京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第21390号 / 情博第676号 / 新制||情||117(附属図書館) / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 鹿島 久嗣, 教授 山本 章博, 教授 下平 英寿 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

No description available.

Indiana University-Purdue University Indianapolis (IUPUI) / The successful treatment and potential eradication of many complex diseases, such as cancer, begins with elucidating the convoluted mapping of molecular profiles to phenotypical manifestation. Our observed molecular profiles (e.g., genomics, transcriptomics, epigenomics) are often high-dimensional and are collected from patient samples falling into heterogeneous disease subtypes. Interpretable learning from such data calls for sparsity-driven models. This dissertation addresses the high dimensionality, sparsity, and heterogeneity issues when analyzing multiple-omics data, where each method is implemented with a concomitant R package. First, we examine challenges in submatrix identification, which aims to find subgroups of samples that behave similarly across a subset of features. We resolve issues such as two-way sparsity, non-orthogonality, and parameter tuning with an adaptive thresholding procedure on the singular vectors computed via orthogonal iteration. We validate the method with simulation analysis and apply it to an Alzheimer’s disease dataset. The second project focuses on modeling relationships between large, matched datasets. Exploring regressional structures between large data sets can provide insights such as the effect of long-range epigenetic influences on gene expression. We present a high-dimensional version of mixture multivariate regression to detect patient clusters, each with different correlation structures of matched-omics datasets. Results are validated via simulation and applied to matched-omics data sets. In the third project, we introduce a novel approach to modeling spatial transcriptomics (ST) data with a spatially penalized multinomial model of the expression counts. This method solves the low-rank structures of zero-inflated ST data with spatial smoothness constraints. We validate the model using manual cell structure annotations of human brain samples. We then applied this technique to additional ST datasets. / 2025-05-22

There exist many real-world applications such as recommendation systems, document retrieval, and computational biology where the correct ordering of instances is of equal or greater importance than predicting the exact value of some discrete or continuous outcome. Learning-to-Rank (LTR) refers to a group of algorithms that apply machine learning techniques to tackle these ranking problems. Despite their empirical success, most existing LTR models are not built to be robust to errors in labeling or annotation, distributional data shift, or adversarial data perturbations. To fill this gap, we develop four LTR frameworks that are robust to various types of perturbations. First, Pairwise Elastic Net Regression Ranking (PENRR) is an elastic-net-based regression method for drug sensitivity prediction. PENRR infers robust predictors of drug responses from patient genomic information. The special design of this model (comparing each drug with other drugs in the same cell line and comparing that drug with itself in other cell lines) significantly enhances the accuracy of the drug prediction model under limited data. This approach is also able to solve the problem of fitting on the insensitive drugs that is commonly encountered in regression-based models. Second, Regression-based Ranking by Pairwise Cluster Comparisons (RRPCC) is a ridge-regression-based method for ranking clusters of similar protein complex conformations generated by an underlying docking program (i.e., ClusPro). Rather than using regression to predict scores, which would equally penalize deviations for either low-quality and high-quality clusters, we seek to predict the difference of scores for any pair of clusters corresponding to the same complex. RRPCC combines these pairwise assessments to form a ranked list of clusters, from higher to lower quality. We apply RRPCC to clusters produced by the automated docking server ClusPro and, depending on the training/validation strategy, we show. improvement by 24%–100% in ranking acceptable or better quality clusters first, and by 15%–100% in ranking medium or better quality clusters first. Third, Distributionally Robust Multi-Output Regression Ranking (DRMRR) is a listwise LTR model that induces robustness into LTR problems using the Distributionally Robust Optimization framework. Contrasting to existing methods, the scoring function of DRMRR was designed as a multivariate mapping from a feature vector to a vector of deviation scores, which captures local context information and cross-document interactions. DRMRR employs ranking metrics (i.e., NDCG) in its output. Particularly, we used the notion of position deviation to define a vector of relevance score instead of a scalar one. We then adopted the DRO framework to minimize a worst-case expected multi-output loss function over a probabilistic ambiguity set that is defined by the Wasserstein metric. We also presented an equivalent convex reformulation of the DRO problem, which is shown to be tighter than the ones proposed by the previous studies. Fourth, Inversion Transformer-based Neural Ranking (ITNR) is a Transformer-based model to predict drug responses using RNAseq gene expression profiles, drug descriptors, and drug fingerprints. It utilizes a Context-Aware-Transformer architecture as its scoring function that ensures the modeling of inter-item dependencies. We also introduced a new loss function using the concept of Inversion and approximate permutation matrices. The accuracy and robustness of these LTR models are verified through three medical applications, namely cluster ranking in protein-protein docking, medical document retrieval, and drug response prediction.

Unmanned vehicles (UVs) are seeing more widespread use in military, scientific, and civil sectors in recent years. These UVs range from unmanned air and ground vehicles to surface and underwater vehicles. Each of these different UVs has its own inherent strengths and weaknesses, from payload to freedom of movement. Research in this field is growing primarily because of the National Defense Act of 2001 mandating that one-third of all military vehicles be unmanned by 2015. Research using small UVs, in particular, is a growing because small UVs can go places that may be too dangerous for humans. Because of the limitations inherent in small UVs, including power consumption and payload, the selection of light weight and low power sensors and processors becomes critical. Low power CMOS cameras and real-time vision processing algorithms can provide fast and reliable information to the UVs. These vision algorithms often require computational power that limits their use in traditional general purpose processors using conventional software. The latest developments in field programmable gate arrays (FPGAs) provide an alternative for hardware and software co-design of complicated real-time vision algorithms. By tracking features from one frame to another, it becomes possible to perform many different high-level vision tasks, including object tracking and following. This thesis describes a vision guidance system for unmanned vehicles in general and the FPGA hardware implementation that operates vision tasks in real-time. This guidance system uses an object following algorithm to provide information that allows the UV to follow a target. The heart of the object following algorithm is real-time rank transform, which transforms the image into a more robust image that maintains the edges found in the original image. A minimum sum of absolute differences algorithm is used to determine the best correlation between frames, and the output of this correlation is used to update the tracking of the moving target. Control code can use this information to move the UV in pursuit of a moving target such as another vehicle.

Let G be a simple graph on n vertices, and let S(G) be the class of all real-valued symmetric nxn matrices whose nonzero off-diagonal entries occur in exactly the positions corresponding to the edges of G. The smallest rank achieved by a matrix in S(G) is called the minimum rank of G, denoted mr(G). The maximum nullity achieved by a matrix in S(G) is denoted M(G). For each graph G, there is an associated minimum rank class, MR(G) consisting of all matrices A in S(G) with rank A = mr(G). Although no restrictions are applied to the diagonal entries of matrices in S(G), sometimes diagonal entries corresponding to specific vertices of G must be zero for all matrices in MR(G). These vertices are known as nil vertices (see [6]). In this paper I discuss some basic results about nil vertices in general and nil vertices in cographs and prove that cographs with a nil vertex of a particular form contain two other nil vertices symmetric to the first. I discuss several open questions relating to these results and a counterexample. I prove that for all cographs G without an induced complete tripartite graph with independent sets all of size 3, the zero-forcing number Z(G), a graph theoretic parameter, is equal to M(G). In fact this result holds for a slightly larger class of cographs and in particular holds for all threshold graphs. Lastly, I prove that the maximum of the minimum ranks of all cographs on n vertices is the floor of 2n/3.

Generative models for creating realistic synthetic graphs constitute a research area that is increasing in popularity, especially as the use of graph data is becoming increasingly common. Generating realistic synthetic graphs enables sharing of the information embedded in graphs without directly sharing the original graphs themselves. This can in turn contribute to an increase of knowledge within several domains where access to data is normally restricted, including the financial system and social networks. In this study, it is examined how existing generative models can be extended to be compatible with directed and weighted graphs, without limiting the models to generating graphs of a specific domain. Several models are evaluated, and all use low-rank approximations to learn structural properties of directed graphs. Additionally, it is evaluated how node embeddings can be used with a regression model to add realistic edge weights to directed graphs. The results show that the evaluated methods are capable of reproducing global statistics from the original directed graphs to a promising degree, without having more than 52% overlap in terms of edges. The results also indicate that realistic directed and weighted graphs can be generated from directed graphs by predicting edge weights using pairs of node embeddings. However, the results vary depending on which node embedding technique is used.

Influence has been studied across many different domains including sociology, statistics, marketing, network theory, psychology, social media, politics, and web search. In each of these domains, being able to measure and rank various degrees of influence has useful applications. For example, measuring influence in web search allows internet users to discover useful content more quickly. However, many of these algorithms measure influence across networks and graphs that are mathematically static. This project explores influence measurement within the context of linear time invariant (LTI) systems. While dynamical networks do have mathematical models for quantifying influence on a node-to-node basis, to the best of our knowledge, there are no proposed mathematical formulations that measure aggregate level influence across an entire dynamical network. The dynamics associated with each link, which can differ from one link to another, add additional complexity to the problem. Because of this complexity, many of the static-graph approaches used in web search do not achieve the desired outcome for dynamical networks. In this work we build upon concepts from PageRank and systems theory introduce two new methods for measuring influence within dynamical networks: 1) Dynamical Responsive Page Rank (DRPR) and 2) Aggregated Targeted Reachability (ATR). We then compare and analyze and compare results with these new methods.

No description available.

No description available.