Natural Language Interfaces (NLIs) have transformed data interaction by enabling natural language querying and visualization of tabular data. Despite the growing importance of NLIs, prior research has examined querying and visualization tasks separately, lacking a unified perspective, especially in the era of Large Language Models (LLMs). To fill this gap, this survey provides a comprehensive analysis of NLIs for tabular data, examining their evolution and fundamental components: datasets, evaluation metrics, and architectural designs. By analyzing over 60 approaches and 38 datasets, we explore recent advancements in Text-to-SQL and Text-to-Vis tasks, focusing on semantic parsing techniques for natural language translation to SQL queries and visualization specifications. We evaluate the impact of LLMs on these systems, discussing their capabilities and limitations. Our systematic review serves as a roadmap for developing NLIs in the foundation model era.

The emergence of natural language processing has revolutionized the way users interact with tabular data, enabling a shift from traditional query languages and manual plotting to more intuitive, language-based interfaces. The rise of large language models (LLMs) such as ChatGPT and its successors has further advanced this field, opening new avenues for natural language processing techniques. This survey presents a comprehensive overview of natural language interfaces for tabular data querying and visualization, which allow users to interact with data using natural language queries. We introduce the fundamental concepts and techniques underlying these interfaces with a particular emphasis on semantic parsing, the key technology facilitating the translation from natural language to SQL queries or data visualization commands. We then delve into the recent advancements in Text-to-SQL and Text-to-Vis problems from the perspectives of datasets, methodologies, metrics, and system designs. This includes a deep dive into the influence of LLMs, highlighting their strengths, limitations, and potential for future improvements. Through this survey, we aim to provide a roadmap for researchers and practitioners interested in developing and applying natural language interfaces for data interaction in the era of large language models.

One of the best ways to deal with the problem of knowledge distillation in unstructured text is applying text mining. This machine learning-based approach can provide extracted useful information from large body of texts, in a reasonable time. However, the results are usually in complicated forms, which meant it is a non-trivial task in term of interpretation. In this paper, we present appropriate visualizations and analyses in order to tackle the tangled network representing relationship between entities (i.e. terms extracted from raw text). Conducted on a case study of information extraction in the business management domain, our results reveal the hidden relationships of inter-organizational success factors in a simple structured way. These results can assist companies to understand and generate business strategies, in terms of the collaboration aspect.

Data visualization plays a significant role in investment activities, since it directly provides the decision-relevant information to investors. The traditional Stock Chart provides limited information to investors as it only displays the price changes of the stock over a period of time. More information could be incorporated in the plot such as the return of a stock, the risk of a stock and the performance of the corporation issuing the stock. Based on the information what an investor concerns most about investment activities, a new style visualization method based on Parallel Coordinates is proposed. This paper will describe this visualization technique in the stock chart context and explain the Data Envelopment Analysis (DEA) model used to construct the Parallel Coordinate. In addition, the benefits of the Parallel Coordinates are illustrated.

We present a method to visualize gene co-expression from microarray data by plotting profiles of dissimilarity within gene-sets of biological pathways. A gene co-expression network is created by computing the correlation between each gene pair in a gene-set. We transform the networks into scale-free networks in order to calculate the dissimilarity weights that are used to create our profiles. Our approach further distinguishes between gene pairs consisting of both, one, or no statistically significant genes. We find that the shapes and density of the profiles provide useful information for identification of disease gene biomarkers. Our results provide a means of visualizing the overall distribution of gene dissimilarity for each gene-set, as well as how gene dissimilarity is linked to the mutual significance of gene pairs within a gene-set.

An integrated network-based visualization tool was designed to aid in the analysis of gene expression data, in which modules or subnetworks that are related to one another are extracted. These subnetworks could be relevant to the cause of disease. This tool may provide a mean to gain better understanding of the disease mechanism and help to discover new disease biomarkers for further in-depth analysis.

The identification of predictive biomarkers of complex disease with robustness and specificity is an ongoing challenge. Gene expressions provide information on how the cell reacts to a particular state and the relationship of genes may lead to novel information. A network-based approach integrating expression data with protein-protein interaction network can be used to identify gene-subnetwork biomarkers for a particular disease. However, cancer datasets are heterogeneous in nature containing unknown or undefined subtypes of cancers. In this study, we propose a gene-subnetwork biomarker identification approach by implementing an Expectation-Maximization (EM) clustering technique to homogenize the dataset. To validate our proposed method. Lung cancer expression datasets are used to identify gene-subnetwork biomarkers. The evaluation of gene-subnetwork biomarkers is done by 5-fold cross-validation on an independent dataset. The comparison between non-clustering and clustering-based gene-subnetwork identification showed that clustering produced improved classification performance at a statistically significant level. Furthermore, preliminary functional analysis results showed more significant subnetworks were identified using the proposed approach.

Ten years ago when Asia-Pacific Bioinformatics Network held the first International Conference on Bioinformatics (InCoB) in Bangkok its theme was North-South Networking. At that time InCoB aimed to provide biologists and bioinformatics researchers in the Asia-Pacific region a forum to meet, interact with, and disseminate knowledge about the burgeoning field of bioinformatics. Meanwhile InCoB has evolved into a major regional bioinformatics conference that attracts not only talented and established scientists from the region but increasingly also from East Asia, North America and Europe. Since 2006 InCoB yielded 114 articles in BMC Bioinformatics supplement issues that have been cited nearly 1,000 times to date. In part, these developments reflect the success of bioinformatics education and continuous efforts to integrate and utilize bioinformatics in biotechnology and biosciences in the Asia-Pacific region. A cross-section of research leading from biological data to knowledge and to technological applications, the InCoB2012 theme, is introduced in this editorial. Other highlights included sessions organized by the Pan-Asian Pacific Genome Initiative and a Machine Learning in Immunology competition. InCoB2013 is scheduled for September 18-21, 2013 at Suzhou, China.

The success of Memetic Algorithms (MAs) has driven many researchers to be more focused on the efficiency aspect of the algorithms such that it would be possible to effectively employ MAs to solve computationally expensive optimization problems where single evaluation of the objective and constraint functions may require minutes to hours of CPU time. One of the important design issues in MAs is the choice of the individuals upon which local search procedure should be applied. Selecting only some potential individuals lessens the demand for functional evaluations hence accelerates convergence to the global optimum. In recent years, advances have been made targeting optimization problems with single equality constraint h(x) = 0. The presence of previously evaluated candidate solutions with different signs of constraint values within some localities thus allows the estimation of the constraint boundary. An individual will undergo local search only if it is sufficiently close to the approximated boundary. Elegant as it may seem, the approach had unfortunately assumed that every constraint function maps the design variables to optimize into unbounded real values. This, however, may not always be the case in practice. In this paper, we present a strategy to efficiently solve constrained problems with a single equality constraint; the function of which maps the design variables into restricted (either strictly non-negative or strictly non-positive) real values only.

Microarray-based classification of disease states is based on gene expression profiles of subjects. Various methods have been proposed to identify diagnostic markers that can accurately discriminate between two classes such as case and control. Many of the methods that used only a subset of ranked genes in the pathway may not be able to fully represent the classification boundaries for the two disease classes. The use of negatively correlated feature sets (NCFS) for identifying phenotype-correlated genes (PCOGs) and inferring pathway activities is used here. The NCFS-based pathway activity inference schemes significantly improved the power of pathway markers to discriminate between normal and cancer, as well as relapse and non-relapse, classes in microarray expression datasets of breast cancer. Furthermore, the use of ranker feature selection methods with top 3 pathway markers has been shown to be suitable for both logistic and NB classifiers. In addition, the proposed single pathway classification (SPC) ranker provided similar performance to the traditional SVM and Relief-F feature selection methods. The identification of PCOGs within each pathway, especially with the use of NCFS based on correlation with ideal markers (NCFS-i), helps to minimize the effect of potentially noisy experimental data, leading to accurate and robust classification results.

This paper proposes an effective method for identifying functional modules of the weighted gene co-expression network using a minimum spanning tree (MST) approach coupled with network neighborhood connectivity. The MST-based gene co-expression network was reconstructed to serve as the backbone of gene co-expression network. Highly connected hub genes were identified based on the connectivity of the backbone network. All sub-networks were extracted by expanding from the hub genes to their neighborhood genes. Finally, functional modules were identified by integrating sub-networks with similar gene expression profiles. We tested the method with both simulated and autism spectrum disorder microarray data sets. The results show that our approach is better in highlighting the hub genes and can effectively identify functional modules with highly enriched pathways.

A systematic statistical process is proposed for cDNA microarray data analysis to counter the potential problems of limited replication and unknown distribution found in some public data sets. The proposed process integrates several existing methods to infer expression patterns of unknown genes. It consists of data normalization, identification of significant genes using a novel within-slide replication method coupled with conditional cluster analysis. This process uses intensity-dependent normalization to reduce dye effect. It subsequently uses one-way ANOVA with a very small cutoff p-value to identify individual genes presenting significant expression patterns across the experiments. The proposed process eventually utilizes the identified significant genes to infer expression patterns of some other genes clustered in the same groups by using k-medoids