Positions for Doctoral Students in computer science (deadline February 2, 2021)

Deadline: Wed, 03.02.2021, this call is closed


Helsinki ICT network: Doctoral student positions in computer science (deadline February 2, 2021)

The Helsinki Doctoral Education Network in Information and Communications Technology (HICT) is a joint initiative by Aalto University and the University of Helsinki, the two leading universities within this area in Finland. The network involves at present over 80 professors and over 200 doctoral students, and the participating units graduate altogether more than 40 new doctors each year.

The participating units of HICT have currently funding available for exceptionally qualified doctoral students. We offer the possibility to join world-class research groups, with 40 research projects to choose from. The activities of HICT and the open positions are structured along five research area specific tracks:

  • Algorithms and machine learning
  • Life science informatics
  • Networks, networked systems and services
  • Software and service engineering and systems
  • User centered and creative technologies

If you wish to be considered as a potential new doctoral student in HICT you can apply to one or a number of doctoral student positions. We welcome applicants with diverse backgrounds, and qualified female candidates are explicitly encouraged to apply. For more information, please see the list of positions and detailed information on the application process below.

The online application form closes February 2, 2021 at midnight Finnish time (23:59 EET, Eastern European Time).

The quality of research and education in both HICT universities is world-class, and the education is practically free as there are no tuition fees for doctoral students in the Finnish university system. In terms of the living environment, Helsinki has been ranked as one of the world's top-10 most livable cities (Economist, 2017), and is also in the third place in the world’s startup city comparison (Valuer, 2019). Finland is among the best countries in the world with respect to many quality of life indicators, ranking at the top in education and subjective wellbeing. Furthermore, Finland is again the world's happiest country in 2020 (World Happiness Report, 2020).

 

Spring call 2021 positions and projects (40 open doctoral student positions!)

Project 1. Large-scale computing on modern architectures and systems: Programming models, Designs and Optimization

Project 2. Doctoral students for ML Engineering research

Project 3: Query processing and optimization for next-generation databases

Project 4: Atmospheric AI (The exploratory data analysis group)

Project 5: Theory of safe algorithms and their applications in Bioinformatics

Project 6: Probabilistic real-time machine learning

Project 7: Deep learning with probabilistic principles

Project 8. Deep learning with differential equations

Project 9: Sample-efficient probabilistic machine learning

Project 10. Probabilistic modelling and Bayesian machine learning

Project 11: Improve drug design with human assisted AI

Project 12: Decomposable latent representations for toxicity prediction

Project 13: Bayesian deep learning

Project 14: Mobile Cross Reality through Immersive Computing (MeXICO)

Project 15: HAIC: Open doctoral student position in Prof. Janne Lindqvist’s group – security engineering and usable security

Project 16: Open doctoral student position in Prof. Janne Lindqvist’s group – understanding video streaming user experiences

Project 17: HAIC: Open doctoral student position in Prof. Janne Lindqvist’s group – artificial intelligence and machine learning for systems security and privacy

Project 18: Open doctoral student position in Prof. Janne Lindqvist’s group – multitasking and productivity tools

Project 19: Open doctoral student position in Prof. Janne Lindqvist’s group – mixed methods HCI and security research

Project 20: Deep learning for electronic health records

Project 21: Theoretical Framework and Deep Learning Algorithms for Large Output Spaces

Project 22: Reconstructing Crisis Narratives for Trustworthy Communication and Cooperative Agency

Project 23. Open doctoral student position in Keijo Heljanko's group - Massively parallel distributing computing

 

FCAI topics

Finnish Center for Artificial Intelligence FCAI is a community of experts that brings together top talents in academia, industry and public sector to solve real-life problems using both existing and novel AI. FCAI’s research mission is to create a new type of AI that is data efficient, trustworthy, and understandable. We aim to build AI systems capable of helping their users in AI-assisted decision-making, design and modeling. We develop AI techniques needed for systems which can help their users make better decisions and design better solutions across a range of tasks from personalized medicine to materials design. A core insight in developing such AIs is that they need to have world models for understanding the world and interacting with it, and user models for understanding the user and interacting with them. Read more about our research here.

Project 24: Computational modeling of human motivation and experience

Project 25: Computational cognitive models for AI assistance

Project 26: AI-assisted modelling of dynamic interactional data

Project 27: Deep Quantum Graphical Models

Project 28: Atmospheric AI

Project 29: Neuroadaptive interfaces for generative brain-computer design

Project 30: Interactive reward elicitation

Project 31: Graph based world models for sample efficient and human friendly reinforcement learning

Project 32: Planning to learn world-models to plan and learn

Project 33: Transferable hierarchical reinforcement learning

Project 34: Context based curriculum learning for safe exploration

Project 35: Data augmentation, noise and active learning — A Bayesian brain approach

Project 36: Bayesian deep active learning for amortized inference of simulator models

Project 37: Fast active-sampling approximate Bayesian inference for everyone

Project 38: Prior constraints in probabilistic programming

Project 39: Visualization in modeling workflow

Project 40: Computer assisted Bayesian workflow

 

Project 1. Large-scale computing on modern architectures and systems: Programming models, Designs and Optimization

Supervisor: Prof. Hong-Linh Truong, Aalto Systems and Services Engineering Analytics (http://rdsea.github.io) (Department of Computer Science, Aalto University)

The topic will focus on researching and developing techniques and tools for large-scale computing on modern architectures and systems. We will investigate emerging programming models, such as function-as-a-service, hybrid workflows of traditional data analysis pipelines and machine learning pipelines, and human-in-the-loop in large-scale, complex analysis. Novel software techniques and methods will be developed for support the developer to design, manage and optimize large-scale computing and data analysis applications. Observability, performance, elasticity, fairness and interpretability will be particular features in the focus of the design and optimization.

Large-scale computing will be targeted to emerging solutions for large-scale and data-intensive applications, which combine traditional workflow-based large-scale data analysis with machine learning, whereas modern systems will be high-performance computing systems with CPU/GPU, large-scale containerized systems, systems with AI accelerators, and potentially quantum computing systems.

 

Project 2. Doctoral students for ML Engineering research

Prof. Jukka K Nurminen (jukka.k.nurminen@helsinki.fi) & Prof. Tommi Mikkonen (tommi.mikkonen@helsinki.fi), Department of Computer Science, University of Helsinki

We are looking for doctoral students to work on tools and methodologies for the software engineering of machine learning systems. To ensure that machine learning systems work for real, new ways are needed to ensure their correct and efficient operation as well as their smooth development, monitoring, and maintenance. At the moment we are running or about to start multiple European and national projects focusing on testing of AI systems, MLOps, and big data analytics.

The aim of our empirical and experimental approach is to come up with new and improved solutions for ML system development and operation. The candidate is expected to analyze, measure, and model alternative approaches and create new ideas and insights based on those. This includes implementing research prototypes to try out ideas and to collect and analyze data. Documenting the results in scientific papers is naturally important. In addition to research, the doctoral student is expected to contribute to other common academic tasks such as teaching, generation of new research ideas, and writing grant proposals.

 

Project 3: Query processing and optimization for next-generation databases

Supervisor: Prof. Jiaheng Lu (Department of Computer Science, University of Helsinki)

As more businesses realized that data is critical to making the best possible decisions, we see the continued growth of systems that support a massive volume of non-relational or unstructured data. We are looking for a new Ph.D. student to develop novel algorithms and theories for a unified database management system to manage both well-structured data and NoSQL data in the era of big data. Link to a webpage with further information: https://www.helsinki.fi/en/researchgroups/unified-database-management-systems-udbms

 

Project 4: Atmospheric AI (The exploratory data analysis group)

Supervisor: Assoc. Prof. Kai Puolamäki (Department of Computer Science and Institute for Atmospheric and Earth System Research (INAR), University of Helsinki)

The exploratory data analysis group is looking for a doctoral student for a project that focuses on the use of artificial intelligence and machine learning on natural sciences, especially in atmospheric and earth sciences. The objective is to build probabilistic models of simulated and measured phenomena and develop methods of interactive artificial intelligence that would enable the substance area experts to build these models. The project can be tailored to focus more on computer science or atmospheric sciences, depending on qualifications and preferences of the applicant. The project will be done in collaboration with the Institute for Atmospheric and Earth System Research (INAR) at the University of Helsinki. Please contact Prof. Kai Puolamäki at kai.puolamaki@helsinki.fi for further information.

 

Project 5: Theory of safe algorithms and their applications in Bioinformatics

Supervisor: Assoc. Prof. Alexandru Tomescu (Department of Computer Science, University of Helsinki) https://www.cs.helsinki.fi/u/tomescu/ 

For an algorithmic problem admitting multiple solutions, a "safe algorithm" is one reporting only those partial solutions that are common to *all* solution to the problem. Thus, the output of a safe algorithm is most likely correct. This is a novel algorithmic perspective from which to tackle both theoretical and practical problems, and could revolutionize the field of Bioinformatics.

We are looking for a PhD student to study safe algorithms for both (1) string problems (aligning biological sequences, with applications in understanding protein function across the tree of life), and (2) graph-theoretic problems, with applications in assembling viral strains (such as for SARS-CoV-2 data). We expect the candidate to have expertise in both algorithms & theory and programming, through previous relevant coursework or projects.

The PhD student will join the Graph Algorithms team (https://www.helsinki.fi/en/researchgroups/algorithmic-bioinformatics/teams/graph-algorithms), which is part of the wider Algorithmic Bioinformatics group at the University of Helsinki.

 

Project 6: Probabilistic real-time machine learning

Supervisor: Prof. Arno Solin (Department of Computer Science, Aalto University)

We are looking for exceptional and highly motivated doctoral students to work on algorithms and applications for real-time machine learning. Central topics and themes in this project include approximate inference methods, stochastic differential equations (incl. neural SDEs), state space modelling, and Gaussian processes. Applications of interests are in online decision-making, sensor fusion, audio/video analysis, control (also linking to RL), and robotics. An intuition of the theoretical backdrop is provided in this ICML tutorial (https://youtu.be/vTRD03_yReI) and this ICML paper (https://arxiv.org/abs/2007.05994).

The work will be done in close collaboration with the supervisor and other members of the team at Aalto University. Doctoral students in the group are encouraged to make research or internship visits to collaborating universities/companies during the course of study. Successful candidates are expected to have completed a Masters degree and have familiarity with machine learning and statistics.

For more information and recent publications and pre-prints, see the research group home page at http://arno.solin.fi

 

Project 7: Deep learning with probabilistic principles

Supervisor: Prof. Arno Solin (Department of Computer Science, Aalto University)

We are looking for exceptional and highly motivated doctoral students to work on algorithms and methods in combining probabilistic methods with deep learning. This project relates to various topics in this space: Bayesian deep learning, computer vision, generative models, meta-learning, Gaussian and neural processes, and normalizing flows. This project is part of a consortium project with Tampere University, with collaborators in Cambridge, Oxford, Prague, and Moscow. Examples of recent work in the group in this area include this ICCV paper (https://aaltoml.github.io/GP-MVS), this NeurIPS paper (https://arxiv.org/abs/1912.10321), and this NeurIPS paper (https://arxiv.org/abs/2010.09494).

The work will be done in close collaboration with the supervisor and other members of the team at Aalto University. Doctoral students in the group are encouraged to make research or internship visits to collaborating universities/companies during the course of study. Successful candidates are expected to have completed a Masters degree and have familiarity with machine learning and statistics.

For more information and recent publications and pre-prints, see the research group home page at http://arno.solin.fi

 

Project 8. Deep learning with differential equations

Supervisor: Markus Heinonen (Department of Computer Science, Aalto University)

We are looking for an exceptional and motivated phd student to push the boundaries of deep learning with differential equations. In conventional deep learning the inputs are transformed by a sequence of layers, while an alternative paradigm emerged recently interpreting learning tasks as continuous flows with ODEs or SDEs. We aim at developing new ways to perform machine learning by repurposing differential equations. Topics range from developing interpretable neural ODEs for supervised tasks, to modelling distributional structures and data augmentations with SDEs and PDEs, to adversarial, robust or probabilistic ODEs.

The work continues on the foundations of our NIPS'19 and AISTATS'19 publications "ODE2VAE" and "Differential Gaussian processes". The work will be done in collaboration with several researchers studying the topic at Aalto and internationally. Sufficient familiarity with statistics, math, physics or machine learning is advantageous.

 

Project 9: Sample-efficient probabilistic machine learning

Supervisor: Prof. Luigi Acerbi (Department of Computer Science, University of Helsinki)

We are looking for an exceptional doctoral student eager to work on new machine learning methods for smart, robust, sample-efficient probabilistic inference, with applications in scientific modeling (e.g., computational and cognitive neuroscience). In our group, we are interested in developing novel approaches for building approximate Bayesian posteriors using only a small number of likelihood evaluations, which can be a game-changer for complex models or when resources are limited. Think of Bayesian optimization, but scaled up to full Bayesian inference. A state-of-the-art framework being developed in our group is Variational Bayesian Monte Carlo (VBMC), which combines Gaussian process surrogates, active learning, variational inference and Bayesian quadrature (Acerbi, NeurIPS, 2018, 2020). 

Promising thesis projects include extending the representational power of VBMC (e.g., discrete variables, more complex posteriors, higher dimension); exploiting recent advances in Gaussian process inference for superior scalability; and exploring the theoretical properties of the framework. Successful candidates are expected to have completed a Masters degree and have familiarity with machine learning and Bayesian statistics.

For more information and recent publications, see the research group home page at http://www.helsinki.fi/machine-and-human-intelligence

 

Project 10. Probabilistic modelling and Bayesian machine learning

Supervisor: Prof. Samuel Kaski, (Department of Computer Science, Aalto University)

I am looking for a doctoral student eager to join the Aalto Probabilistic Machine Learning Group, to work on new probabilistic models and inference techniques. Particularly promising thesis topics are: (1) simulator-based inference, for combining first-principles models with learning from data, (2) Bayesian deep learning, (3) Bayesian reinforcement learning and inverse reinforcement learning, (4) multi-agent modelling, and (4) privacy-preserving machine learning and synthetic data generation. The group has excellent opportunities for collaboration with topnotch partners in multiple applications, including user interaction, health, genomics, and neuroscience. Links: http://research.cs.aalto.fi/pml

 

Project 11: Improve drug design with human assisted AI

Supervisors: Prof. Samuel Kaski, Research Fellow Markus Heinonen (Department of Computer Science, Aalto University)

Not all relevant knowledge is explicitly accessible and usable for machine learning modelling in drug design. A new and emerging area in machine learning is knowledge elicitation from human experts to improve the prediction accuracy of models so called human-in-the-loop-modelling. The goal of this PhD studentship is to develop human-in-loop machine learning models applicable for drug design. 

The PhD student will join an ongoing collaboration between Aalto and AstraZeneca (Gothenburg, Sweden). The main task will be to develop human-in-the-loop modelling so it can be used to guide deep learning based drug design. The student will be developing systems to query drug experts for information to improve machine learning models. The candidates are expected to have background in machine learning, statistics or mathematics, with keen interest on learning chemistry.

 

Project 12: Decomposable latent representations for toxicity prediction

Supervisors: Prof. Samuel Kaski, Research Fellow Markus Heinonen (Department of Computer Science, Aalto University)

We are looking for an exceptional and motivated phd student to develop machine learning methods for drug discovery, with a specific target of developing decomposable latent representations for in-vivo toxicity prediction. We aim at ameliorating the key challenges of drug modelling under scarce data by developing interpretable generative models with intuitive chemical or clinical explanations. The developed methods will be based on Bayesian and deep learning principles, following the successes of VAEs and GANs.

The research will be conducted and supervised jointly by the two sites of Aalto and Janssen pharmaceuticals (Belgium) under an ongoing collaboration. The candidates are expected to have background in machine learning, statistics or mathematics, with keen interest on learning chemistry.

 

Project 13: Bayesian deep learning

Supervisors: Prof. Samuel Kaski (samuel.kaski@aalto.fi), Research Fellow Markus Heinonen (markus.o.heinonen@aalto.fi) (Department of Computer Science, Aalto University)

We are looking for a doctoral student to join the Aalto Probabilistic Machine Learning Group, to work on developing state-of-the-art Bayesian deep learning. Key research questions are about more useful neural parameterisations, process priors on function spaces and more efficient probabilistic inference methods for deep neural networks. Possible applications range from large-scale image classification to sample-efficient Bayesian reinforcement learning and robotics.

This work will build on top of existing research lines in the group on RL and BNNs, with a recent highlight work of implicit BNNs with state-of-the-art ImageNet performance while maintaining Bayesian principles. The group has excellent collaboration and application opportunities. Background in machine learning, statistics or math is expected.

 

Project 14: Mobile Cross Reality through Immersive Computing (MeXICO)

Supervisors: Prof. Mario Di Francesco, Matti Siekkinen (Department of Computer Science, Aalto University)

The MeXICO project aims to overcome the limited resources of mobile devices for novel applications in cross-reality (XR): virtual, augmented, and mixed reality. Current solutions for interactive and mobile XR require real-time rendering but they are constrained by the limited resources of mobile devices. A promising approach to overcome this issue is to offload most of the heavy computing tasks from the mobile device to a remote processing unit in the cloud or at the edge.

Unfortunately, this approach faces challenges in terms of both latency and bandwidth. In fact, a noticeable motion-to-photon latency is highly detrimental in interactive applications, as it may even cause severe discomfort in addition to a poor user experience. Moreover, transmitting high-quality graphic content from remote servers to mobile devices requires a large amount of network bandwidth. Motivated by these challenges, MeXICO explores solutions for mobile and distributed XR to enable novel and effective applications.

 

Project 15: HAIC: Open doctoral student position in Prof. Janne Lindqvist’s group – security engineering and usable security
 

Supervisor: Prof. Janne Lindqvist https://www.aalto.fi/en/janne-lindqvist (Aalto University, Department of Computer Science, Director Helsinki-Aalto Institute for Cybersecurity (HAIC) HAIC.fi)

We are looking for PhD students interested in security engineering, usable security and human-computer interaction. Background and interest in systems security, security engineering, data science, machine learning, modeling, human-computer interaction or social and behavioral sciences is required. PhD topic will be agreed together with the applicant. Examples of work done in the group can be found at the website for the group https://www.lindqvistlab.org/. The students will also get to participate in the activities of HAIC. Please contact me at the aalto.fi email address about these positions.



Project 16: Open doctoral student position in Prof. Janne Lindqvist’s group – understanding video streaming user experiences

Supervisor: Prof. Janne Lindqvist https://www.aalto.fi/en/janne-lindqvist (Aalto University, Department of Computer Science, Director Helsinki-Aalto Institute for Cybersecurity(HAIC) HAIC.fi)

We are looking for doctoral students interested in understanding video streaming user experiences. Background and interest in measuring user experience, modeling, human-computer interaction, computer science or social and behavioral sciences is required.  Examples of work done in the group can be found at the website for the group https://www.lindqvistlab.org/. Please contact me at the aalto.fi email address about these positions.

 

Project 17: HAIC: Open doctoral student position in Prof. Janne Lindqvist’s group – artificial intelligence and machine learning for systems security and privacy

Supervisor: Prof. Janne Lindqvist https://www.aalto.fi/en/janne-lindqvist  (Aalto University, Department of Computer Science, Director Helsinki-Aalto Institute for Cybersecurity (HAIC) HAIC.fi)

We are looking for doctoral students interested in developing novel artificial intelligence and machine learning approaches to security engineering and systems security and privacy.. Background and interest in data science, machine learning, statistics and computational approaches to computer science are required. Examples of work done in the group can be found at the old website for the group https://www.lindqvistlab.org/. The students will also get to participate in the activities of HAIC. Please see specific examples also http://jannelindqvist.com/publications/IMWUT19-fails.pdf http://jannelindqvist.com/publications/NDSS19-robustmetrics.pdf  Please contact me at the aalto.fi email address about these positions.

 

Project 18: Open doctoral student position in Prof. Janne Lindqvist’s group – multitasking and productivity tools

Supervisor: Prof. Janne Lindqvist https://www.aalto.fi/en/janne-lindqvist (Aalto University, Department of Computer Science, Director Helsinki-Aalto Institute for Cybersecurity (HAIC) HAIC.fi)

We are looking for doctoral students interested in understanding productivity tools and multitasking. Background and interest in measuring user experience, modeling, human-computer interaction, computer science or social and behavioral sciences is required.  Examples of work done in the group can be found at the website for the group https://www.lindqvistlab.org/. Please contact me at the aalto.fi email address about these positions.

 

Project 19: Open doctoral student position in Prof. Janne Lindqvist’s group – mixed methods HCI and security research

Supervisor: Prof. Janne Lindqvist https://www.aalto.fi/en/janne-lindqvist (Aalto University, Department of Computer Science, Director Helsinki-Aalto Institute for Cybersecurity (HAIC) HAIC.fi)

We are looking for doctoral students interested in pushing the envelope in mixed methods HCI and security research. Background and interest in measuring either qualitative methods or quantitative methods, and interested to learning new methods, user experience, modeling, human-computer interaction, computer science or social and behavioral sciences is required.  Examples of work done in the group can be found at the website for the group https://www.lindqvistlab.org/. Please contact me at the aalto.fi email address about these positions.

 

Project 20: Deep learning for electronic health records

Supervisor: Prof. Pekka Marttinen (pekka.marttinen@aalto.fi) (Department of Computer Science, Aalto University)

We will develop novel deep learning models for healthcare time series data. This work is done in collaboration with the Finnish Institute for Health and Welfare and other healthcare data holder in the Helsinki region. In particular, we will focus on 1) probabilistic deep learning, to address uncertainty in the predictions, 2) causal inference, required for decision making on individual and population levels, and 3) interpretability, which helps communicate the results to patients and policy makers. Also important are topics related to privacy and anonymization. The advances made in the project are central for the trustworthiness and acceptability of the methods in practice.

 

Project 21. Theoretical Framework and Deep Learning Algorithms for Large Output Spaces

Supervisor: Prof. Rohit Babbar (Department of Computer Science, Aalto University)

There is a growing interest in supervised learning for classification of data with large number of outputs or labels. However, there still lacks good theoretical understanding of algorithms in this domain with only few works such as [1,2]. Unlike other domains, the algorithmic success of deep learning methods for large output spaces has also been somewhat limited. The goal of the PhD thesis would be to explore novel theoretical frameworks and deep architectures in this domain. The scope of the project also includes exploring connections with adversarial robustness analysis of learning algorithms [3,4]. Requirements: Background in Linear Algebra, probability and Optimization. Programming experience with python. Deep learning frameworks such as PyTorch, and working with large datasets. 

  • [1] Stochastic Negative Mining for Learning with Large Output Spaces, AISTATS 2019
  • [2] A no-regret generalization of hierarchical softmax to extreme multi-label classification, NIPS 2018
  • [3] Robustness May Be at Odds with Accuracy, ICLR 2019
  • [4] Data scarcity, robustness and extreme multi-label classification, Machine Learning Journal, 2019

 

Project 22. Reconstructing Crisis Narratives for Trustworthy Communication and Cooperative Agency

Supervisor: Nitin Sawhney, Professor of Practice, Department of Computer Science, Aalto University

The research project, jointly conducted between Aalto University and the Finnish Institute for Health and Welfare (THL), proposes to analyze and reconstruct crisis narratives using mixed-methods, combining qualitative research for narrative inquiry with computational data analytics of crisis discourses in news and social media to understand global pandemics. Candidates will work at the intersection of Human-Computer Interaction (HCI), design research, computational social sciences, and public health for critical societal impact. We expect the candidate to have background in computer science, media and communication studies, social science, or similar disciplines.

Potential duties and tasks may include the following (to be conducted as part of the research team):

1. Examining the narratives emerging in crisis-related communication using qualitative research methods across various data sources including organization communications, news/media coverage, and social media exchange among diverse publics.

2. Automating content analysis for narrative work using suitable machine learning techniques for Natural Language Processing (NLP) such as Conversation Analysis, Content Classification, and/or Sentiment Analysis. This includes collecting and curating datasets, devising suitable methodologies, setting up the research infrastructure and tools, and a pipeline for data extraction, analysis and validation.

3. Representing and visualizing crisis narratives to support understanding and collaborative sensemaking among key stakeholders and diverse publics. These tools should support browsing, searching, and exploring information for selected crisis themes and narratives. Work in this area includes not only developing prototypes of visualizations, but also conducting design research, user experience (UX) evaluation, and pilot assessment of such tools.

The candidate is not expected to master all these domains, but work closely with a multi-disciplinary research team to lead design and development efforts, while learning and contributing to ongoing work in specific research areas of interest.

The position belongs to the Aalto career system and the selected person will be appointed for a two-year fixed term appointment with an option for renewal.

 

Project 23. Open doctoral student position in Keijo Heljanko's group - Massively parallel distributing computing

Supervisor: Prof. Keijo Heljanko, Department of Computer Science, University of Helsinki and HiDATA Helsinki Centre for Data Science (https://www.helsinki.fi/en/helsinki-centre-for-data-science)

The PhD Student position is in the context of the Academy of Finland project "Design and Verification Methods for Massively Parallel Distributed Systems (DeVeMaPa)". The main focus of the project is on the use of GPU computing to accelerate Big Data processing and the theoretical foundations of these systems.

The project will develop methodology for the design and verification of massively parallel heterogeneous computing. We need new methods to support the massive increase in the amount of parallelism at all levels of the hardware/software stack. Such massive increases in parallelism will make some currently used programming paradigms infeasible and thus new methods need to be devised to cope with industrial Big Data use cases. These methods must also be accompanied with solid theoretical foundations, allowing for the development of automated testing and verification tools that are required to validate the parallelization runtime software before production deployment. A key challenge is the need for seamless integration of heterogeneous computing with GPUs and hardware accelerators (e.g., neural network accelerators), how can they all be handled in a unified Big Data programming framework?

 

FCAI topics

Finnish Center for Artificial Intelligence FCAI is a community of experts that brings together top talents in academia, industry and public sector to solve real-life problems using both existing and novel AI. FCAI’s research mission is to create a new type of AI that is data efficient, trustworthy, and understandable. We aim to build AI systems capable of helping their users in AI-assisted decision-making, design and modeling. We develop AI techniques needed for systems which can help their users make better decisions and design better solutions across a range of tasks from personalized medicine to materials design. A core insight in developing such AIs is that they need to have world models for understanding the world and interacting with it, and user models for understanding the user and interacting with them. Read more about our research here.

 

Project 24: Computational modeling of human motivation and experience

One of our aims at FCAI is ‘artificial human understanding’, a type of AI that much better estimate what human collaborators want or prefer or experience. The goal is to estimate the deeper (latent) motivational processes and subjective constructs like perceived aesthetics, and thereby offer help to human collaborators more effectively. We are looking for an outstanding researcher to help us construct computational models of motivations and experience rooted in most current psychological theory, and use those models to boost the inferential capability of AI assistants. In particular, we will model designers as agents that maximize expected utility in their choices but are limited by their own bounds and that of the design tools (see Gershman et al. Science 2015). This area, called computational rationality, is an exciting convergence point of machine learning and cognitive science. The ideal candidate has demonstrated track record modeling user behavior or cognition, with publications on reinforcement learning, probabilistic models, or cognitive models.

Supervision: Antti Oulasvirta (Aalto University); potential other supervisors: Perttu Hämäläinen (Aalto University), Arto Klami (University of Helsinki), Elisa Mekler (Aalto University)

Keywords: computational modeling, human motivation, computational rationality, interactive AI, reinforcement learning, cognitive models

Previous papers: Roohi, Shaghayegh, et al. "Predicting Game Difficulty and Churn Without Players." Proceedings of the Annual Symposium on Computer-Human Interaction in Play. 2020.

 

Project 25: Computational cognitive models for AI assistance

A prime goal for FCAI's research is to develop a new form of AI that can better work with people and assist them in everyday tasks. We believe that deep integration of human cognition into the technical principles that govern the AI’s operation is necessary for ethically acceptable applications. We are now looking for an outstanding, methodologically oriented candidate to help us develop computational cognitive models based on the theory of computational rationality (Gershman 2015 Science). The ideal candidate has previous background in MDP/POMDP-based modeling, decision theory, multi-agent systems, or reinforcement learning.

Previous papers:

  • Kangasrääsiö, Antti, et al. "Parameter inference for computational cognitive models with Approximate Bayesian Computation." Cognitive Science 43.6 (2019): e12738.
  • Gebhardt, C., Oulasvirta, A., & Hilliges, O. (2020). Hierarchical Reinforcement Learning as a Model of Human Task Interleaving. arXiv preprint arXiv:2001.02122.

Supervision: Antti Oulasvirta (Aalto University); Samuel Kaski (Aalto University)

Keywords: Computational cognitive modeling, AI assistance, interactive AI

 

Project 26: AI-assisted modelling of dynamic interactional data

This project is concerned with modelling on graphs where the environment is uncertain and can change, possibly due to user interaction [1]. Graph neural networks, the state-of-the-art models for embedding graphs, have almost exclusively focused on non-strategic settings. However, several important applications involve agents on networks that compete/collaborate [2] as part of decision-making. This project is concerned with modeling the uncertainties due to the different players. We are looking for doctoral students to work on the intersection of graph neural networks, Gaussian processes, and dynamical systems (see [3] for an overview). The ideal candidate will have a strong mathematical/statistical training and good programming skills.

References:

  • [1] Ruotsalo, Tuukka, Jacucci, Giulio, Myllymäki, Petri, and Kaski, Samuel. ”Interactive intent modeling: information discovery beyond search”. Communications of the ACM, 58(1):86-92, 2015.
  • [2] Garg, Vikas, and Jaakkola, Tommi. "Predicting deliberative outcomes." In International Conference on Machine Learning, pp. 3408-3418. PMLR, 2020. Link: https://www.mit.edu/~vgarg/DeliberativeOutcomes_CameraReady.pdf
  • [3] Solin, Arno. “Machine Learning with Signal Processing”. ICML 2020 tutorial. Link: https://youtu.be/vTRD03_yReI

Supervision: Samuel Kaski (Aalto University), Vikas Garg (Aalto University) and Arno Solin (Aalto University)

Keywords: graph neural networks, dynamical models, Gaussian processes, stochastic differential equations, Bayesian methods                              

 

Project 27: Deep Quantum Graphical Models

Probabilistic graphical models (PGMs) such as Markov Random Fields and Bayesian networks allow us to encode the conditional dependencies between random variables succinctly. However, these models are designed to run on the classical computer. This project is aimed at designing quantum counterparts for PGMs, and implementing them via deep architectures such as graph neural networks [1] that can be executed on a quantum computer. Another goal of the project is to investigate whether it might be possible to simulate these algorithms on a classical computer under some restrictions [2]. We are looking for doctoral students to work at the intersection of quantum machine learning, graphical models, and deep learning. The ideal candidate will have a strong mathematical inclination and training, and excellent programming skills.

References:

Supervision: Vikas Garg (Aalto University) and Juho Kannala (Aalto University)

Keywords: graphical models, quantum computing, deep learning

 

Project 28: Atmospheric AI

Artificial intelligence (AI) and machine learning (ML) are making their inroads to atmospheric and earth sciences. There are lots of opportunities to do research in physical sciences more efficiently and to obtain novel results of high impact—both in atmospheric and computer sciences—by developing and applying novel AI methods to solve scientific problems. In this project, we plan to build probabilistic models of measured and simulated natural world phenomena, trained by using simulator outputs or real-world observations, which allow us for example replace computationally expensive simulator runs with faster ML computations, to fill in missing data from observations, and to better understand complex systems and processes and underlying causal relations. Our objective is to also model the interactive data analysis and model building process of the substance area experts (here atmospheric scientists), which allows us to address problems such as how to design the exploratory data analysis workflows and systems and how to best incorporate the knowledge and insights of the experts into the model building process. We are looking for an atmospheric scientist with interest in AI, or a computer scientist who wants to develop AI methodology and work with physics-related applications. We can adjust the work plan and the supervision arrangement depending on the qualifications and interests of the hired person.

Supervision: Kai Puolamäki (University of Helsinki); potential co-supervisors Hanna Vehkamäki (University of Helsinki), Leena Järvi (University of Helsinki), Tuomo Nieminen (University of Helsinki)

Keywords: Atmospheric and earth sciences; exploratory data analysis; automatic experimental design; interactive user modelling; causal inference

 

Project 29: Neuroadaptive interfaces for generative brain-computer design

The project aims to establish the scientific foundations for generative brain-computer design. The project combines 1) human preference and critique estimation directly from the human brain manifested as implicit natural human reactions evoked in response to generative model 2) use critique as input for generative models producing that information to a) derive objectives for learning generative models, b) to adjust generative model outputs, and c) develop self-supervised brain-computer interfaces for 3) interactive design. The work is based on our fundamental research on novel neuroadaptive brain-computer interfaces [1,2,3,4] and is conducted in close collaboration with neuroscientists.

The PhD student will develop new types of deep learning methods that can jointly decode human brain responses and associate the decoding process to adapt and learn generative latent models that produce stimuli information evoking the brain responses. and evaluate and test the models in human-computer and brain-computer interaction settings. The student should have experience in deep learning and the associated software development with TensorFlow or similar frameworks. A basic knowledge, or interest to learn, cognitive neuroscience, especially EEG and fNIRS brain imaging methods, are an advantage.

  • [1] Kangassalo Lauri; Spapé, Michiel; Ruotsalo, Tuukka. Neuroadaptive modelling for generating images matching perceptual categories. Scientific reports (Nature), 2020, 10.1: 1-10.
  • [2] de la Torre-Ortiz, Carlos, et al. Brain Relevance Feedback for Interactive Image Generation. In: Proceedings of the 33rd Annual ACM Symposium on User Interface Software and Technology. 2020. p. 1060-1070.
  • [3] Kangassalo, L., Spapé, M., Ravaja, N., & Ruotsalo, T. information gain modulates brain activity evoked by reading. Scientific reports (Nature), 2020, 10.1, 1-10.
  • [4] Davis III, Keith M., et al. Brainsourcing: Crowdsourcing Recognition Tasks via Collaborative Brain-Computer Interfacing. In: Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems. 2020. p. 1-14.

Supervision: Tuukka Ruotsalo (University of Helsinki) and Jaakko Lehtinen (Aalto University). More information about related research is available at: https://www.cs.helsinki.fi/group/intercom/. The project also involves cooperation with University of Copenhagen.

Keywords: brain-computer interfaces, brain feedback, generative adversarial neural networks, deep autoencoders, design

 

Project 30: Interactive reward elicitation

A major problem in RL is to determine suitable reward functions. Reward design often requires a significant amount of trial-and-error even for experts with experience. Our core idea is to include a model of the user in the reward elicitation process such that the process can take into account the user’s limitations in addition to maximizing information gain, thus moving beyond the noisy oracle paradigm. Moreover, we plan to integrate behavioural (explicit feedback from user) as well as implicit physiological measurements.

Supervision: Ville Kyrki (Aalto University), Simo Särkkä (Aalto University), Joni Pajarinen (Aalto University), Antti Oulasvirta (Aalto University), Samuel Kaski (Aalto University)

Keywords: Reinforcement learning, reward design, reward elicitation, interactive AI, inverse RL

 

Project 31: Graph based world models for sample efficient and human friendly reinforcement learning

Reinforcement learning has shown promise in computer game play and robotics but learning long-term behavior directly from visual input has been limited to simple tasks and actual task specification has been hard for regular users. Meanwhile deep learning has been able to infer semantic information about objects and their dependencies from visual input in the form of object graphs. To make long-term planning more efficient we will use graphs as dynamic states in reinforcement learning. We will learn models of how the state of the system, that is, the graph changes on agent actions. Using the learned world model together with a reinforcement learning policy representation such as a graph neural network allows the system to generalize over different world scenes. Moreover, the approach allows for measuring closeness of the current scene to the desired one in the form of graph similarity measures. Thus, a non-expert user can specify the task objective in terms of visual scenes, for example, by uploading a set of photos of the desired end state or of states that the algorithm should avoid which are then converted to graphs. We provide good opportunities for applying the methods in mobile robotic manipulation and autonomous driving.

Supervision: Joni Pajarinen (Aalto University), Alexander Ilin (Aalto University), Juho Kannala (Aalto University)

Keywords: Reinforcement learning, model learning, planning, computer vision, decision-making, human feedback, robotics               

 

Project 32: Planning to learn world-models to plan and learn

Planning to reach long-term goals has allowed for super-human performance in tasks with accurate models [1] while learning has allowed for solving tasks with complex inputs and outputs [2]. However, tasks with complex inputs and outputs that require long term exploration and planning are out of reach for current methods. We will go further and integrate learning and planning. We will use meta-learning [3] to learn a world model and a policy that mixes planning and reinforcement learning such that planning is used for problem parts that require principled exploration and learning is used for parts that require pattern recognition. We will develop methods that integrate learned dynamics models and planning with both model-free and model-based reinforcement learning. We expect the developed methods to enable efficient long-term decision making in high-dimensional continuous and discrete control tasks. We provide good opportunities for applying the methods in mobile robotic manipulation and autonomous driving.

  • [1] Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., Hubert, T., Baker, L., Lai, M., Bolton, A. and Chen, Y., 2017. Mastering the game of Go without human knowledge. Nature, 550(7676), pp.354-359.
  • [2] Hessel, M., Modayil, J., van Hasselt, H., Schaul, T., Ostrovski, G., Dabney, W., Horgan, D., Piot, B., Azar, M.G. and Silver, D., 2018. Rainbow: Combining Improvements in Deep Reinforcement Learning. In AAAI.
  • [3] Finn, C., Abbeel, P. and Levine, S., 2017. Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks. In ICML.

Supervision: Joni Pajarinen (Aalto University), Alexander Ilin (Aalto University), Simo Särkkä (Aalto University), Ville Kyrki (Aalto University)

Keywords: Reinforcement learning, machine learning, learning dynamics, control, planning, decision-making

 

Project 33: Transferable hierarchical reinforcement learning

We tackle the general problem of learning to make high value decisions in dynamic systems at different time-scales. Major challenges of current approaches include high sample complexity, problem specificity of learned solutions, and opaquity of models. We will develop efficient hierarchical decision making models that allow 1) learning parts of the model separately for efficiency, 2) transferring the model to novel situations including transfer from simulation to the real world, and 3) demonstrating model parts and their interplay to users to bridge the human-computer gap. Contrary to prior work that does not learn models or learns only low level models we learn models on all levels allowing for quick transfer to new tasks and interpretability for users. We are looking for a researcher interested in developing the new decision making models and methods, with options on applying the techniques to autonomous driving and robotics.

Supervision: Joni Pajarinen (Aalto University), Alexander Ilin (Aalto University)

Keywords: Reinforcement learning, learning dynamics, control, machine learning, planning, decision making under uncertainty

 

Project 34: Context based curriculum learning for safe exploration

Contrary to state-of-the-art curriculum learning that incrementally adapts the underlying task to make learning more efficient [1], we will use curriculum learning to also improve safety in exploration on real systems. We will use context based curriculum learning where the agent controls the context. By controlling the context the agent can explore different behaviors without risking the system or users. In human-robot interaction, the context can be defined in terms of the user, for example, user pose or pose of an object the user is holding. To change the context, the robot asks the human to execute the desired pose or hand out a specific object. With this approach we expect to be able to learn behavior safely on challenging physical systems. Technically the approach will require probabilistic modeling of a safety index (for example, safety margin distance between user and robot) that is learned interactively with a user that controls the context. Then exploration in reinforcement learning will need to integrate the safety index such that a minimum required safety is guaranteed, which can be achieved by controlling the learning curriculum.

[1] Klink, P., D'Eramo, C., Peters, J. and Pajarinen, J., 2020. Self-Paced Deep Reinforcement Learning. Advances in Neural Information Processing Systems (NeurIPS)

Supervision: Joni Pajarinen (Aalto University), Ville Kyrki (Aalto University)

Keywords: Reinforcement learning, curriculum learning, safety, planning, decision making under uncertainty

 

Project 35: Data augmentation, noise and active learning — A Bayesian brain approach

Data augmentation is crucial for modern deep learning methods. However, it is usually done in a very heuristic and manually designed way. Recently, the possibility of learning to do such data augmentation from data, especially in a probabilistic framework, has received increasing attention. Interestingly, something akin to data augmentation occurs naturally in biological brains, which rely on noisy sensory inputs and noisy neural circuits. Biological systems learn to push such noise in directions that leave the underlying inference invariant, effectively learning the local equivariance structure of the inputs. Another important strategy adopted by biological systems is active learning. Here, our goal is to develop probabilistic models which enable learning of data augmentation, in particular by drawing inspiration from Bayesian approaches to brain function, and exploit such probabilistic representations to perform active learning. This project aims at making existing machine learning algorithms more efficient while at the same time elucidating deep connections between data augmentation, noise and Bayesian active learning in neural networks, both artificial and biological. The candidate should have an MSc degree containing a lot of mathematics.

Supervision: Aapo Hyvärinen (University of Helsinki) and Luigi Acerbi (University of Helsinki)

Keywords: Probabilistic modelling, Bayesian brain, Data augmentation, Active learning

 

Project 36: Bayesian deep active learning for amortized inference of simulator models

Recent approaches to inference in simulator models exploit the power of flexible deep neural density estimators to iteratively learn a direct mapping from summary statistics of the data to the posterior distribution, skipping the intermediate steps of approximate inference. In some limited cases, the trained networks can be immediately used on new data, achieving the holy grail of amortized Bayesian inference — inference at virtually no cost at runtime. However, training these networks requires a very large number of samples from the model, and the mapping to the posterior has no notion of uncertainty, meaning that the network could fail silently in unseen regions of parameter space. This project is concerned with applying Bayesian principles of uncertainty estimation and active learning to develop a new generation of algorithms for sample-efficient training of robust, safe emulator networks for simulator-based inference. The ideal candidate has prior experience with deep learning and Bayesian methods.

Supervision: Luigi Acerbi (University of Helsinki); Jukka Corander (University of Helsinki), Samuel Kaski (Aalto University)

Keywords: Simulator-based inference; Bayesian deep learning; active learning; neural density estimators; amortized inference

 

Project 37: Fast active-sampling approximate Bayesian inference for everyone

In recent years, a new approach to approximate Bayesian inference has emerged, on the side of the traditional workhorses (MCMC and variational inference). Active-sampling Bayesian inference aims to build posterior distributions (and approximations of the model evidence) in a sample-efficient way, by constructing a statistical surrogate of the posterior (or likelihood), such as via a Gaussian process, and then actively evaluating the log-likelihood or log-joint distribution where needed to efficiently update the surrogate model [1-3]. This approach is similar to Bayesian optimization, but the goal differs in that the goal is to learn the posterior distribution (and/or the marginal likelihood). Crucially, thanks to recent advances in Bayesian nonparametrics, this approach is not limited anymore to “expensive” models, but it could become part of the standard Bayesian workflow for many models, affording calculation of cheap, uncertainty-aware posterior approximations with only a small number of evaluations. This project is concerned with pushing the state-of-the-art of active-sampling Bayesian inference algorithms, in terms of both theory and implementation, to obtain a new instrument for approximate inference which would be widely accessible, fast and failsafe. The ideal candidate has prior experience with Gaussian processes and active learning (e.g., Bayesian optimization), both in theory and with modern software implementations (e.g., GPyTorch).

  • [1] Acerbi L (2018). Variational Bayesian Monte Carlo, NeurIPS.
  • [2] Järvenpää M., Gutmann MU, Vehtari A., and Marttinen P (2020). Parallel Gaussian process surrogate Bayesian inference with noisy likelihood evaluations. Bayesian Analysis.
  • [3] Acerbi L (2020). Variational Bayesian Monte Carlo with Noisy Likelihoods, NeurIPS.

Supervision: Luigi Acerbi (University of Helsinki); Aki Vehtari (Aalto University), Samuel Kaski (Aalto University), Arto Klami (University of Helsinki)

Keywords: Bayesian inference; active learning; Gaussian processes; Bayesian optimization

 

Project 38: Prior constraints in probabilistic programming

Using prior domain knowledge on model parameters and transformations is at the core of Bayesian modeling, helping to build more interpretable models that can be estimated from less data. This project develops richer ways of encoding prior knowledge, focusing on incorporating (soft and hard) constraints into probabilistic programs. The current tools support simple constraints (non-negativity of a parameters, or linearity of a function) but often the prior knowledge is in form of more complex constraints (e.g. monotonicity or near-linearity of a function, or permutation invariance) that remain challenging. Building on existing theoretical foundations for specific cases, you will work on developing both the theory and practical inference algorithms for handling such constraints.

The position is ideal for candidates with strong background in Bayesian modeling or machine learning. Your main task is to develop the required computational methods and ideally proceed to implement them into existing probabilistic programming tools in collaboration with others. We already have several concrete applications with such prior knowledge (e.g. physical knowledge in material design and cognitive theories in decision-making), and you will work in collaboration with other FCAI projects to apply the methods in selected interesting use-cases.

Supervision: Arto Klami (University of Helsinki); Aki Vehtari (Aalto University)

Keywords: Probabilistic programming, Bayesian modeling, Prior knowledge

 

Project 39: Visualization in modeling workflow

We develop AI techniques needed for systems which can help their users make better decisions and design better solutions across a range of tasks from personalized medicine to materials design. A core insight in developing such AIs is that they need to have world models for understanding the world and interacting with it, and user models for understanding the user and interacting with them. Many parts of the probabilistic modeling workflow benefit from visualization. This project develops tools for AI-assisted visualizations using AI which has a theory of mind of the user. The work will be built on existing theory in cognitive sciences and human-computer interaction. The goal is to generate visually appealing, task-specific, and informative visualizations with controllable complexity depending on the amount of information that is available, required, and sensible given the expertise of the user.

Supervision: Aki Vehtari (Aalto University); Antti Oulasvirta (Aalto University); Arto Klami (University of Helsinki)

Keywords: Interactive probabilistic modeling, modeling workflow, visualization, uncertainty quantification, decision-making

 

Project 40: Computer assisted Bayesian workflow

Probabilistic programming languages make it easier to specify and fit Bayesian models, but this still leaves us with many options regarding constructing, evaluating, and using these models, along with many remaining challenges in computation. Using Bayesian inference to solve real-world problems requires not only statistical skills, subject matter knowledge, and programming, but also awareness of the decisions made in the process of data analysis. All of these aspects can be understood as part of a tangled workflow of applied Bayesian statistics. Beyond inference, the workflow also includes iterative model building, model checking, validation and troubleshooting of computational problems, model understanding, and model comparison. The goal is to develop for different parts of the workflow self-diagnosing tools that can be used as part of interactive computer assisted workflow for probabilistic model building and Bayesian data analysis.

[1] Gelman, Vehtari, Simpson et al (2020). Bayesian workflow. https://arxiv.org/abs/2011.01808

Supervision: Aki Vehtari (Aalto University); Arto Klami (University of Helsinki); Antti Oulasvirta (Aalto University)

Keywords: probabilistic modeling workflow, Bayesian workflow, diagnostics

 

Application process

The Spring call 2021 includes a number of doctoral student positions in specific research projects by several HICT professors and researchers. If you wish to be considered as a potential new doctoral student in HICT, you can apply directly to the specific research projects.

An applicant can choose one or multiple projects in the application form. Enter the number(s) of the project(s) that you are interested in. Please use only numbers and list the projects in your order of preference (e.g. 9, 3, 7). Express your motivation towards the project(s) in your motivation letter (compulsory attachment). You do not have to write several motivation letters in case you apply for multiple projects, but if you prefer you can attach separate letters for individual projects and supervisors.

The online application form closes on February 2, 2021 at midnight Finnish time (23:59 EET, Eastern European Time), after which applications will be reviewed. Incomplete applications or applications arriving after the deadline will not be considered. Based on the results of the review, top candidates will be invited to interviews.

All the supervisors you indicate on your application form will be informed of your interest, and others also have access to your application documents. If your application is considered strong enough, given the limited resources and intense competition, you will be contacted for a skype interview in February or March 2021.

 

How are the applications submitted?

Applications need to be submitted through the online electronic application system of Aalto University. Applications sent through any other means will not be processed.

 

What material is required in the application?

Note that there are compulsory information/attachments that you need to include in your application (all text documents need to be provided in a single pdf file named "lastname_firstname”.pdf).

 

Compulsory attachments

Please submit your attachments as single pdf file containing (all documents in English):

  1. Letter of motivation (max. one page) Please describe your background and future plans, and in particular the reasons for selecting the project(s) (you can get more information on the projects and supervisors through their web pages). Try to make your motivation letter as convincing as possible, so that the potential supervisors get interested. You do not have to write several motivation letters in case you apply for multiple projects, but if you prefer you can attach separate letters for individual projects.
  2. A curriculum vitae and list of publications (with complete study and employment history, please see an example CV at Europass pages)
  3. A study transcript provided by the applicant's university that lists studies completed and grades achieved.
  4. A copy of the M.Sc. degree certificate. If the degree is still pending, then a plan for its completion must be provided. (The letter describing the completion plan can be free-format.)
  5. Contact details of possible referees. Please, provide names, positions, affiliations, and e-mail addresses of 2-3 senior academic people available for providing recommendation letters upon request from HICT. We will contact you and the recommenders separately afterwards, if and when recommendation letters are required. 

 

Eligibility and required documents later in the recruitment process

In the Finnish university system, a person must have a Master's degree in order to enroll for doctoral studies. In case you wish to pursue graduate studies with a B.Sc. background, please apply first to one of the participating units' Master's programmes (Aalto University School of Science (SCI) or School of Electrical Engineering (ELEC), and University of Helsinki). A number of these programmes provide special “doctoral tracks” with some financial support and study plans oriented towards continuing to doctoral education after the M.Sc. degree.

In order to get a study right for doctoral studies in Aalto University or University of Helsinki, an applicant with a Master’s degree outside of Aalto University/University of Helsinki needs to meet certain eligibility requirements and present some mandatory documents. A successful applicant must have an excellent command of Finnish, Swedish, or English. The universities participating in HICT have strict language skills requirements for doctoral students (Aalto University, University of Helsinki). All international applicants applying for doctoral studies must demonstrate their proficiency in English. For example, an English language proficiency certificate (TOEFL, IELTS, CAE/CPE) is required later in case you will proceed to the recruitment process and apply for a doctoral study right.

Only the following applicant groups can be exempted from the language test requirement: applicants who have completed a higher education degree 1) taught in Finnish, Swedish or English in a higher education institution in Finland or 2) in an English-medium programme at a higher education institution in an EU/EEA country, provided that all parts of the degree were completed in English or 3) an English-medium higher education degree requiring a physical on-site presence at a higher education institution in the United States, Canada, Great Britain, Ireland, Australia or New Zealand. More information on minimum language requirements and language test scores can be found at Master programmes admission webpage (see "Language requirements" and "demonstrating proficiency in English").

Please, be prepared to check the eligibility requirements for doctoral studies and present additional documents in case you will proceed to the recruitment and apply for doctoral study right in Aalto University or University of Helsinki.

HICT itself does not award doctoral study rights or doctoral degrees. All doctoral students within the HICT network are doctoral students of either Aalto University or University of Helsinki. Please, find below more detailed information about the eligibility requirements and doctoral studies in general at Aalto University and University of Helsinki:

For the Aalto Doctoral Programme in Science (SCI)

For the Aalto Doctoral Programme in Electrical Engineering (ELEC)

For the University of Helsinki

 

Important about COVID-19

Please, note that the coronavirus situation may constrain recruitments from abroad and thus affect exact starting dates. We also understand that because of the coronavirus situation there can be difficulties or delays in getting the required language proficiency certificates. We will solve potential situations in terms of the missing documents if you will proceed in the recruitment process and start as a new doctoral student at Aalto University or University of Helsinki.

 

Problems with the application?

First read all the above material carefully, and if your problem is still unsolved, only then send email to HICT coordinator: hict-apply@hiit.fi

Please note that our reply may take longer than usual due to holidays.

 

---

General HICT Call FAQ

 

Who is eligible to apply?

Prospective new doctoral students who are willing to start their doctoral studies in Aalto University or University of Helsinki under one of the HICT supervisors. In order to be successful in the call, the applicant needs to be an exceptional student. Although the number of positions is relatively high, we expect to receive hundreds of good applications, and the process is extremely competitive.

In the Finnish university system, a person must have a Master's degree in order to enroll for doctoral studies. Please, read more information on eligibility above from the call announcement.

Current doctoral students in Aalto University or University of Helsinki cannot participate in this call.

While all applicants who have submitted an application by the deadline will be appropriately considered, Aalto University and the University of Helsinki reserve the right to consider also other candidates for the announced positions.

 

When does the funding period start?

The exact starting date can be negotiated between the student and the supervisor. The student must have completed his/her M.Sc. degree by the time of starting doctoral studies.

New students also need to go through the standard doctoral student enrolment process of the hosting university/school before the start of the funding period. The supervisors will help in this process, once the best candidates have been identified and linked to a supervisor.

Important about COVID-19: Please, note that the coronavirus situation may constrain recruitments from abroad and thus affect exact starting dates.

 

How long is the funding period?

The maximum length of the funding period is four years.

 

How much is the grant?

The exact amount of monthly salary depends on the stage of the doctoral studies and varies between 2,000 and 3,000 euros/month. The level of the salary is sufficient for a funded student to focus on his/her doctoral studies full-time, without need to resort to other sources of income.

 

What are the duties and benefits of a doctoral student?

Funded doctoral students are typically hired as full-time employees for the duration of their doctoral studies. The contract includes the normal occupational health benefits of the employing university, and Finland has a comprehensive social security system.

The annual total workload of research and teaching staff at Finnish universities is 1624 hours. In addition to doctoral studies, persons hired are expected to participate in the supervision of students and teaching following the standard practices of the recruiting unit.

 

Where are the doctoral studies to take place?

Physically the main work location is at the department of the supervisor: in case of the Aalto departments, these are located at the Otaniemi campus while the Department of Computer Science of University of Helsinki is located at the Kumpula Science Campus. The joint research institute HIIT that coordinates the activities of HICT operates on both campuses.

Both campuses are easily reachable from Helsinki city centre by public transportation.

The distance between Otaniemi and Kumpula is only some 10 kilometres, and the participating departments collaborate frequently in research and education (in doctoral education through the HICT network). It is administratively very easy to incorporate courses from the other university as part of a doctoral degree in the other university. HICT also organises joint seminars, lecture series, workshops and other events for its students.

 

What does HICT doctoral student membership mean?

All doctoral students within the HICT network are doctoral students of either Aalto University or University of Helsinki. When a new doctoral student starts his/her doctoral studies at a HICT partner department, he or she by default becomes a HICT student member. Most of our doctoral students belong also to other doctoral education networks or programs in their home university or national and international networks. HICT itself does not award doctoral degrees. Our doctoral students graduate either from Aalto University or University of Helsinki.

 

If I come to study in Helsinki, do I need to learn Finnish?

No: the working environment of doctoral students is highly international, and the working language is English. You can normally also cope in English outside work as most Finns have a very good command of English.

More information for international applicants: