Active learning

Also see most of my papers on reinforcement learning. These include active learning (i.e. exploration) of some type.
  • Christos Dimitrakakis, Christian Savu-Krohn, 2007
    Cost-minimising strategies for data labelling: optimal stopping and active learning
    Foundations of Information and Knowledge Systems, FOIKS 2008, [Abstract] [PDF] [BibTeX] [arXiv]

Algorithmic analysis

Papers on computational complexity and loss/regret bounds in various contexts.
  • Aikaterini Mitrokotsa, Pedro Peris-Lopez, Christos Dimitrakakis and Serge Vaudenay, 2012
    On selecting the nonce length in distance bounding protocols
    The Computer Journal
  • Christos Dimitrakakis and Aikaterini Mitrokotsa, 2012
    Near-optimal Node Blacklisting in Adversarial Networks
    GameSec 2012 (poster) [Abstract] [PDF]
  • Bramert Ottens, Christos Dimitrakakis and Boi Faltings, 2012
    DUCT: An Upper Confidence Bound Approach to Distributed Constraint Optimization Problems
    AAAI 201 [PDF]
  • Christos Dimitrakakis
    Robust Bayesian reinforcement learning through tight lower bounds
    [PDF] [C++ Code] EWRL 2011.
  • Christos Dimitrakakis, Constantin Rothkopf
    Bayesian multitask inverse reinforcement learning
    [PDF] [C++ Code] EWRL 2011.
  • Christos Dimitrakakis, 2012
    Sparse Reward Processes
    arXiv:1201.2555v1 [Abstract] [PDF]
  • Christos Dimitrakakis, 2009
    Complexity of stochastic branch and bound methods for belief tree search in Bayesian reinforcement learning.
    International conference on Agents and Artificial intelligence (ICAART 2010), pages 259-254.
    [PDF] [Abstract] [PDF (short version)]
  • Christos Dimitrakakis and Michail Lagoudakis, 2008
    Algorithms and Bounds for Sampling-based Approximate Policy Iteration, [Abstract] [PDF] [BibTeX]
    EWRL'08, Lilles, France
  • Christos Dimitrakakis, Aikaterini Mitrokotsa and Serge Vaudenay, 2012
    Expected loss bounds for authentication in constrained channels
    Proceedings of the 31st IEEE International Conference on Computer Communications (INFOCOM 2012), 478-485 [Draft PDF]

Cryptography, Privacy and Security

Papers on security applications of statistics/machine learning, or on the analysis of cryptographic protocols.
  • Christos Dimitrakakis, Aikaterini Mitrokotsa and Serge Vaudenay, 2014
    Expected loss bounds for authentication in constrained channels.
    Journal of Computer Security (to appear).
  • Christos Dimitrakakis, Aikaterini Mitrokotsa
    Distance bounding protocols: are you close enough?
    IEEE Security and Privacy (to appear).
  • Elena Pagnin, Christos Dimitrakakis, Aysajan Abidin, Aikaterini Mitrokotsa. On the Leakage of Information in Biometric Authentication
    Indocrypt 2014.
  • Aikaterini Mitrokotsa, Pedro Peris-Lopez, Christos Dimitrakakis and Serge Vaudenay, 2012
    On selecting the nonce length in distance bounding protocols
    The Computer Journal
  • Christos Dimitrakakis and Aikaterini Mitrokotsa, 2012
    Near-optimal Node Blacklisting in Adversarial Networks
    GameSec 2012. [Abstract] [PDF]
  • A. Mitrokotsa, C. Dimitrakakis, 2012
    Intrusion detection in MANET using classification algorithms: the effects of cost and model selection.
    Ad-Hoc Networks, in press.
  • Christos Dimitrakakis, Aikaterini Mitrokotsa and Serge Vaudenay
    Expected loss bounds for authentication in constrained channels
    INFOCOM 2012, 478-485 [Draft PDF]
  • A. Mitrokotsa, C. Dimitrakakis, P. Peris-Lopez, J. C. Hernandez-Castro, 2009
    Reid et al.'s Distance Bounding Protocol and Mafia Fraud Attacks over Noisy Channels
    IEEE Communication Letters 14(2):121-123, 2010. [PDF] [Abstract]
  • Christos Dimitrakakis and Aikaterini Mitrokotsa, 2009
    Statistical decision making for authentication and intrusion detection
    Proceedings of the 8th international conference on machine learning and applications (ICMLA 2009), pp. 409-414. [PDF]. [Abstract].
  • Aikaterini Mitrokotsa, Christos Dimitrakakis, Christos Douligeris, 2007
    Intrusion Detection using cost-sensitive classification
    EC2ND'07

Decision Making, Reinforcement Learning and Control

This is the bulk of my work and includes standard reinforcement learning and control, as well as other problems in decision making such as inverse reinforcement learning and apprenticeship learning.
  • Christos Dimitrakakis and Nikolaos Tziortziotis, 2014. Usable ABC Reinforcement Learning
    NIPS 2014, ABC in Montreal workshop.
  • Emmanouil Androulakis and Christos Dimitrakakis, 2014. Generalised entropy MDPs and Minimax Regret
    NIPS 2014, From bad models to good policies workshop.
  • Aristide Toussou, Christos Dimitrakakis 2013
    Probabilistic inverse reinforcement learning in unknown environments
    UAI 2013
  • Christos Dimitrakakis, Nikolaos Tziortziotis,2013
    ABC Reinforcement Learning
    arXiv:1303.2506, ICML 2013 [Abstract] [PDF]
  • Nikolaos Tziortziotis, Christos Dimitrakakis, Konstantinos Blekas, 2013
    Cover tree Bayesian reinforcement learning
    (draft) arXiv:305.1809 [Abstract] [PDF]
  • Nikolaos Tziortziotis, Christos Dimitrakakis, Konstantinos Blekas, 2013
    Linear Bayesian reinforcement learning
    IJCAI 2013. [PDF]
  • Christos Dimitrakakis, 2013
    Monte-Carlo utility estimates for Bayesian reinforcement learning
    IEEE Conference on Decision and Control, arXiv:1303.2506 [Abstract] [PDF]
  • Bramert Ottens, Christos Dimitrakakis and Boi Faltings, 2012
    DUCT: An Upper Confidence Bound Approach to Distributed Constraint Optimization Problems
    AAAI 2012
  • Christos Dimitrakakis, 2012
    Sparse Reward Processes
    arXiv:1201.2555v1 [Abstract] [PDF]
  • Christos Dimitrakakis
    Robust Bayesian reinforcement learning through tight lower bounds
    [PDF] EWRL 2011.
  • Christos Dimitrakakis, Constantin Rothkopf
    Bayesian multitask inverse reinforcement learning
    [PDF] EWRL 2011.
  • Constantin Rothkopf, Christos Dimitrakakis
    Preference elicitation and inverse reinforcement learning
    arXiv 1104.5687 [Abstract] [PDF] ECML 2011.
  • Christos Dimitrakakis, 2010
    Context model inference for large or partially observable MDPs
    ICML Workshop on reinforcement learning and search in very large spaces. [PDF] [Extended PDF]
  • C. Dimitrakakis, 2009. Efficient Methods for Near-Optimal Sequential Decision Making Under Uncertainty. In Robert Babuska and Frans Groen, editors, Interactive Collaborative Information Systems, SCI 81, pages 125-153, Springer, 2010.
  • Christos Dimitrakakis, 2009
    Complexity of stochastic branch and bound methods for belief tree search in Bayesian reinforcement learning.
    International conference on Agents and Artificial intelligence (ICAART 2010), pages 259-254.
    [PDF] [Abstract] [PDF (short version)]
  • Christos Dimitrakakis and Michail Lagoudakis, 2008
    Rollout Sampling Approximate Policy Iteration
    Machine Learning 72(3), [Abstract] [PDF] [BibTeX]
  • Christos Dimitrakakis 2008
    Tree Exploration for Bayes RL Exploration
    International Conference on Computational Intelligence for Modelling, Control and Automation, CIMCA08, Vienna, Austria. [arXiv]
  • Christos Dimitrakakis and Michail Lagoudakis, 2008
    Algorithms and Bounds for Sampling-based Approximate Policy Iteration, [Abstract] [PDF] [BibTeX]
    EWRL'08, Lilles, France
  • PhD thesis: Ensembles for Sequence Learning [Gzipped PostScript] [PDF] EPFL, Lausanne, Switzerland, 2006.
  • Christos Dimitrakakis, 2006
    Nearly optimal exploration-exploitation decision thresholds [PDF]
    ICANN 2006
  • Christos Dimitrakakis, Samy Bengio, 2005
    Online Policy Adaptation for Ensemble Classifiers[Gzipped PostScript][PDF]
    Neurocomputing 64 (2005) 211-221.
    Note that this is an extended version of the conference paper with the same title.
  • Christos Dimitrakakis, Samy Bengio, 2005
    Gradient-based Estimates of Return Distributions
    PASCAL workshop on Principled Methods of Trading Exploration and Exploitation. [PDF]

Neuroscience

Mainly statistical analysis for some problems in neuroscience.
  • Pengsheng Zhang, C. Dimitrakakis and J. Triesch, 2012
    Network Self-organization Explains the Statistics and Dynamics of Synaptic Connection Strengths in Cortex
    PLoS Computational Biology
  • L. Lonini, C. Dimitrakakis, C. A. Rothkopf, J. Triesch
    Generalization and interference in human motor control
    In Computational and Robotic Models of the Hierarchical Organization of Behavior, Springer
  • Pengsheng Zhang, Christos Dimitrakakis and Jochen Triesch, 2011
    Network Self-organization Explains the Distribution of Synaptic Efficacies in Neocortex
    Proceedings of DevLeaNN: A Workshop on Development and Learning in Artificial Neural Networks, p. 8-9. [PDF]

Speech recognition

The focus here is on ensemble models for speech recognition, done during my PhD thesis.
  • C. Dimitrakakis, S. Bengio, 2011
    Phoneme and sentence-level ensembles for speech recognition
    EURASIP Journal on Audio, Speech and Music Processing, Article ID 426792, 17 pages, 2011. [Abstract] [PDF]
  • PhD thesis: Ensembles for Sequence Learning [Gzipped PostScript] [PDF] EPFL, Lausanne, Switzerland, 2006.
  • Christos Dimitrakakis, Samy Bengio, 2005
    Boosting Word Error Rates[PDF]
    ICASSP 2005
  • Christos Dimitrakakis, Samy Bengio, 2004
    Boosting HMMs with an application to Speech Recognition[PDF]
    ICASSP 2004

Statistical inference

Statistical modelling and analysis.
  • Christos Dimitrakakis, Nikolaos Tziortziotis,2013
    ABC Reinforcement Learning
    arXiv:1303.2506, ICML 2013 [Abstract] [PDF]
  • C. Dimitrakakis, 2011
    Context models on sequences of covers [Talk] [PDF] [Abstract] arXiv:1005.2263.
  • Christos Dimitrakakis, 2010
    Bayesian variable order Markov models.
    13th International Conference on Artificial Intelligence and Statistics (AISTATS 2010). [PDF] [Abstract][BibTeX]
    See also technical report: TR-UVA-09-04.
  • Christos Dimitrakakis and Aikaterini Mitrokotsa, 2009
    Statistical decision making for authentication and intrusion detection
    Proceedings of the 8th international conference on machine learning and applications (ICMLA 2009), pp. 409-414. [PDF]. [Abstract].