness learning based on terminological cluster trees

In the context of the Semantic Web regarded as a Web of Data, research efforts have been devoted to improving the quality of the ontologies that are used as vocabularies to enable complex services based on automated reasoning. From various surveys it emerges that many domains would require better ontologies that include non-negligible constraints for properly conveying the intended semantics. In this respect, disjointness axioms are representative of this general problem: these axioms are essential for making the negative knowledge about the domain of interest explicit yet they are often overlooked during the modeling process (thus affecting the efficacy of the reasoning services). To tackle this problem, automated methods for discovering these axioms can be used as a tool for supporting knowledge engineers in modeling new ontologies or evolving existing ones. The current solutions, either based on statistical correlations or relying on external corpora, often do not fully exploit the terminology. Stemming from this consideration, we have been investigating on alternative methods to elicit disjointness axioms from existing ontologies based on the induction of terminological cluster trees, which are logic trees in which each node stands for a cluster of individuals which emerges as a sub-concept. The growth of such trees relies on a divide-and-conquer procedure that assigns, for the cluster representing the root node, one of the concept descriptions generated via a refinement operator and selected according to a heuristic based on the minimization of the risk of overlap between the candidate sub-clusters (quantified in terms of the distance between two prototypical individuals). Preliminary works have showed some shortcomings that are tackled in this paper. To tackle the task of disjointness axioms discovery we have extended the terminological cluster tree induction framework with various contributions: 1) the adoption of different distance measures for clustering the individuals of a knowledge base; 2) the adoption of different heuristics for selecting the most promising concept descriptions; 3) a modified version of the refinement operator to prevent the introduction of inconsistency during the elicitation of the new axioms. A wide empirical evaluation showed the feasibility of the proposed extensions and the improvement with respect to alternative approaches.

In vitro, ex vivo and in vivo methods of lung absorption for inhaled drugs.

The assessment and prediction of lung absorption and disposition are an increasingly essential preclinical task for successful discovery and product development of inhaled drugs for both local and systemic delivery. Hence, in vitro, ex vivo and in vivo preclinical methods of lung absorption continue to evolve with several technical, methodological and analytical refinements. As in vitro lung epithelial cell monolayer models, the air-liquid interface (ALI)-cultured Calu-3 cells have most frequently been used, but the NCI-H441 and hAELVi cells have now been proposed as the first immortalized human alveolar epithelial cells capable of forming highly-restricted monolayers. The primary ALI-cultured three-dimensional (3D) human lung cell barriers have also become available; efforts to incorporate aerosol drug deposition into the in vitro lung cell models continue; and stem cell-derived lung epithelial cells and "lung-on-a-chip" technology are emerging. The ex vivo isolated perfused rat lung (IPRL) methods have increasing been used, as they enable the kinetic determination of tissue/organ-level diffusive and membrane protein-mediated absorption and competing non-absorptive loss; the assessment of "pre-epithelial" aerosol biopharmaceutical events in the lung, such as dissolution and release; and the ex vivo-to-in vivo extrapolation and prediction. Even so, in vivo small rodent-based methods have been of mainstay use, while large animal-based methods find an additional opportunity to study region-dependent lung absorption and disposition. It is also exciting that human pharmacokinetic (PK) profiles and systemic exposures for inhaled drugs/molecules may be able to be predicted from these in vivo rodent PK data following lung delivery using kinetic modeling approach with allometric scaling. Overall, the value of these preclinical assessments appears to have shifted more to their translational capability of predicting local lung and systemic exposure in humans, in addition to rationalizing optimal inhaled dosage form and delivery system for drugs/molecules in question. It is critically important therefore to make appropriate selection and timely exploitation of the best models at each stage of drug discovery and development program for efficient progress toward product approval and clinical use.

Multimodal Word Discovery and Retrieval With Spoken Descriptions and Visual Concepts

In the absence of dictionaries, translators, or grammars, it is still possible to learn some of the words of a new language by listening to spoken descriptions of images. If several images, each containing a particular visually salient object, each co-occur with a particular sequence of speech sounds, we can infer that those speech sounds are a word whose definition is the visible object. A multimodal word discovery system accepts, as input, a database of spoken descriptions of images (or a set of corresponding phone transcriptions) and learns a mapping from waveform segments (or phone strings) to their associated image concepts. In this article, four multimodal word discovery systems are demonstrated: three models based on statistical machine translation (SMT) and one based on neural machine translation (NMT). The systems are trained with phonetic transcriptions, MFCC and multilingual bottleneck features (MBN). On the phone-level, the SMT outperforms the NMT model, achieving a 61.6% F1 score in the phone-level word discovery task on Flickr30k. On the audio-level, we compared our models with the existing ES-KMeans algorithm for word discovery and present some of the challenges in multimodal spoken word discovery.

RCorp: a resource for chemical disease semantic extraction in Chinese

BackgroundTo robustly identify synergistic combinations of drugs, high-throughput screenings are desirable. It will be of great help to automatically identify the relations in the published papers with machine learning based tools. To support the chemical disease semantic relation extraction especially for chronic diseases, a chronic disease specific corpus for combination therapy discovery in Chinese (RCorp) is manually annotated.MethodsIn this study, we extracted abstracts from a Chinese medical literature server and followed the annotation framework of the BioCreative CDR corpus, with the guidelines modified to make the combination therapy related relations available. An annotation tool was incorporated to the standard annotation process.ResultsThe resulting RCorp consists of 339 Chinese biomedical articles with 2367 annotated chemicals, 2113 diseases, 237 symptoms, 164 chemical-induce-disease relations, 163 chemical-induce-symptom relations, and 805 chemical-treat-disease relations. Each annotation includes both the mention text spans and normalized concept identifiers. The corpus gets an inter-annotator agreement score of 0.883 for chemical entities, 0.791 for disease entities which are measured by F score. And the F score for chemical-treat-disease relations gets 0.788 after unifying the entity mentions.ConclusionsWe extracted and manually annotated a chronic disease specific corpus for combination therapy discovery in Chinese. The result analysis of the corpus proves its quality for the combination therapy related knowledge discovery task. Our annotated corpus would be a useful resource for the modelling of entity recognition and relation extraction tools. In the future, an evaluation based on the corpus will be held.

Unsupervised Speech Representation Learning Using WaveNet Autoencoders

We consider the task of unsupervised extraction of meaningful latent representations of speech by applying autoencoding neural networks to speech waveforms. The goal is to learn a representation able to capture high level semantic content from the signal, e.g.\ phoneme identities, while being invariant to confounding low level details in the signal such as the underlying pitch contour or background noise. Since the learned representation is tuned to contain only phonetic content, we resort to using a high capacity WaveNet decoder to infer information discarded by the encoder from previous samples. Moreover, the behavior of autoencoder models depends on the kind of constraint that is applied to the latent representation. We compare three variants: a simple dimensionality reduction bottleneck, a Gaussian Variational Autoencoder (VAE), and a discrete Vector Quantized VAE (VQ-VAE). We analyze the quality of learned representations in terms of speaker independence, the ability to predict phonetic content, and the ability to accurately reconstruct individual spectrogram frames. Moreover, for discrete encodings extracted using the VQ-VAE, we measure the ease of mapping them to phonemes. We introduce a regularization scheme that forces the representations to focus on the phonetic content of the utterance and report performance comparable with the top entries in the ZeroSpeech 2017 unsupervised acoustic unit discovery task.

Combining URL and HTML Features for Entity Discovery in the Web

The web is a large repository of entity-pages. An entity-page is a page that publishes data representing an entity of a particular type, for example, a page that describes a driver on a website about a car racing championship. The attribute values published in the entity-pages can be used for many data-driven companies, such as insurers, retailers, and search engines. In this article, we define a novel method, called SSUP , which discovers the entity-pages on the websites. The novelty of our method is that it combines URL and HTML features in a way that allows the URL terms to have different weights depending on their capacity to distinguish entity-pages from other pages, and thus the efficacy of the entity-page discovery task is increased. SSUP determines the similarity thresholds on each website without human intervention. We carried out experiments on a dataset with different real-world websites and a wide range of entity types. SSUP achieved a 95% rate of precision and 85% recall rate. Our method was compared with two state-of-the-art methods and outperformed them with a precision gain between 51% and 66%.

Fuzzy functional dependencies and linguistic interpretations employed in knowledge discovery tasks from relational databases

Knowledge discovery from databases copes with several problems including the heterogeneity of data and interpreting the solution in an understandable and convenient form for domain experts. Fuzzy logic approaches based on the computing with words paradigm are very appealing since they offer the possibility to express useful knowledge from a large volume of data by linguistic terms, which are easily understandable for diverse users. In this paper, the novel descriptive data mining algorithm based on fuzzy functional dependencies has been proposed. In the first step, data are fuzzified, which ensures the same manipulation of crisp and fuzzy data. The data mining step is based on revealing fuzzy functional dependencies among considered attributes. In the final step, the mined knowledge is interpreted linguistically by the fuzzy modifiers and quantifiers. The proposed algorithm has been explained on illustrative data and tested on real-world dataset. Finally, its benefits, weak points and possible future research topics are discussed.