Publications

This paper presents recent results of studies in application of sequence-based pattern structures and emerging patterns to analysis of demographic sequences in Russia. This study is performed on data of 11 generations from 1930 till 1984 for the panel of three waves of the Russian part of Generation and Gender Survey, which took place in 2004, 2007, and 2011. The main goal is to develop methods of extracting emerging patterns (EP) with the following restrictions: the obtained patterns need to be (closed) frequent contiguous prefixes of the input sequences. These constraints were required by demographers for proper interpretation and understanding of early life course events that lead to adulthood. To fulfil this requirement we used modified FP-trees based on pattern structures of contiguous prefixes. After extraction of EP we use CAEP(Classification by Aggregating Emerging Patterns) classifier to predict gender of respondents using their demographic sequences of the first life course events. The best results in terms of TPR-FPR have been obtained for large values of minimum growth-rate parameter (with some objects left without classification).

This book constitutes the refereed proceedings of the 28th International Conference on Data Analytics and Management in Data Intensive Domains, DAMDID/RCDL 2016, held in Ershovo, Moscow, Russia, in October 2016.

The 16 revised full papers presented together with one invited talk and two keynote papers were carefully reviewed and selected from 57 submissions. The papers are organized in topical sections on semantic modeling in data intensive domains; knowledge and learning management; text mining; data infrastructures in astrophysics; data analysis; research infrastructures; position paper.

In this paper, we discuss a semi-dense  depth map interpolation method based on convolutional neural network. We propose a compact  neural network architecture with loss function defined as Euclidean distance in the feature space of VGG-16 neural network used for deep visual recognition. The suggested solution shows state-of-art performance on synthetic and real datasets. Together with LSD-SLAM, the method could be used to provide a dense depth map for interaction purposes, such as creating a first person game in AR/MR or perception module for autonomous vehicle.

Dualization of a monotone Boolean function on a finite lattice can be represented by transforming the set of its minimal 1 values to the set of its maximal 0 values. In this paper we consider finite lattices given by ordered sets of their meet and join irreducibles (i.e., as a concept lattice of a formal context). We show that in this case dualization is equivalent to the enumeration of so-called minimal hypotheses. In contrast to usual dualization setting, where a lattice is given by the ordered set of its elements, dualization in this case is shown to be impossible in output polynomial time unless P = NP. However, if the lattice is distributive, dualization is shown to be possible in subexponential time.

PRIMARY THERAPY OF EARLY BREAST CANCER

Evidence, Controversies, Consensus 

15th St.Gallen International Breast Cancer Conference

Vienna, Austria, 15–18 March 2017 

 

This paper is devoted to mathematical modelling of the progression and stages of breast cancer. The Consolidated mathematical growth Model of primary tumor (PT) and secondary distant metastases (MTS) in patients with lymph nodes MTS (Stage III) (CoM-III) is proposed as a new research tool. The CoM-III rests on an exponential tumor growth model and consists of a system of determinate nonlinear and linear equations. The CoM-III describes correctly primary tumor growth (parameter T) and distant metastases growth (parameter M, parameter N). The CoM-III model and predictive software: a) detect di erent growth periods of primary tumor and distant metastases in patients with lymph nodes MTS; b) make forecast of the period of the distant metastases appearance in patients with lymph nodes MTS; c) have higher average prediction accuracy than the other tools; d) can improve forecasts on survival of breast cancer and facilitate optimisation of diagnostic tests. The CoM-III enables us, for the rst time, to predict the whole natural history of PT and secondary distant MTS growth of patients with/without lymph nodes MTS on each stage relying only on PT sizes.

Nowadays,  mind  mapping  is  rather  popular  educational  technique.  Like any other learning tools, mind maps became  a part of modern educational  trends like blended learning and computer-supported collaborative learning. Lots  of mind mapping software tools are adopted to teaching and learning routines  such as educational content delivery or assessment. This paper focuses on the  additional automatic evaluation of digital educational mind maps gained from the  existing procedures of assessments. The review of automatic graders which support the evaluation process demonstrates that some systematical work is done in  automation grading by comparing students’ mind maps with a template. But lots  of questions about automatic mind maps’  scoring by retrieving  the  data from a  scored mind map are still open. This paper introduces the automatic grader for  educational mind maps (AGEMM) which acts like a teacher’s assistant and calculates several quantitative metrics. The AGEMM is implemented as a web-service and interacted with mind maps prepared in the Coggle web-service through  its API. The AGEMM is adopted to a bachelor course. Results demonstrate that  scores from the AGEMM may be transformed to scales or criterial levels which  are  used  to  evaluation.  Moreover,  the  AGEMM  application  revealed  several  problems and shew lines of development which we discuss in the paper.

Thispaperreportsonourongoingworkthataimsataclassificationof conjunctive queries q according to the data complexity of answering ontology- mediated queries ({A ⊑ T ⊔ F }, q). We give examples of queries from the complexity classes C ∈ {AC0, L, NL, P, CONP}, and obtain a few syntactical conditions for C-membership and C-hardness. 

We revisit the notion of probably approximately correct implication bases from the literature and present a first formulation in the language of formal concept analysis, with the goal to investigate whether such bases represent a suitable substitute for exact implication bases in practical use cases. To this end, we quantitatively examine the behavior of probably approximately correct implication bases on artificial and real-world data sets and compare their precision and recall with respect to their corresponding exact implication bases. Using a small example, we also provide evidence suggesting that implications from probably approximately correct bases can still represent meaningful knowledge from a given data set.

We consider a propositional language for describing parameterized ceteris paribus preferences over atomic conjunctions. Such preferences are only required to hold when the alternatives being compared agree on a specified subset of propositional variables. Regarding the expressivity of the language in question, we show that a parameterized preference statement is equivalent to a conjunction of an exponential number of classical, non-parameterized, ceteris paribus statements. Next, we present an inference system for parameterized statements and prove that the problem of checking the semantic consequence relation for such statements is coNP-complete. We propose an approach based on formal concept analysis to learning preferences from data by showing that ceteris paribus preferences valid in a particular model correspond to implications of a special formal context derived from this model. The computation of a complete preference set is then reducible to the computation of minimal hypergraph transversals. Finally, we adapt a polynomial-time algorithm for abduction using Horn clauses represented by their characteristic models to the problem of determining preferences over new alternatives from preferences over given alternatives (with ceteris paribus preferences as the underlying model).

In this paper, we present novel winning team predicting models and compare the accuracy of the obtained prediction with TrueSkill model of ranking individual players impact based on their impact in team victory for the two most popular online games: Dota 2 and Counter-Strike: Global Offensive.

The cornerstone of retail banking risk management is the estimation of the expected losses when granting a loan to the borrower. The key driver for loss estimation is probability of default (PD) of the borrower. Assessing PD lies in the area of classification problem. In this paper we apply FCA query-based classification techniques to Kaggle open credit scoring data. We argue that query based classification allows one to achieve higher classification accuracy as compared to applying classical banking models and still to retain interpretability of model results, whereas black-box methods grant better accuracy but diminish interpretability. 

Modern co-authorship networks contain hidden patterns of researchers interaction and publishing activities. We aim to provide a system for selecting a collaborator for joint research or an expert on a given list of topics. We have improved a recommender system for finding possible collaborator with respect to research interests and predicting quality and quantity of the anticipated publications. Our system is based on a co-authorship network derived from the bibliographic database, as well as content information on research papers obtained from SJR Scimago, staff information and the other features from the open data of researchers profiles. We formulate the recommendation problem as a weighted link prediction within the co-authorship network and evaluate its prediction for strong and weak ties in collaborative communities.

With advances of recent technologies, augmented reality systems and autonomous vehicles gained a lot of interest from academics and industry. Both these areas rely on scene geometry understanding, which usually requires depth map estimation. However, in case of systems with limited computational resources, such as smartphones or autonomous robots, high resolution dense depth map estimation may be challenging. In this paper, we study the problem of semi-dense depth map interpolation along with low resolution depth map upsampling. We present an end-to-end learnable residual convolutional neural network architecture that achieves fast interpolation of semi-dense depth maps with different sparse depth distributions: uniform, sparse grid and along intensity image gradient. We also propose a loss function combining classical mean squared error with perceptual loss widely used in intensity image super-resolution and style transfer tasks. We show that with some modifications, this architecture can be used for depth map super-resolution. Finally, we evaluate our results on both synthetic and real data, and consider applications for autonomous vehicles and creating AR/MR video games.

AIST is a scientific conference on Analysis of Images, Social Networks, and Texts. The conference is intended for computer scientists and practitioners whose research interests involve Internet mathematics and other related fields of data science. Similar to the previous year, the conference will be focused on applications of data mining and machine learning techniques to various problem domains: image processing, analysis of social networks, and natural language processing. We hope that the participants will benefit from the interdisciplinary nature of the conference and exchange experience.

As the number of digital texts increases rapidly, there is a pressing need for more advanced and diverse tools of natural language processing. While purely statistical approaches proved powerful and efficient for many NLP tasks, there are many applications that would benefit from the formal models and approaches traditional language science has to offer. With hopes to facilitate this interaction between theory and practical implementation, we are pleased to announce the workshop on Computational Linguistics and Language Science to be held in Moscow, Russia on April 25, 2016 (11 AM to 6 PM).

This paper provides the reader with a report on 9th Russian Summer School in Information Retrieval (RuSSIR 2015).

We propose a new algorithm for consensus clustering, FCA-Consensus, based on Formal Concept Analysis. As the input, the algorithm takes T partitions of a certain set of objects obtained by k-means algorithm after T runs from different initialisations. The resulting consensus partition is extracted from an antichain of the concept lattice built on a formal context objects×classes, where the classes are the set of all cluster labels from each initial k-means partition. We compare the results of the proposed algorithm in terms of ARI measure with the state-of-the-art algorithms on synthetic datasets. Under certain conditions, the best ARI values are demonstrated by FCA-Consensus.