### Systems biology

"... A boosting approach to structure learning of graphs with and without prior knowledge ..."

Abstract
- Add to MetaCart

A boosting approach to structure learning of graphs with and without prior knowledge

### Reconstruction of Biological Networks from Limited and Uneven Reliable Interactions

"... Motivation: An important problem in systems biology is reconstructing complete networks of interactions between biological objects by extrapolating from a few known interactions as examples. While there are many computational techniques proposed for this network reconstruction task, their accuracy i ..."

Abstract
- Add to MetaCart

Motivation: An important problem in systems biology is reconstructing complete networks of interactions between biological objects by extrapolating from a few known interactions as examples. While there are many computational techniques proposed for this network reconstruction task, their accuracy is consistently limited by the small number of high-confidence examples, and the uneven distribution of these examples across the potential interaction space, with some objects having many known interactions and others few. Results: To address this issue, we propose two computational methods based on the concept of training set expansion. They work particularly effectively in conjunction with kernel approaches, which are a popular class of approaches for fusing together many disparate types of features. Both our methods are based on semisupervised learning and involve augmenting the limited number of

### Positive Definite Kernels in Machine Learning

, 2009

"... This survey is an introduction to positive definite kernels and the set of methods they have inspired in the machine learning literature, namely kernel methods. We first discuss some properties of positive definite kernels as well as reproducing kernel Hibert spaces, the natural extension of the set ..."

Abstract
- Add to MetaCart

This survey is an introduction to positive definite kernels and the set of methods they have inspired in the machine learning literature, namely kernel methods. We first discuss some properties of positive definite kernels as well as reproducing kernel Hibert spaces, the natural extension of the set of functions {k(x, ·), x ∈ X} associated with a kernel k defined on a space X. We discuss at length the construction of kernel functions that take advantage of well-known statistical models. We provide an overview of numerous data-analysis methods which take advantage of reproducing kernel Hilbert spaces and discuss the idea of combining several kernels to improve the performance on certain tasks. We also provide a short cookbook of different kernels which are particularly useful for certain datatypes such as images, graphs or speech segments. Remark: This report is a draft. Comments and suggestions will be highly appreciated. Summary We provide in this survey a short introduction to positive definite kernels and the set of methods they have inspired in machine learning, also known as kernel methods. The main idea behind kernel methods is the following. Most datainference tasks aim at defining an appropriate decision function f on a set of objects of interest X. When X is a vector space of dimension d, say R d, linear functions fa(x) = a T x are one of the easiest and better understood choices, notably for regression, classification or dimensionality reduction. Given a positive definite kernel k on X, that is a real-valued function on X × X which quantifies effectively how similar two points x and y are through the value k(x, y), kernel methods are algorithms which estimate functions f of the form f: x ∈ X → f(x) = ∑ αik(xi, x), (1) i∈I

### Structured Prediction of Generalized Matching Graphs

"... A structured prediction approach is proposed for completing missing edges in a graph using partially observed connectivity between n nodes. Unlike previous approaches, edge predictions depend on the node attributes (features) as well as graph topology. To overcome unrealistic i.i.d. edge prediction ..."

Abstract
- Add to MetaCart

A structured prediction approach is proposed for completing missing edges in a graph using partially observed connectivity between n nodes. Unlike previous approaches, edge predictions depend on the node attributes (features) as well as graph topology. To overcome unrealistic i.i.d. edge prediction assumptions, the structured prediction framework is extended to an output space of directed subgraphs that satisfy in-degree and out-degree constraints. An efficient cutting plane algorithm is provided which interleaves the estimation of an edge score function with exact inference of the maximum weight degree-constrained subgraph. Experiments with social networks, protein-protein interaction graphs and citation networks are shown.

### unknown title

"... Vol. 24 ISMB 2008, pages i232–i240 doi:10.1093/bioinformatics/btn162 Prediction of drug–target interaction networks from the integration of chemical and genomic spaces ..."

Abstract
- Add to MetaCart

Vol. 24 ISMB 2008, pages i232–i240 doi:10.1093/bioinformatics/btn162 Prediction of drug–target interaction networks from the integration of chemical and genomic spaces

### unknown title

"... Vol. 24 ISMB 2008, pages i232–i240 doi:10.1093/bioinformatics/btn162 Prediction of drug–target interaction networks from the integration of chemical and genomic spaces ..."

Abstract
- Add to MetaCart

Vol. 24 ISMB 2008, pages i232–i240 doi:10.1093/bioinformatics/btn162 Prediction of drug–target interaction networks from the integration of chemical and genomic spaces

### IEEE Transactions on Knowledge and Data Engineering, vol.22, no.7, pp.957–968, 2010. 1 Conic Programming for Multi-Task Learning

"... When we have several related tasks, solving them simultaneously has been shown to be more effective than solving them individually. This approach is called multi-task learning (MTL). In this paper, we propose a novel MTL algorithm. Our method controls the relatedness among the tasks locally, so all ..."

Abstract
- Add to MetaCart

When we have several related tasks, solving them simultaneously has been shown to be more effective than solving them individually. This approach is called multi-task learning (MTL). In this paper, we propose a novel MTL algorithm. Our method controls the relatedness among the tasks locally, so all pairs of related tasks are guaranteed to have similar solutions. We apply the above idea to support vector machines and show that the optimization problem can be cast as a second-order cone program, which is convex and can be solved efficiently. The usefulness of our approach is demonstrated in ordinal regression, link prediction and collaborative filtering, each of which can be formulated as a structured multi-task problem.

### International Journal of Knowledge Discovery in Bioinformatics, vol.1, no.1, pp.66-80, 2010. A Transfer Learning Approach and Selective Integration of Multiple Types of Assays for Biological Network Inference

"... Inferring the relationship among proteins is a central issue of computational biology and a diversity of biological assays are utilized to predict the relationship. However, as experiments are usually expensive to perform, automatic data selection is employed to reduce the data collection cost. Alth ..."

Abstract
- Add to MetaCart

Inferring the relationship among proteins is a central issue of computational biology and a diversity of biological assays are utilized to predict the relationship. However, as experiments are usually expensive to perform, automatic data selection is employed to reduce the data collection cost. Although data useful for link prediction are different in each local sub-network, existing methods cannot select different data for different processes. This paper presents a new algorithm for inferring biological networks from multiple types of assays. The proposed algorithm is based on transfer learning and can exploit local information effectively. Each assay is automatically weighted through learning and the weights can be adaptively different in each local part. Our algorithm was favorably examined on two kinds of biological networks: a metabolic network and a protein interaction network. A statistical test confirmed that the weight that our algorithm assigned to each assay was meaningful.

### STRUCTURED PRIORS FOR SUPERVISED LEARNING IN COMPUTATIONAL BIOLOGY

"... All Rights ReservedOnly for you, children of doctrine and learning, have we written this work. Examine this book, ponder the meaning we have dispersed in various places and gathered again; what we have concealed in one place we have disclosed in another, that it may be understood by your wisdom. H. ..."

Abstract
- Add to MetaCart

All Rights ReservedOnly for you, children of doctrine and learning, have we written this work. Examine this book, ponder the meaning we have dispersed in various places and gathered again; what we have concealed in one place we have disclosed in another, that it may be understood by your wisdom. H. C. A. von Nettesheim, De occulta philosophia, 3, 65. I have understood. And the certainty that there is nothing to understand should be my peace, my triumph. But I am here, and They are looking for me, thinking I possess the revelation They sordidly desire. It isn’t enough to have understood, if others refuse and continue to interrogate. U. Eco, Foucault’s Pendulum., Malkhut, 120. iiRemerciements Je tiens en premier lieu à remercier Jean-Philippe, qui m’a encadré pendant la durée de cette thèse. Jean-Philippe a toujours su m’indiquer de bonnes directions de recherche quand il le fallait, me laisser chercher seul quand il le fallait, et a ce talent de savoir expliquer les concepts les plus techniques en des termes très intuitifs. Je pense avoir énormément appris à son contact, j’ai sincèrement apprécié de travailler avec lui et lui suis reconnaissant pour le temps qu’il ma consacré et toutes les opportunités qu’il m’a données au cours de cette thèse.