Jan. 9 
Unsupervised Machine Learning for Matrix Decomposition Abstract: Unsupervised learning is a classical approach in pattern recognition and data analysis. Its importance is growing today, due to the increasing data volumes and the difficulty of obtaining statistically sufficient amounts of labelled training data. Typical analysis techniques using unsupervised learning are principal component analysis, independent component analysis, and cluster analysis. They can all be presented as decompositions of the data matrix containing the unlabeled samples. Starting from the classical results, the author reviews some advances in the field up to the present day. Speaker: Erkki Oja Affiliation: Professor Emeritus, Aalto University Place of Seminar: Aalto University 

Jan. 16 
Probabilistic Programming: Bayesian Modeling Made Easy Abstract: Probabilistic models are principled tools for understanding data, but difficulty of inference limits the complexity of models we can actually use. Often we need to develop specific inference algorithms for new models (which might take months), and need to restrict ourselves to tractable model families that might not match our beliefs about the data. Probabilistic programming promises to fix this, by separating the model description from the inference: With probabilistic programming languages we can specify complex models using a highlevel programming language, letting a blackbox inference engine take care of the tricky details. This talk covers the basic idea of probabilistic programming and discusses how well its promises hold now and in the future. Speaker: Arto Klami Affiliation: Academy Research Fellow, University of Helsinki Place of Seminar: University of Helsinki 

Jan. 23 
Metabolite Identification Through Machine Learning Abstract: Identification of small molecules from biological samples remains a major bottleneck in understanding the inner working of biological cells and their environment. Machine learning on data from large public databases of tandem mass spectrometric data has transformed this field in recent years, witnessing an increase of identification rates by 150%. In this presentation, I will outline the key machine learning methods behind this development: kernelbased learning of molecular fingerprints, multiple kernel learning, structured prediction as well as some recent advances. Speaker: Juho Rousu Affiliation: Associate Professor, Aalto University Place of Seminar: Aalto University 

Jan. 30 
Likelihoodfree Inference and Predictions for Computational Epidemiology Abstract: Simulatorbased models often allow inference and predictions under more realistic assumptions than those employed in standard statistical models. For example, the observation model for an underlying stochastic process can be more freely chosen to reflect the characteristics of the data gathering procedure. A major obstacle for such models is the intractability of the likelihood, which has to a large extent hampered their practical applicability. I will discuss recent advances in likelihoodfree inference that greatly accelerate the model fitting process by exploiting a combination of machine learning techniques. Applications to several novel models in infectious disease epidemiology are used to illustrate the potential offered by this approach. Speaker: Jukka Corander Affiliation: Professor, University of Helsinki and University of Oslo Place of Seminar: University of Helsinki Slides 

Feb. 6 
Towards Perfect Density Estimation Abstract: We start by addressing a most simple problem, estimation of a one dimensional density function, and argue that despite of the apparent simplicity of the problem, it is surprisingly difficult to solve it in a holistic manner that is both computationally feasible and theoretically justifiable without strong distributional or other assumptions. We demonstrate how the informationtheoretic MDL framework can be used for reaching this goal (almost) perfectly, and show how this simple setup gives interesting perspectives on the fundamental concepts in probabilistic modelling and statistical inference. We also discuss ideas for extending the framework to more complex models with additional practical applications. Speaker: Petri Myllymäki Affiliation: Professor, University of Helsinki Place of Seminar: Aalto University 

Feb. 13 
Variable Selection From Summary Statistics Abstract: With increasing capabilities to measure a massive number of variables, efficient variable selection methods are needed to improve our understanding of the underlying data generating processes. This is evident, for example, in human genomics, where genomic regions showing association to a disease may contain thousands of highly correlated variants, while we expect that only a small number of them are truly involved in the disease process. I outline recent ideas that have made variable selection practical in human genomics and demonstrate them through our experiences with the FINEMAP algorithm (Benner et al. 2016, Bioinformatics). (1) Compressing data to lightweight summaries to avoid logistics and privacy concerns related to complete data sharing and to minimize the computational overhead. (2) Efficient implementation of sparsity assumptions. (3) Efficient stochastic search algorithms. (4) Use of public reference databases to complement the available summary statistics. Speaker: Matti Pirinen Affiliation: Academy Research Fellow, Institute for Molecular Medicine Finland, University of Helsinki Place of Seminar: University of Helsinki Slides 

Feb. 20 
Compressed Sensing for SemiSupervised Learning From Big Data Over Networks Abstract: In this talk I will present some of our most recent work on the application of compressed sensing to semisupervised learning from massive networkstructured datasets, i.e., big data over networks. We expect the user of compressed sensing ideas to be gamechanging for machine learning from big data in a similar manner as it was for digital signal processing. In particular, I will present a sparse label propagation algorithm which efficiently learn from large amounts of networkstructured unlabeled data by leveraging the information provided by a few initially labelled training data points. This algorithm is inspired by compressed sensing recovery methods and allows for a simple sufficient condition on the network structure which guarantees accurate learning. Speaker: Alexander Jung Affiliation: Assistant Professor, Aalto University Place of Seminar: Aalto University 

Feb. 27 
Inverse Modeling in Behavioral Sciences and HCI Abstract: Can one make deep inferences about a person based only on observations of how she acts? I discuss methodology for inverse modeling in behavioral sciences, where the goal is to estimate a cognitive model from limited behavioral data. Given substantial diversity in people's intentions, strategies and abilities, this is a difficult problem and previously unaddressed. I discuss advances achieved with an approach that combines (1) computational rationality, to predict how a person adapts to a task when her capabilities are known, and (2) Approximate Bayesian Computation (ABC) to estimate those capabilities. The benefit is that model parameters are conditioned on both prior knowledge and observations, which improves model validity and helps identify causes for observations. Inverse modeling methods can advance theoryformation by bringing complex behavior within reach of modeling. This talk is based on ongoing collaborations with Antti Kangasraasio, Samuel Kaski, Jukka Corander, Andrew Howes, Kumaripaba Athukorala, Jussi Jokinen, Sayan Sarcar, and Xiangshi Ren. Speaker: Antti Oulasvirta Affiliation: Associate Professor, Aalto University Place of Seminar: University of Helsinki 

March. 6 
Differentially Private Bayesian Learning Abstract: Many applications of machine learning for example in health care would benefit from methods that can guarantee data subject privacy. Differential privacy has recently emerged as a leading framework for private data analysis. Differenctial privacy guarantees privacy by requiring that the results of an algorithm should not change much even if one data point is changed, thus providing plausible deniability for the data subjects. In this talk I will present methods for efficient differentially private Bayesian learning. In addition to asymptotic efficiency, we will focus on how to make the methods efficient for moderatelysized data sets. The methods are based on perturbation of sufficient statistics for exponential family models and perturbation of gradients for variational inference. Unlike previous stateoftheart, our methods can predict drug sensitivity of cancer cell lines using differentially private linear regression with better accuracy than using a very small nonprivate data set. Speaker: Antti Honkela Affiliation: Assistant Professor, University of Helsinki Place of Seminar: Aalto University 

March. 13 
Small Data AUC Estimation of Machine Learning Methods: Pitfalls and Remedies Abstract: Asking whether two populations can be distinguished from each other is one of the most fundamental questions in data analysis and area under ROC curve (AUC) is one of the simplest and most practical tools for answering it. Also known as the WilcoxonMannWhitney U statistic, it can be associated with a pvalue indicating how likely one would obtain as good AUC value if the two populations would not be stochastically different. Estimating AUC of a predictive model and its statistical significance has a huge practical importance in fields like medicine, where one often has access to only small amounts of labeled data but large number of features. Leavepairout crossvalidation (LPOCV) is an almost unbiased AUC estimator of machine learning methods that has also been empirically shown to be the most reliable of the crossvalidation (CV) based estimators. We further study the properties of LPOCV and show some serious pitfalls one can encounter when estimating AUC with CV and how to avoid them. In particular, we show how one can produce very promising results with high AUC values even if there is no signal in the data. Finally, we show how to counter these risks with new Wilcoxon–Mann–Whitney U type of permutation tests adjusted for LPOCV, thus upgrading one of the classical statistical tools for CV estimates. Speaker: Tapio Pahikkala Affiliation: Assistant Professor, University of Turku Place of Seminar: University of Helsinki 

March. 20 
Future AI: Autonomous machine learning and beyond Abstract: Many researchers have identified autonomous machine learning (unsupervised, semisupervised and reinforcement learning) as an important cornerstone of advanced artificial intelligence. The Curious AI Company is developing such autonomous learning systems. We already have stateoftheart results in several semisupervised classification tasks but we are also working on bringing autonomy to learning segmentation and hierarchical control, both of them tasks that take a lot of human work when developing for instance selfdriving cars. However, we believe there's an even more important blocker on the way to advanced AI: the fundamental inability of currently used parallel distributed neural coding to properly represent objects and their interactions. We are working on deep learning networks whose neurosymbolic representations will hopefully allow neural networks to understand the world not only in terms of a collection of features but in terms of objects and their interactions, too. This is necessary for many tasks such as communication, reasoning and complex decision making. Speaker: Harri Valpola Affiliation: CEO of the Curious AI Company Place of Seminar: Aalto University 

March. 27 
Learning to Rank: Applications to Bioinformatics Abstract: Learning To Rank (LTR) has been developed in information retrieval for ranking documents regarding the relevance to a given query. Typically LTR builds a ranking model from given relevant (or irrelevant) querydocument pairs. Generally, in some respect, LTR can be thought as an attempt to solve a multilabel classification problem, where queries are labels. A lot of settings in bioinformatics can be turned into multilabel classification problems having relatively similar properties. One typical example is biomedical document annotation. Currently PubMed, a database of 26 million biomedical citations, has around 30,000 keywords, called MeSH (Medical Subject Headings) terms, i.e. labels in multilabel classification, where the number of articles per MeSH term is extremely diverse, ranging from only 20 to more than eight million. This large, biased dataset already goes beyond the general sense of settings expected by regular multilabel classifiers. In this talk, I will start with introduction and a brief review of LTR. I then raise three bioinformatics multilabel classification problems that share real dataderived, practical properties, which hamper the application of regular multilabel classifiers. Finally I will show that LTR nicely addresses such largescale, challenging bioinformatics multilabel classification problems. A large portion of this talk appeared in ISMB in 2015 and 2016. Speaker: Hiroshi Mamitsuka Affiliation: Professor, Kyoto University Place of Seminar: University of Helsinki 

April. 03 
Multilayer Networks Abstract: Network science has been very successful in investigations of a wide variety of applications from biology and the social sciences to physics, technology, and more. In many situations, it is already insightful to use a simple (and typically naive) representation as a simple, binary graph in which nodes are entities and unweighted edges encapsulate the interactions between those entities. This allows one to use the powerful methods and concepts for example from graph theory, and numerous advances have been made in this way. However, as network science has matured and (especially) as ever more complicated data has become available, it has become increasingly important to develop tools to analyse more complicated structures. For example, many systems that were typically initially studied as simple graphs are now often represented as timedependent networks, networks with multiple types of connections, or interdependent networks. This has allowed deeper and more realistic analyses of complex networked systems, but it has simultaneously introduced mathematical constructions, jargon, and methodology that are specific to research in each type of system. Recently, the concept of “multilayer networks” was developed in order to unify the aforementioned disparate language (and disparate notation) and to bring together the different generalised network concepts that included layered graphical structures. In this talk, I will introduce multilayer networks and discuss how to study their structure. Generalisations of the clustering coefficient for multiplex networks and graph isomorphism for general multilayer networks are used as illustrative examples. Speaker: Mikko Kivelä Affiliation: Postdoctoral Researcher, Aalto University Place of Seminar: Aalto University 

April. 10 
Learning With Spectral Kernels Abstract: Machine learning algorithms learn models that automatically infer data representations and generalise into new data. Gaussian processes are Bayesian kernelbased models with a key advantage of being able to efficiently learn kernel functions from data. All kernel functions can be decomposed into sinusoidal components, which provide a highly expressive basis for learning arbitrary representations. In this talk I will discuss how we can exploit spectral kernel learning for largescale multitask learning. We also generalise spectral learning into learning nonstationary kernels with inputspecific behavior Speaker: Markus Heinonen Affiliation: Department of Computer Science, Aalto University Place of Seminar: University of Helsinki 

April. 24 
Nintendo Wii FitBased Balance Testing to Detect Sleep Deprivation: Approximate Bayesian Computation Approach Abstract: Sleep deprivation deteriorates health and causes accidents. Measuring a person’s postural steadiness may be used to determine his/hers state of alertness. Posturographic measurements are easy to conduct: a person’s body sway is measured during upright stance on a balance board for 60 s. The Nintendo Wii Fit balance board is a portable and affordable alternative to expensive clinical force plates. Body sway may be modeled with a singlelink inverted pendulum (Asai et al. 2009). The model parameters, such as time delay and noise intensity in the nervous system, are physiologically relevant. The pendulum is kept upright with controllers, that include stiffness and damping gain parameters. Level of control determines how often the active controller is ON. The model cannot be solved analytically in closed form. Therefore, inferring model parameters and their confidence limits is nontrivial. We used sequential Monte Carlo approximate Bayesian computation (SMCABC) algorithm to infer the model parameters. The inferred parameters may allow determining a person’s state of alertness. Speaker: Aino Tietäväinen Affiliation: Department of Physics, University of Helsinki Place of Seminar: Aalto University 

May. 08 
Empirical Parameterization of Exploratory Search Systems Based on Bandit Algorithms Abstract: Exploratory searches are where a user has insufficient knowledge to define exact search criteria or does not otherwise know what they are looking for. Reinforcement learning techniques have demonstrated great potential for supporting exploratory search in information retrieval systems as they allow the system to tradeoff exploration (presenting the user with alternatives topics) and exploitation (moving toward more specific topics). Users of such systems, however, often feel that the system is not responsive to user needs. This problem is not an inherent feature of such systems, but is caused by the exploration rate parameter being inappropriately tuned for a given system, dataset or user. In this talk, we discuss two approaches how to optimise exploratory search systems based on bandit algorithms. First, we show that the tradeoff between exploration and exploitation can be modelled as a direct relationship between the exploration rate parameter from the reinforcement learning algorithm and the number of relevant documents returned to the user over the course of a search session. We define the optimal exploration/exploitation tradeoff as where this relationship is maximised and show this point to be broadly concordant with user satisfaction and performance. Our second approach aims to dynamically adapt exploration and exploitation in a manner commensurate with the user's individual requirements for each search session. We present a novel study design together with a regression model for predicting the optimal exploration rate based on simple metrics from the first iteration, such as clicks and reading time. We perform model selection based on the data collected from a user study and show that predictions are consistent with user feedback. Speaker: Dorota Glowacka Affiliation: Department of Computer Science, University of Helsinki Place of Seminar: University of Helsinki 

May. 15 
Machine Learning for ImageBased Localization Abstract: Imagebased localization refers to a problem where the camera position and orientation for a given query image is computed with respect to a known visual 3D map of the scene. This problem is relevant for applications such as robot selflocalization, pedestrian navigation, and augmented reality. Another related problem is the relative pose estimation between two camera views which is required for computing imagebased 3D models from a collection of 2D images. Traditionally both of these problems have been approached by using handcrafted local image features and descriptors, such as the widely used SIFT keypoint detector. However, recently several deep learning based localization approaches have been proposed. They omit local feature matching and directly try to regress the camera pose. In this presentation, we will describe an overview of the problem area and explore some recent deep learning based approaches. We will also present some of our own recent results in this area. Speaker: Juho Kannala Affiliation: Professor of Computer Science, Aalto University Place of Seminar: Aalto University 

May. 22 
On Priors and Bayesian Variable Selection in Large p, Small n Regression Abstract: The Bayesian approach is well known for using priors to improve inference, but equally important part is the integration over the uncertainties. I first present recent development in hierarchical shrinkage priors for presenting sparsity assumptions in covariate effects. I then present a projection predictive variable selection approach, which is a Bayesian decision theoretical approach for variable selection which can preserve the essential information and uncertainties related to all variables in the study. I also present recent excellent experimental results and easy to use software. Speaker: Aki Vehtari Affiliation: Professor of Computer Science, Aalto University Place of Seminar: University of Helsinki 

May. 29 
Graphics Meets Vision Meets Machine Learning Abstract: Realistic threedimensional modeling and animation are key bottlenecks in the production film, games, VR, and other applications of computer graphics. In this talk, I will describe our recent research that makes use of machine learning techniques for solving hard inference problems for generating 3D content: capture and reproduction of photorealistic surface appearance, facial performance capture, and turning audio into facial animation. These works both push the state of the art forward in research – two of the three projects have been published at ACM SIGGRAPH – and are surprisingly ready for production use already now. Speaker: Jaakko Lehtinen Affiliation: Professor of Computer Science, Aalto University Place of Seminar: Aalto University 

Jun. 5 
Statistical Ecology with Gaussian Processes Abstract: Ecology studies the distribution and abundance of species, and their interactions with other species and the environment. Key questions in ecology include what are the environmental factors and interspecies dependencies that drive species distributions, how these processes together affect species community structures and how environmental changes, such as climate change, affect species distribution and species communities. These questions are essentially about variable selection and causal and predictive inference. Hence, statistics has a central role in answering them. The species distribution models (SDMs) used for these analyses are traditionally based on generalized linear and additive models. In this talk I will present how Gaussian processes (GPs) can be used in SDMs and what benefits and challenges this provides. I will present recent results on GP based species distribution modeling in the Baltic Sea and Great Barrier Reef, Australia. I will discuss the potential future development and current challenges related to computation and model building. Speaker: Jarno Vanhatalo Affiliation: Professor of Statistics, University of Helsinki Place of Seminar: University of Helsinki 

Jun. 12 
Learning Data Representation by LargeScale Neighbor Embedding Abstract: Machine learning, the stateoftheart data science, has been increasingly influencing our life. Encoding data in a suitable vector space is the fundamental starting point for machine learning. A good vector coding should respect the relations among the data items. However, conventional methods that preserve pairwise or higher order relationship are very slow and consequently they can handle only smallscale data sets. We have been developing a family of unsupervised methods called largescale Neighbor Embedding (NE) which substantially accelerate the vector coding. Our method can thus learn lowdimensional vector representation for megascale data according to their neighborhoods in the original space. With our efficient algorithms and a wealth of neighborhood information, Neighbor Embedding significantly outperforms smallscale NE and many other existing approaches for learning data representation. Besides generic feature extraction, our work also delivers two important tools as special cases of Neighbor Embedding for data visualization and cluster analysis, which scales up these applications by an order of magnitude and enables the currentsized visualization and clustering for interactive use. Because neighborhood information is naturally and massively available in many areas, our method has wide applications as a critical component in scientific research, nextgeneration DNA sequence analysis, natural language processing, educational cloud, financial data analysis, market studies, etc. Speaker: Zhirong Yang Affiliation: Department of Computer Science, Aalto University Place of Seminar: Aalto University 

Sep. 11 
Latent Stochastic Models for Comparing Tumor Samples of Unknown Purity Abstract: A challenge in analyzing data from tumor samples is that the biopsies contain an mixture of various cells, including cancer cells, immune cells and stromal cells. This hinders the discovery of clinically relevant information and can lead to systematically biased results. A few recent analysis techniques control for such factors, but only accommodate specific types of data, or require controls which cannot be obtained from each patient. I will present our developments on statistical methods for controlling for the latent and varying fraction of tumor cells in nextgeneration methylation and RNA sequencing data, which aim to enable unbiased and more accurate comparison of patientderived samples. Speaker: Antti Häkkinen Affiliation: Postdoctoral fellow, GenomeScale Biology Program, Faculty of Medicine, University of Helsinki Place of Seminar: Aalto University 

Sep. 18 
Fast Nearest Neighbor Search in High Dimensions by Multiple Random Projection Trees Abstract: Efficient index structures for fast approximate nearest neighbor queries are required in many applications such as recommendation systems. In highdimensional spaces, many conventional methods suffer from excessive usage of memory and slow response times. We propose a method where multiple random projection trees are combined. We demonstrate by extensive experiments on a wide variety of data sets that the method is faster than existing partitioning tree or hashing based approaches, making it the fastest available technique on high accuracy levels. Speaker: Teemu Roos Affiliation: Associate Professor, Department of Computer Science, University of Helsinki Place of Seminar: University of Helsinki Hyvönen et al., "Fast Nearest Neighbor Search through Sparse Random Projections and Voting", IEEE Big Data Conference 2016: [link] 

Sep. 25 
Computational creativity and machine learning. Abstract: Computational creativity has been defined as the art, science, philosophy and engineering of computational systems which, by taking on particular responsibilities, exhibit creative behaviours. In this talk I first try to elaborate on what creative responsibilities could be and why they are interesting. I then outline ways in which machine learning can be used to take on some of these responsibilities, helping computational systems become more creative. Speaker: Hannu Toivonen Affiliation: Professor of Computer Science, University of Helsinki Place of Seminar: Aalto University 

Oct. 2 
Machine learning for Materials Research. Abstract: In materials research, we have learnt to predict the evolution of microstructure starting with the atomic level processes. We know about defects  point and extended,  and we know that these can be crucial for the final structural (and related mechanical and electrical) properties. Often simple macroscopic differential equations, which are used for the purpose, fail to predict simple changes in materials. Many questions remain unanswered. Why a ductile material suddenly becomes brittle? Why a strong concrete bridge suddenly cracks and eventually collapses after serving for tens of years? Why the wall of high quality steels in fission reactors suddenly crack? Or, why the clean smooth surface roughens under applied electric fields? All these questions can be answered, if one peeks in to atom's behavior imagining it jumping inside the material. But how the atoms "choose" where to jump amongst the numerous possibilities in complex metals? Tedious parameterization can help to deal with the problem, but machine learning can provide a better and more elegant solution to this problem. Speaker: Flyura Djurabekova Affiliation: Department of Physics, University of Helsinki Place of Seminar: University of Helsinki 

Oct. 9 
Does my algorithm work? Abstract: It is easy to propose a new algorithm for solving a Machine Learning problem. It is much harder to convince other people that the proposed algorithm actually works. The "gold standard" of tight theoretical guarantees is often out of reach. So what do we do? Typically, an algorithm is validated on a couple of test problems and its output is compared with that of algorithms that are known to work. This is not a great strategy. Speaker: Daniel Simpson Affiliation: Professor of Stastical Sciences, University of Toronto Place of Seminar: Aalto University 

Oct. 16 
Probabilistic preference learning with the Mallows rank model Abstract: Ranking and comparing items is crucial for collecting information about preferences in many areas, from marketing to politics. The Mallows rank model is among the most successful approaches to analyze rank data, but its computational complexity has limited its use to a form based on Kendall distance. Here, new computationally tractable methods for Bayesian inference in Mallows models are developed that work with any rightinvariant distance. The method performs inference on the consensus ranking of the items, also when based on partial rankings, such as topk items or pairwise comparisons. When assessors are many or heterogeneous, a mixture model is proposed for clustering them in homogeneous subgroups, with clusterspecific consensus rankings. Approximate stochastic algorithms are introduced that allow a fully probabilistic analysis, leading to coherent quantification of uncertainties. The method can be used, for example, for making probabilistic predictions on the class membership of assessors based on their ranking of just some items, and for predicting missing individual preferences, as needed in recommendation systems. Speaker: Elja Arjas Affiliation: Professor (emeritus) of Mathematics and Statistics, University of Helsinki Place of Seminar: University of Helsinki 

Oct. 23 
Computational Challenges in Analyzing And Moderating Online Social Discussions Abstract: Online social media are a major venue of public discourse today, hosting the opinions of hundreds of millions of individuals. Social media are often credited for providing a technological means to break information barriers and promote diversity and democracy. In practice, however, the opposite effect is often observed: users tend to favor content that agrees with their existing worldview, get less exposure to conflicting viewpoints, and eventually create "echo chambers" and increased polarization. Arguably, without any kind of moderation, current socialmedia platforms gravitate towards a state in which netcitizens are constantly reinforcing their existing opinions. In this talk we present a ongoing line of work on analyzing and moderating online social discussions. We first consider the questions of detecting controversy using network structure and content, tracking the evolution of polarized discussions, and understanding their properties over time. We then address the problem of designing algorithms to break filter bubbles and reduce polarization. We discuss a number of different strategies such as user and content recommendation, as well as viral approaches. Speaker: Aristides Gionis Affiliation: Professor of Computer Science, Aalto University Place of Seminar: Aalto University 

Oct. 30 
Learning Markov Equivalence Classes of Directed Acyclic Graphs: an Objective Bayes Approach Abstract: A Markov equivalence class contains all the Directed Acyclic Graphs (DAGs) encoding the same conditional independencies, and is represented by a Completed Partially Directed DAG (CPDAG), also named Essential Graph (EG). We approach the problem of model selection among noncausal sparse Gaussian DAGs by directly scoring EGs, using an objective Bayes method. Specifically, we construct objective priors for model selection based on the Fractional Bayes Factor, leading to a closed form expression for the marginal likelihood of an EG. Next we propose an MCMC strategy to explore the space of EGs, possibly accounting for sparsity constraints, and illustrate the performance of our method on simulation studies, as well as on a real dataset. Our method is fully Bayesian and thus provides a coherent quantification of inferential uncertainty, requires minimal prior specification, and shows to be competitive in learning the structure of the datagenerating EG when compared to alternative stateoftheart algorithms. Speaker: Guido Consonni Affiliation: Professor of Statistics, Universita Cattolica del Sacro Cuore Place of Seminar: University of Helsinki 

Nov. 6 
Efficient and accurate approximate Bayesian computation Abstract: Approximate Bayesian computation (ABC) is a method for calculating a posterior distribution when the likelihood is intractable, but simulating the model is feasible. It has numerous important applications, for example in computational biology, material physics, user interface design, etc. However, many ABC algorithms require a large number of simulations, which can be costly. To reduce the cost, Bayesian optimisation (BO) and surrogate models such as Gaussian processes have been proposed. Bayesian optimisation enables deciding intelligently where to simulate the model next, but standard BO approaches are designed for optimisation and not for ABC. Here we address this gap in the existing methods. We model the uncertainty in the ABC posterior density which is due to a limited number of simulations available, and define a loss function that measures this uncertainty. We then propose to select the next model simulation to minimise the expected loss. Experiments show the proposed method is often more accurate than the existing alternatives. Speaker: Pekka Marttinen Affiliation: Academy Research Fellow, Department of Computer Science, Aalto University Place of Seminar: Aalto University 

Nov. 13 
Learning of Ultra HighDimensional Potts Models for Bacterial Population Genomics Abstract: The potential for genomewide modeling of epistasis has recently surfaced given the possibility of sequencing densely sampled populations and the emerging families of statistical interaction models. Direct coupling analysis (DCA) has earlier been shown to yield valuable predictions for single protein structures, and has recently been extended to genomewide analysis of bacteria, identifying novel interactions in the coevolution between resistance, virulence and core genome elements. However, earlier computational DCA methods have not been scalable to enable model fitting simultaneously to 10000100000 polymorphisms, representing the amount of core genomic variation observed in analyses of many bacterial species. Here we introduce a novel inference method (SuperDCA) which employs a new scoring principle, efficient parallelization, optimization and filtering on phylogenetic information to achieve scalability for up to 100000 polymorphisms. Using two large population samples of Streptococcus pneumoniae, we demonstrate the ability of SuperDCA to make additional significant biological findings about this major human pathogen. We also show that our method can uncover signals of selection that are not detectable by genomewide association analysis, even though our analysis does not require phenotypic measurements. SuperDCA thus holds considerable potential in building understanding about numerous organisms at a systems biological level. Speaker: Jukka Corander Affiliation: Professor of Statistics, University of Helsinki and University of Oslo Place of Seminar: University of Helsinki 

Nov. 20 
Towards Intelligent Exergames Abstract: Exergames  video games that require physical activity  hold promise of solving the societal hard problem of motivating people to move. At the same time, artificial intelligence and machine learning are transforming how video games are designed, produced, and tested. Work combining both computational intelligence and exergames is sparse, however. In my talk, I delineate the challenges, opportunities, and my group's research towards intelligent exergames, building on our previous research on both exergame design (e.g., Augmented Climbing Wall, Kick Ass KungFu) and intelligent control of embodied simulated agents. Video and examples: http://perttu.info Speaker: Perttu Hämäläinen Affiliation: Professor of Computer Science, Aalto University Place of Seminar: Aalto University 

Nov. 27 
CorrelationCompressed Direct Coupling Analysis Abstract: Direct Coupling Analysis (DCA) is a powerful tool to find pairwise dependencies in large biological data sets. It amounts to inferring coefficients in a probabilistic model in an exponential family, and then using the largest such inferred coefficients as predictors for the dependencies of interest. The main computational bottleneck is the inference. As described recently by Jukka Corander in this seminar series DCA has be done on bacterial wholegenome data, at the price of significant compute time, and investment in code optimization. We have looked at if DCA can be speeded up by first filtering the data on correlations, an approach we call CorrelationCompressed Direct Coupling Analysis (CCDCA). The computational bottleneck then moves from DCA to the more standard task of finding a subset of most strongly correlated vectors in large data sets. I will describe results obtained so far, and outline what it would take to do CCDCA on wholegenome data in human and other higher organisms. This is joint work with ChenYi Gao and HaiJun Zhou, available as arXiv:1710.04819. Speaker: Erik Aurell Affiliation: Professor of Biological Physics, KTHRoyal Institute of Technology Place of Seminar: University of Helsinki 

Dec. 18 
Ecofriendly Bayes: Reusing Computationally Costly Posteriors in Posthoc Analyses Abstract: Bayesian inference has many attractive features, but a major challenge is its potentially very high computational cost. While sampling from the prior distribution is often straightforward, the most expensive part is typically conditioning on the data. In many problems, a single data set data may not be informative enough to enable reliable inference for a given quantity of interest. This can be difficult to assess in advance and may require a considerable amount of computation to discover, resulting in a weakly informative posterior distribution "gone to waste". On the other hand, borrowing strength across multiple related data sets using a hierarchical model may for very costly models be computationally infeasible. As an alternative approach to traditional hierarchical models, we develop in this work a framework which reuses and combines posterior distributions computed on individual data sets to achieve posthoc borrowing of strength, without the need to redo expensive computations on the data. As a byproduct, we also obtain a notion of metaanalysis for posterior distributions. By adopting the view that posterior distributions are beliefs which reflect the uncertainty about the value of some quantity, we formulate our approach as Bayesian inference with uncertain observations. We further show that this formulation is closely related to belief propagation. Finally, we illustrate the framework with posthoc analyses of likelihoodfree Bayesian inferences. Speaker: Paul Blomstedt Affiliation: Department of Computer Science, Aalto University Place of Seminar: Aalto University 

Jan. 15 
Scalable Algorithms for Extreme MultiClass and MultiLabel Classificiation in Big Data Abstract: In the era of big data, largescale classification involving tens of thousand target categories is not uncommon. Also referred to as Extreme Classification, it has also been recently shown that the machine learning challenges arising in ranking, recommendation systems and webadvertising can be effectively addressed by reducing it to extreme multilabel classification framework. In this talk, I will discuss my two recent works, and present TerseSVM and DiSMEC algorithms for extreme multiclass and multilabel classification. The precision@k and nDCG@k results using DiSMEC improve by upto 20% on benchmark datasets over stateoftheart methods, which are used by Microsoft in production system of Bing Search. The training process for these algorithms makes use of openMP based distributed architectures, and is able to leverage thousands of cores for computation. Speaker: Rohit Babbar Affiliation: Professor of Computer Science, Aalto University Place of Seminar: Aalto University 

Jan. 22 
Confident Bayesian Learning of Graphical Models Abstract: Confident Bayesian learning amounts to computing summaries of a Speaker: Mikko Koivisto Affiliation: Professor of Computer Science, University of Helsinki Place of Seminar: University of Helsinki 

Jan. 29 
Special Seminar This week we will have multiple machine learning talks at Aalto. We cancel the normal Monday session so that we can meet in them according to the following plan: Jan 29, Glowacka Dorota
Jan 30, Heinonen Markus
Feb 1, Marttinen Pekka
Feb 2, Solin Arno
NOTE: The TUAS building address is Maarintie 8 (next to the CS building) 

Feb. 5 
Studying mutational processes in cancer Abstract: Somatic mutations in cancer have accumulated during its evolution and are caused by different exposures to carcinogens and therapeutic agents, as well as, intrinsic errors that occur during DNA replication. Analysing a set of cancer samples jointly allows to explain their somatic mutations as a linear combination of (to be learned) mutational signatures. In this presentation I will discuss the problem of learning mutational signatures from cancer data using probabilistic modelling and nonnegative matrix factorisation. I further describe our ongoing work using mutational signatures in the context of drug response prediction and extensions of the basic model to explicitly include DNA repair processes. Speaker: Ville Mustonen Affiliation: Professor of Mathematics and Natural Science, University of Helsinki Place of Seminar: University of Helsinki 

Feb. 12 
Here Be Dragons: HighDimensional Spaces and Statistical Computation Abstract: With consistently growing data sets and increasingly complex models, the frontiers of applied statistics is found in highdimensional spaces. Unfortunately most of the intuitions that we take for granted in our lowdimensional, routine experiences don’t persist to these highdimensional spaces which makes the development of scalable computational methodologies and algorithms all the more challenging. In this talk I will discuss the counterintuitive behavior of highdimensional spaces and the consequences for statistical computation. Speaker: Michael Betancourt His research focuses on the development of robust statistical workflows, computational tools, and pedagogical resources that bridge statistical theory and practice and enable scientists to make the most out of their data. The pursuit of general but scalable statistical computation has lead him to the intersection of differential geometry and probability theory where exploiting the inherent geometry of highdimensional problems naturally leads to algorithms such Hamiltonian Monte Carlo and its generalizations. He is developing both the theoretical foundations and the practical implementations of these algorithms, the latter specifically in the software ecosystem Stan. Affiliation: Columbia University Place of Seminar: Aalto University 

Feb. 19 
Learning and Stochastic Control in Gaussian Process Driven Physical Systems Abstract: Traditional machine learning is often overemphasising problems, where we wish to automatically learn everything from the problem at hand solely using a set of training data. However, in many physical systems we already know much about the physics, typically in form of partial differential equations. For efficient learning in this kind of systems, it is beneficial to use graybox models where only the unknown parts are modeled with datatrained machine learning models. This talk is concerned with learning and stochastic control in physical systems which contain unknown input or force signals that we wish to learn from data. These unknown signals are modeled using Gaussian processes (GP) from machine learning. The resulting latent force models (LFMs) can be seen as hybrid models that contain a firstprinciples physical model part and a nonparametric GP model part. We present and discuss methods for learning and stochastic control in this kind of models. Speaker: Simo Särkkä Simo Särkkä is an Associate Professor and Academy Research Fellow with Aalto University, Technical Advisor of IndoorAtlas Ltd., and an Adjunct Professor with Tampere University of Technology and Lappeenranta University of Technology. His research interests are in multisensor data processing systems with applications in artificial intelligence, machine learning, inverse problems, location sensing, health technology, and brain imaging. He has authored or coauthored around 100 peerreviewed scientific articles and his book "Bayesian Filtering and Smoothing" along with its Chinese translation were published via the Cambridge University Press in 2013 and 2015, respectively. He is a Senior Member of IEEE and serving as an Associate Editor of IEEE Signal Processing Letters. Affiliation: Professor of Electrical Engineering and Automation, Aalto University Place of Seminar: University of Helsinki 

Feb. 26 
Bayesian Deep Learning for Image Data Abstract: Deep learning is the paradigm that lies at the heart of stateoftheart machine learning approaches. Despite their groundbreaking success on a wide range of applications, deep neural nets suffer from: i) being severely prone to overfitting, ii) requiring intensive handcrafting in topology design, iii) being agnostic to model uncertainty, iv) and demanding large volumes of labeled data. The Bayesian approach provides principled solutions to all of these problems. Bayesian deep learning converts the loss minimization problem of conventional neural nets into a posterior inference problem by assigning prior distributions on synaptic weights. This talk will provide a recap of recent advances in Bayesian neural net inference and detail my contributions to the solution of this problem. I will demonstrate how Bayesian neural nets can achieve groundbreaking performance in weaklysupervised learning, active learning, fewshot learning, and transfer learning setups when applied to medical image analysis and core computer vision tasks. I will conclude by a summary of my ongoing research in reinforcement active learning, videobased imitation learning, and reconciliation of Bayesian Program Learning with Generative Adversarial Nets. Speaker: Melih Kandemir Dr. Kandemir studied computer science in Hacettepe University and Bilkent University between 2001 and 2008. Later on, he pursued his doctoral studies in Aalto University (former Helsinki University of Technology) on the development of machine learning models for mental state inference until 2013. He worked as a postdoctoral researcher in Heidelberg University, Heidelberg Collaboratory for Image Processing (HCI) between 2013 and 2016. As of 2017, he is an assistant professor at Özyeğin University, Computer Science Department. Throughout his career, he took part in various research projects in funded collaboration with multinational corporations including Nokia, Robert Bosch GmbH, and Carl Zeiss AG. Bayesian deep learning, fewshot learning, active learning, reinforcement learning, and application of these approaches to computer vision are among his research interests. Affiliation: Professor of Computer Science, Özyeğin University Place of Seminar: Aalto University 

March 5 
Artificial Intelligence for Mobility Studies in Urban And Natural Areas Abstract: Understanding how people use and move in space is important for planning, both in urban and natural areas. Recent research has shown that locationbased social media data may reveal spatial and temporal patterns of the use of space, and reveal areas where human activities might be detrimental. We have shown that social media data corresponds to reallife spatial and temporal patterns of visitors in national parks and is able to bring light to use of space in cities, by providing meaningful information about the activities and preferences of people. The overwhelming magnitudes of social media data require special filtering and cleaning and tested analyses approaches. We are now using geospatial analysis methods together with machine learning to understand where, when, how and by whom areas are being used and how people and goods move about and why. Automated text and image content analysis is needed to leverage the full potential of social media data in spatial planning. Also new applications are yet to be discovered. Speaker: Tuuli Toivonen Affiliation: Professor of Geoinformatics, University of Helsinki Place of Seminar: University of Helsinki 

March 12 
Finding Outlier Correlations Abstract: Finding strongly correlated pairs of observables is one of the basic tasks in data analysis and machine learning. Assuming we have N observables, there are N(N1)/2 pairs of distinct observables, which gives rise to quadratic scalability in N if our approach is to explicitly compute all pairwise correlations. In this talk, we look at algorithm designs that achieve subquadratic scalability in N to find pairs of observables that are strongly correlated compared with the majority of the pairs. Our plan is to start with an exposition of G. Valiant's breakthrough design [FOCS'12,JACM'15] and then look at subsequent improved designs, including some of our own work. Based on joint work with M. Karppa, J. Kohonen, and P. Ó Catháin, cf. https://arxiv.org/abs/1510.03895 (ACM TALG, to appear) and https://arxiv.org/abs/1606.05608 (ESA'16). Speaker: Petteri Kaski Affiliation: Professor of Computer Science, Aalto University Place of Seminar: Aalto University 

March 19 
Fun With Relative Distance Comparisons Abstract: The distance between two data points is a fundamental ingredient of many data analysis methods. In this talk I review some of my work on “human computation” algorithms that use relative distance comparisons only. I.e., statements of the form “of items a, b, and c, item c is an outlier”, or “item a is closer to item b than to item c”. Such statements are easier to elicit from human annotators than absolute judgements of distance. I consider the problems of centroid computation (Heikinheimo & Ukkonen, HCOMP 2013), density estimation (Ukkonen et al, HCOMP 2015), embeddings (Amid & Ukkonen, ICML 2015), clustering (Ukkonen, ICDM 2017), as well as give a few sneak previews of ongoing work. Speaker: Antti Ukkonen Antti Ukkonen is an Academy research fellow at University of Helsinki. He obtained his doctoral degree at Aalto university in 2008, and has since held positions at Yahoo! Research, Helsinki Institute for Information Technology HIIT, and Finnish Institute of Occupational Health. Currently he is the PI in the “Data Science for the Masses” project funded by Academy of Finland. His research interests include algorithmic aspects of (distributed) human computation and machine learning, as well as applied data science. Affiliation: Professor of Computer Science, University of Helsinki Place of Seminar: University of Helsinki 

March 26 


April 9 


April 16 
Machine Learning using Unreliable Components: From Matrix Operations to Neural Networks and Stochastic Gradient Descent Abstract: Reliable computation at scale is one key challenge in largescale machine learning today.Unreliability in computation can manifest itself in many forms, e.g. (i) "straggling" of a few slow processing nodes which can delay your entire computation, e.g., in synchronous gradient descent; (ii) processor failures; (iii) "softerrors," which are undetected errors where nodes can produce garbage outputs. My focus is on the problem of training using unreliable nodes. First, I will introduce the problem of training model parallel neural networks in the presence of softerrors. This problem was in fact the motivation of von Neumann's 1956 study, which started the field of computing using unreliable components. We propose "CodeNet", a unified, errorcorrection codingbased strategy that is weaved into the linear algebraic operations of neural network training to provide resilience to errors in every operation during training. I will also survey some of the notable results in the emerging area of "coded computing," including my own work on matrixvector and matrixmatrix products, that outperform classical results in faulttolerant computing by arbitrarily large factors in expected time. Next, I will discuss the errorruntime tradeoffs of various data parallel approaches in training machine learning models in presence of stragglers, in particular, synchronous and asynchronous variants of SGD. Finally, I will discuss some open problems in this exciting and interdisciplinary area. Parts of this work is accepted at AISTATS 2018 and ISIT 2018. Speaker: Sanghamitra Dutta Affiliation: PhD Candidate, Carnegie Mellon University Place of Seminar: Aalto University 

April 23 
Deep Learning Spectroscopy: Neural Networks for Molecular Excitation Spectra Abstract: For the study of molecules and materials, conventional theoretical and experimental spectroscopies are well established in the natural sciences, but they are slow and expensive. Our objective is to launch a new era of artificial intelligence (AI) enhanced spectroscopy that learns from the plethora of already available experimental and theoretical spectroscopy data. Once trained, the AI can make predictions of spectra instantly and at no further cost. In this new paradigm, AI spectroscopy would complement conventional theoretical and experimental spectroscopy to greatly accelerate the spectroscopic analysis of materials, make predictions for novel and hitherto uncharacterized materials, and discover entirely new materials. In this presentation, I will introduce the two approaches we have used to learn spectroscopic properties: kernel ridge regression (KRR) and deep neural networks (NN). The models are trained and validated on data generated by accurate stateofthe art quantum chemistry computations for diverse subsets of the GBD13 and GBD17 molecular datasets [1,2]. The molecules are represented by a simple, easily attainable numerical description based on nuclear charges and cartesian coordinates [3,4]. The complexity of the molecular descriptor [4] turns out to be crucial for the learning success, as I will demonstrate for KRR. I will then show, how we can learn spectra (i.e. continuous target quantities) with NNs. We design and test three different NN architectures: multilayer perceptron (MLP) [5], convolutional neural network (CNN) and deep tensor neural network (DTNN) [6]. Already the MLP is able to learn spectra, but the learning quality improves significantly for the CNN and reaches its best performance for the DTNN. Both CNN and DTNN capture even small nuances in the spectral shape. * This work was performed in collaboration with A. Stuke, K. Ghosh, L. Himanen, M. Todorovic, and A. Vehtari [1] L. C. Blum et al., J. Am. Chem. Soc. 131, 8732 (2009) Speaker: Patrick Rinke Affiliation: Professor of Physics, Aalto University Place of Seminar: University of Helsinki 

May 7 
Fairnessaware machine intelligence: foundations and challenges Abstract: Algorithmic decision making is pervasive, the prices we pay, the news or movies we see, the jobs or credits we get are advised by algorithms. Not so long ago the public used to think that decision making by computers is inherently objective, but realization that models learned from data are not more objective than the data on which they have been trained is becoming common. Fairnessaware machine learning has emerged as a discipline about ten years ago with the main goal to correct algorithmically for potential biases towards sensitive groups of people. The talk will discuss the main challenges, existing solutions and current trends in this research area. Speaker: Indre Zliobaite Affiliation: University of Helsinki Place of Seminar: University of Helsinki 
Last updated on 9 May 2018 by Homayun Afrabandpey  Page created on 12 Dec 2016 by Homayun Afrabandpey