Cem Tekin

Associate Professor, Department of Electrical and Electronics Engineering

Head of Cognitive Systems, Bandits, and Optimization Research Group (CYBORG)

Bilkent University

cemtekin(at)ee.bilkent.edu.tr

+90-312-290-2584

Visit our research group's webpage here for up-to-date information.

Cem is an Associate Professor in the Department of Electrical and Electronics Engineering and Head of Cognitive Systems, Bandits and Optimization Research Group (CYBORG) at Bilkent University.

He received his PhD degree in Electrical Engineering: Systems from the University of Michigan, Ann Arbor, in 2013 (advisor: Mingyan Liu). He also received his MS degree in Mathematics and MSE degree in Electrical Engineering: Systems, from the University of Michigan in 2011 and 2010, respectively. Prior to attending the University of Michigan, He received his BS in Electrical and Electronics Engineering (valedictorian) from METU in 2008. From February 2013 to January 2015 he was a postdoctoral scholar in Electrical Engineering Department, UCLA (advisor: Mihaela van der Schaar). He received the Fred W. Ellersick award for the best paper in MILCOM 2009, the Science Academy Association of Turkey Distinguished Young Scientist (BAGEP) Award in 2019, Parlar Foundation Research Incentive Award in 2019, and IEEE Turkey Chapter Research Incentive Award in 2020. He is a Senior Member of IEEE.

Cem has authored or coauthored over 60 research papers, 5 book chapters and a research monograph. He has served as a reviewer for numerous journals including IEEE Transactions on Information Theory, IEEE Transactions on Automatic Control, IEEE/ACM Transactions on Networking, IEEE Transactions on Signal Processing, IEEE Transactions on Image Processing, IEEE Transactions on Mobile Computing, IEEE Transactions on Wireless Communications, IEEE JSTSP and IEEE JSAC. He has served as a reviewer for NeurIPS-22, ICML-22, ICLR-22, AISTATS-22, NeurIPS-21, ICML-21, AISTATS-21, NeurIPS-20, ICML-20 and TPC member for AAAI-21, AAAI-18, ACM Mobihoc-17, AAAI-17, AAAI-16, ISM-16, ECAI-16, MLSP-15 and GlobalSIP-15.

Cem's research interests include bandit problems, reinforcement learning, data science for personalized medicine, multi-agent systems, stream mining, influence maximization and cognitive communications.

Joint work with C. Ararat (Bilkent University), titled "Vector Optimization with Stochastic Bandit Feedback" has been accepted at AISTATS 2023.

Our work (by Ilker Demirel, Ahmet Alparslan Celik and Cem Tekin), titled "ESCADA: Efficient Safety and Context Aware Dose Allocation for Precision Medicine" has been accepted at NeurIPS 2022.

Our work (by Andi Nika, Sepehr Elahi and Cem Tekin), titled "Online context-aware task assignment in mobile crowdsourcing via adaptive discretization" will appear in IEEE Transactions on Network Science and Engineering, 2022.

Prof. Cem Tekin recently gave a tutorial on Multi-Armed Bandits in Healthcare in the summer school organized by The Cambridge Center for AI in Medicine (CCAIM), 2022.

Joint work with M. Abroshan (Alan Turing Institute), K. H. Yip (UCL), M. van der Schaar (University of Cambridge) titled "Conservative policy construction using variational autoencoders for logged data with missing values" appeared in IEEE Transactions on Neural Networks and Learning Systems, 2022.

Research Topics

Combinatorial Optimization in Unknown Environments

How should an advertiser promote its products in a social network to reach to a large set of users with a limited budget? How should a search engine suggest a ranked list of items to its users to maximize the click-through rate? How should a base station allocate its users to channels to maximize the system throughput? How should a mobile crowdsourcing platform dynamically assign available tasks to its workers to maximize the performance? How can we identify the most reliable paths from source to destination under probabilistic link failures? All of these problems require optimizing decisions among a vast set of alternatives. When the probabilistic description of the environment is fully specified, these problems-and many others-are solved using computationally efficient exact or approximation algorithms. In this paper, we focus on a much difficult and more realistic problem: How should we learn the optimal decisions in these complex problems via repeated interaction with the environment when the probabilistic description of the environment is unknown or only partially known?

It is natural to assume that the environment is unknown in many real-world applications. For instance, the advertiser may not know with what probability user i will influence its neighbor j in the social network or the search engine may not know with what probability user i will click the item shown on position j beforehand. Moreover, decisions are need to made sequentially over time. For instance, the recommender system should show a new list of items to each arriving user and the base station should reallocate network resources when the channel conditions change or the users leave/enter the system. Obviously, future decisions of the learner must be guided based on what it has observed thus far, i.e., the trajectory of actions, observations and rewards generated by the learner's past decisions. Importantly, both the cumulative reward of the learner and what it has learned so far also depend on this trajectory. Therefore, the learner needs to balance how much it earns (by exploiting the actions it believes to be the best) and how much it learns (by exploring actions it does not know much about) in order to maximize its long-term performance. We solve the formidable task of combinatorial optimization in unknown environments by modeling it as a combinatorial multi-armed bandit.

Related Publications

A. Huyuk and C. Tekin " Thompson sampling for combinatorial network optimization in unknown environments".
A. Huyuk and C. Tekin, " Analysis of Thompson sampling for combinatorial multi-armed bandit with probabilistically triggered arms", in Proc. 22nd International Conference on Artificial Intelligence and Statistics (AISTATS), April 2019.
O. Saritac, A. Karakurt and C. Tekin " Online contextual influence maximization with costly observations", IEEE Transactions on Signal and Information Processing over Networks, 5(2): 273-289, June 2019.
A. O. Saritac, C. Tekin, "Combinatorial multi-armed bandit problem with probabilistically triggered arms: A case with bounded regret", in Proc. IEEE GlobalSIP, November 2017.
A. O. Saritac, A. Karakurt and C. Tekin, "Online contextual influence maximization in social networks", in Proc. 54th Allerton Conference, September 2016, Monticello, Illinois.

Multi-objective Multi-armed Bandit

The multi-armed bandit (MAB) is extensively used to model sequential decision-making problems with uncertain rewards. While many real-world applications ranging from cognitive radio networks to recommender systems to medical diagnosis require intelligent decision-making mechanisms that learn from the past, majority of these applications involve side-observations that can guide the decision-making process, which does not fit into the classical MAB model. This issue is resolved by proposing new MAB models, called contextual bandits, that learn how to act optimally based on side-observations. On the other hand, the aforementioned real-world applications also involve multiple and possibly conflicting objectives. For instance, these objectives include throughput and reliability in a cognitive radio network, semantic match and job-seeking intent in a talent recommender system, and sensitivity and specificity in medical diagnosis. Motivated by the applications and the challenges above, in this research we develop new MAB models that address the learning challenges that arise from the presence of multiple and possibly conflicting objectives.

Related Publications

C. Tekin and E. Turgay " Multi-objective contextual multi-armed bandit problem with a dominant objective", IEEE Transactions on Signal Processing, 66(14): 3799-3813, July, 2018.
E. Turgay, D. Oner and C. Tekin, " Multi-objective contextual bandit problem with similarity information", in Proc. 21st International Conference on Artificial Intelligence and Statistics (AISTATS), April 2018.
C. Tekin and E. Turgay, "Multi-objective contextual bandits with a dominant objective", in Proc. IEEE MLSP, September 2017.

Big Data Stream Mining

Huge amounts of data streams are now being produced by more and more sources and in increasingly diverse formats: sensor readings, physiological measurements, GPS events, network traffic information, documents, emails, transactions, tweets, audio files, videos etc. These streams are then mined in real-time to assist numerous applications: patient monitoring, online recommendation systems, social networks, targeted advertisement, network security etc. Hence, online data mining systems have emerged that enable such applications to analyze, extract actionable intelligence and make decisions in real-time, based on the correlated, high-dimensional and dynamic data captured by multiple heterogeneous data sources. To mine the data streams, the following questions need to be answered continuously: Which classifiers should process the data? How many and in which order, configuration or topology? What are the costs (e.g. delay) and benefits (accuracy of predictions) in invoking a specific classifier?

In this research, we formalize the real-time mining of Big Data streams as an online learning and sequential decision problem, where classifiers and their configurations are chosen online to make predictions based on the gathered data, and subsequently focus on multi-armed bandits (MABs) as an important class of solutions for solving this problem. In the considered systems, data from the sources are processed by an online learner which determines on-the-fly how to classify the different data streams and make decisions based on the predictions. For this, the learner uses one of its available classifiers (or classifier chains) to make a prediction. Since the prediction accuracy of the classifiers changes dynamically, over time, based on the characteristics of the collected data streams, this needs to be learned online. Hence, the learner needs to continuously learn while at the same time make accurate predictions and decisions, i.e. the learning and decision making are coupled and concurrent.

The objective of this research is to highlight the key challenges associated with implementing real-time Big Data stream mining systems, show how to formalize the online learning and decision making in stream mining systems as MAB problems, present different MAB-based methods to solve these problems, and discuss their performance.

Related Publications

C. Tekin J. Yoon and M. van der Schaar " Adaptive ensemble learning with confidence bounds", IEEE Transactions on Signal Processing, 65(4): 888-903, February 2017.
C. Tekin and M. van der Schaar, " Distributed online learning via cooperative contextual bandits", IEEE Transactions on Signal Processing, 63(14):3740–3754, July 2015.
J. Xu, C. Tekin, S. Zhang and M. van der Schaar, " Distributed multi-agent online learning based on global feedback", IEEE Transactions on Signal Processing, 63(9):2225–2238, May 2015.
O. Atan, Y. Andreopoulos, C. Tekin, and M. van der Schaar, " Bandit framework for systematic learning in wireless video-based face recognition", IEEE Journal of Selected Topics in Signal Processing (JSTSP) , 9(1):180–194, February 2015.

Discovering, Learning and Exploiting Relevance in Big Data

Relevance relations arise naturally in many practical applications. For example, when treating patients with a particular disease, many contexts may be available: the patients' age, weight, blood tests, imaging, medical history etc. But often only a few of these contexts are relevant in choosing/not choosing a particular treatment or medication. For instance, surgery may be strongly contra-indicated in patients with clotting problems; drug therapies that require close monitoring may be strongly contra-indicated in patients who do not have committed care-givers, etc. Similarly, in recommender systems, a product recommendation may sometimes depend on many characteristics of the user such as gender, occupation, history of past purchases etc., but will often depend only (or most strongly) on a few characteristics such as location and home-ownership.

In this research we develop and analyze algorithms that allow efficient learning and decision-making while avoiding the curse of dimensionality. We formalize the information available to the learner/decision-maker at a particular time as a context vector which the learner should consider when taking actions. In general the context vector is very high dimensional, but in many settings, the most relevant information is embedded into only a few relevant dimensions. If these relevant dimensions were known in advance, the problem would be simple -- but they are not. Moreover, the relevant dimensions may be different for different actions. Our algorithms learn the relevant dimensions for each action, and make decisions based on what they have learned. Formally, we build on the structure of a contextual multi-armed bandit by adding and exploiting a relevance relation. We prove a general regret bound for our algorithms whose time order depends only on the maximum number of relevant dimensions among all the actions. In the absence of a relevance relation, the best known contextual bandit algorithms achieve regret that scales with the full dimension of the context vector. Our algorithms alternate between exploring and exploiting and do not require observing outcomes during exploitation (so allows for active learning). Moreover, during exploitation, suboptimal actions are chosen with arbitrarily low probability. The performance of our algorithms are tested on datasets arising from network security and online news article recommendations.

Related Publications

E. Turgay, C. Bulucu and C. Tekin " Exploiting relevance for online decision-making in high-dimensions".
C. Tekin and M. van der Schaar, " RELEAF: An algorithm for learning and exploiting relevance", IEEE Journal of Selected Topics in Signal Processing (JSTSP), 9(4):716–727, June 2015.
C. Tekin O. Atan and M. van der Schaar, " Discover the expert: Context-adaptive expert selection for medical diagnosis", IEEE Transactions on Emerging Topics in Computing, 3(2):220–234, June 2015.
C. Tekin and M. van der Schaar, " Discovering, learning and exploiting relevance", in 28th Annual Conference on Neural Information Processing Systems (NIPS), December 2014, Montreal, Canada.

Online Learning for Social Networks and Recommender Systems

One of the most powerful benefits of a social network is the ability for cooperation and coordination on a large scale over a wide range of different agents. By forming a network, agents are able to share information and opportunities in a mutually beneficial fashion. For example, companies can collaborate to sell products, charities can work together to raise money, and a group of workers can help each other search for jobs. Through such cooperation, agents are able to attain much greater rewards than would be possible individually. But sustaining efficient cooperation can also prove extremely challenging. First, agents operate with only incomplete information, and must learn the environment parameters slowly over time. Second, agents are decentralized and thus uncertain about their neighbor's information and preferences.Finally, agents are selfish in the sense that, they don't want to reveal their inventory, observations and actions to other agents, unless they benefit.

In this research we produce a class of algorithms that effectively addresses all of these issues: at once allowing decentralized agents to take near-optimal actions in the face of incomplete information, while still incentivizing them to fully cooperate within the network.

The framework we consider is very broad and applicable to a wide range of social networking situations. We analyze a group of agents that are connected together via a fixed network, each of whom experiences inflows of users to its page. Each time a user arrives, an agent chooses from among a set of items to offer to that user, and the user will either reject or accept each item. These items can represent a variety of things, from a good that the agent is trying to sell, to a cause that the agent is trying to promote, to a photo that the agent is trying to circulate. In each application, the action of accepting or rejecting by the user will likewise have a distinct meaning. When choosing among the items to offer, the agent is uncertain about the user's acceptance probability of each item, but the agent is able to observe specific background information about the user, such as the user's gender, location, age, etc. Users with different backgrounds will have different probabilities of accepting each item, and so the agent must learn this probability over time by making different offers.

We produce a class of algorithms that allows agents to take near-optimal actions even with decentralized learning. We prove specific bounds for the regret, which is the difference between the total expected reward of an agent using a learning algorithm and the total expected reward of the optimal policy for the agent, which is computed given perfect knowledge about acceptance probabilities for each context. We show that the regret is sublinear in time in all cases. We further show that our algorithms can operate regardless of the specific network topology, including the degree of connectivity, degree distribution, clustering coefficient, etc., although the performance is better if the network is more connected since each agent will have access to more items of other agents.



Related Publications

L. Song, C. Tekin and M. van der Schaar, " Online learning in large-scale contextual recommender systems", IEEE Transactions on Services Computing, 9(3): 433-445, 2016.
C. Tekin, S. Zhang, and M. van der Schaar, " Distributed online learning in social recommender systems", IEEE Journal of Special Topics in Signal Processing (JSTSP), 8(4): 638-652, August 2014.

Students

MSc Students

İlker Demirel [Moving to MIT EECS for PhD]

Ilker is a second-year MS student in the the Department of Electrical and Electronics Engineering at Bilkent University. He is interested in statistical machine learning and sequential decision-making under uncertainty with applications to healthcare, distributed learning, and optimization.


Kerem Bozgan [Moving to Virginia Tech for PhD]

Kerem's research interests include multi-objective and robust Bayesian optimization.


Undergraduate Students

Baran Atalar [Moving to CMU ECE for PhD]

Baran is a senior undergraduate student in the Department of Electrical and Electronics Engineering at Bilkent University. He is interested in reinforcement learning, multi-armed bandit problems, online learning and machine learning in general. He is currently working on contextual combinatorial volatile bandits with thresholds, and mmWave beam selection problem for vehicular communications.


Batu Arda Düzgün [Moving to ETH for Master in Electrical Engineering and Information Technology]

Batu is a senior undergraduate student in the Department of Electrical and Electronics Engineering at Bilkent University. He works on bandit algorithms and PAC best arm identification.


Sepehr Elahi [Moving to EPFL for PhD]

Sepehr is a senior undergraduate double major student studying Electrical & Electronics Engineering and Mathematics at Bilkent University, Ankara. At Bilkent CYBORG, he has authored papers on online decision making, contextual bandits, Gaussian processes, multi-objective optimization, and their applications to crowdsourcing/sensing, content recommendation, and item recommendation. Besides the previously mentioned, he is also interested in federated learning, Bayesian learning, and in general statistical machine learning.


Sevda Ögüt [Moving to EPFL for PhD]

Sevda is a senior undergraduate student in the Department of Electrical and Electronics Engineering at Bilkent University. Her research interests include multi-armed bandits, especially in contextual, combinatorial and volatile settings, Gaussian processes, thresholding and applications of these to content caching and mmWave communications.


Mehmet Ufuk Özdemir [Moving to UCLA ECE for PhD]

Ufuk is a senior undergraduate student in the Department of Electrical and Electronics Engineering at Bilkent University. His research interests include safe bandit algorithms and vector optimization with stochastic bandit feedback.


Alumni

Ahmet Alparslan Çelik [Former MS student]

Andi Nika [Former MS student, now PhD student at Max Planck Institute for Software Systems]

Alireza Javanmardi [Former MS student, now PhD student at University of Tubingen]

Muhammad Anjum Qureshi [Former PhD student, now General Manager at Ministry of Science and Technology, Pakistan]

Alihan Hüyük [Former undergraduate student, now PhD student at University of Cambridge]

Doruk Öner [Former undergraduate student, now PhD student at EPFL]

Anıl Ömer Sarıtaç [Former MS student, now PhD student at London Business School]

Nima Akbarzadeh [Former MS student, now PhD student at McGill University]

Cem Bulucu [Former MS student]

Ümitcan Şahin [Former MS student, now Research Engineer at ASELSAN Research Center]

Eralp Turğay [Former MS student, now Research Engineer at ASELSAN Research Center]

Kubilay Ekşioğlu [Former MS student]

Publications

Preprints

[I4] S. Elahi, B. Atalar, S. Ogut, C. Tekin, "Contextual combinatorial volatile bandits with satisfying via Gaussian processes".
[I3] I. Demirel, Y. Yildirim, C. Tekin, "Federated multi-armed bandits under Byzantine attacks".
[I2] A. Nika, S. Elahi and C. Tekin, "Contextual combinatorial volatile bandits via Gaussian processes".
[I1] A. Nika, K. Bozgan, S. Elahi, C. Ararat and C. Tekin, "Pareto active learning with Gaussian processes and adaptive discretization".

Monograph

[M1] C. Tekin and M. Liu, "Online learning methods in networking", Foundations and Trends in Networking, January 2015.

Book chapters

[B5] R. Ducasse, C. Tekin, and M. van der Schaar, "Finding it now: Networked classifiers in real-time stream mining systems", in Handbook of Signal Processing Systems, Springer, 2018.
[B4] C. Tekin, S. Zhang, J. Xu, and M. van der Schaar, "Multi-agent systems: Learning, strategic behavior, cooperation, and network formation", in Cooperative and Graph Signal Processing, Elsevier, 2018.
[B3] C. Tekin, and M. van der Schaar, "Actionable intelligence and online learning for semantic computing", in Encyclopedia with Semantic Computing and Robotic Intelligence, 01, 1630011, 2017.
[B2] S. S. Bhattacharyya, M. van der Schaar, O. Atan, C. Tekin, and K. Sudusinghe, "Data-driven stream mining systems for computer vision", in Advances in Embedded Computer Vision, Springer, November 2014.
[B1] C. Tekin and M. Liu, "Performance and convergence of multiuser online learning and its application in dynamic spectrum sharing", in Mechanisms and Games for Dynamic Spectrum Allocation, Cambridge University Press, February 2014.

Journal papers

[J33] A. Nika, S. Elahi, C. Tekin, "Online context-aware task assignment in mobile crowdsourcing via adaptive discretization", IEEE Transactions on Network Science and Engineering, 2022.
[J32] M. Abroshan, K. H. Yip, C. Tekin, M. van der Schaar "Conservative policy construction using variational autoencoders for logged data with missing values", IEEE Transactions on Neural Networks and Learning Systems, 2022.
[J31] A. Javanmardi, M. A. Qureshi and C. Tekin "Decentralized dynamic rate and channel selection over a shared spectrum", IEEE Transactions on Communications, 69(6): 3787-3801, June 2021.
[J30] M. A. Qureshi, A. Nika and C. Tekin "Multi-user small base station association via contextual combinatorial volatile bandits", IEEE Transactions on Communications, 69(6): 3726-3740, June 2021.
[J29] A. Huyuk and C. Tekin "Multi-objective multi-armed bandit with lexicographically ordered and satisficing objectives", Machine Learning, 110: 1233-1266, May 2021.
[J28] E. Turgay, C. Bulucu and C. Tekin "Exploiting relevance for online decision-making in high-dimensions", IEEE Transactions on Signal Processing, 69: 1438-1451, December 2020.
[J27] C. Tekin, S. Elahi and M. van der Schaar, "Feedback adaptive learning for medical and educational application recommendation", to appear in IEEE Transactions on Services Computing, 2020.
[J26] A. Huyuk and C. Tekin "Thompson sampling for combinatorial network optimization in unknown environments", IEEE/ACM Transactions on Networking, 28(6): 2836-2849, December 2020.
[J25] M. A. Qureshi and C. Tekin "Rate and channel adaptation in cognitive radio networks under time-varying constraints", IEEE Communications Letters, 24(12): 2979-2983, December 2020.
[J24] M. A. Qureshi and C. Tekin "Fast learning for dynamic resource allocation in AI-enabled radio networks", IEEE Transactions on Cognitive Communications and Networking, 6(1): 95-110, March 2020.
[J23] O. Saritac, A. Karakurt and C. Tekin "Online contextual influence maximization with costly observations", IEEE Transactions on Signal and Information Processing over Networks, 5(2): 273-289, June 2019.
[J22] N. Akbarzadeh, C. Tekin and M. van der Schaar "Online learning in limit order book trade execution", IEEE Transactions on Signal Processing, 66(17): 4626-4641, September, 2018.
[J21] H. S. Lee, C. Tekin, M. van der Schaar and J. W. Lee "Adaptive contextual learning for unit commitment in microgrids with renewable energy sources", IEEE Journal of Selected Topics in Signal Processing, 12(4): 688-702, August, 2018.
[J20] C. Tekin and E. Turgay "Multi-objective contextual multi-armed bandit problem with a dominant objective", IEEE Transactions on Signal Processing, 66(14): 3799-3813, July, 2018.
[J19] S. Muller, C. Tekin, M. van der Schaar and A. Klein "Context-aware hierarchical online learning for performance maximization in mobile crowdsourcing", IEEE/ACM Transactions on Networking, 26(3): 1334-1347, June, 2018.
[J18] O. Atan, C. Tekin and M. van der Schaar "Global bandits", IEEE Transactions on Neural Networks and Learning Systems, 29(12): 5798-5811, December 2018.
[J17] R. B. Hellman, C. Tekin, M. van der Schaar and V. J. Santos "Functional contour-following via haptic perception and reinforcement learning", IEEE Transactions on Haptics, 11(1): 61-72, January-March, 2018.
[J16] C. Shen, R. Zhou, C. Tekin and M. van der Schaar "Generalized global bandit and its application in cellular coverage optimization", IEEE Journal of Selected Topics in Signal Processing, 12(1): 218-232, January, 2018.
[J15] C. Tekin J. Yoon and M. van der Schaar "Adaptive ensemble learning with confidence bounds", IEEE Transactions on Signal Processing, 65(4): 888-903, February 2017.
[J14] C. Shen, C. Tekin and M. van der Schaar "A non-stochastic learning approach to energy efficient mobility management", IEEE Journal on Selected Areas in Communications (JSAC), 34(12): 3854-3868, December 2016.
[J13] K. Kanoun, C. Tekin, D. Atienza and M. van der Schaar, "Big-data streaming applications scheduling based on staged multi-armed bandits", IEEE Transactions on Computers, 65(12): 3591-3605, December 2016.
[J12] S.D. Amuru, C. Tekin, M. van der Schaar and R. Buehrer, "Jamming bandits - A novel learning method for optimal jamming", IEEE Transactions on Wireless Communications, 15(4):2792-2808, April 2016.
[J11] C. Tekin and M. van der Schaar, "Distributed online learning via cooperative contextual bandits", IEEE Transactions on Signal Processing, 63(14):3740–3754, July 2015.
[J10] C. Tekin and M. van der Schaar, "Active learning in context-driven stream mining with an application to image mining", IEEE Transactions on Image Processing, 24(11): 3666-3679, June 2015.
[J9] C. Tekin and M. van der Schaar, "RELEAF: An algorithm for learning and exploiting relevance", IEEE Journal of Selected Topics in Signal Processing (JSTSP), 9(4):716–727, June 2015.
[J8] C. Tekin O. Atan and M. van der Schaar, "Discover the expert: Context-adaptive expert selection for medical diagnosis", IEEE Transactions on Emerging Topics in Computing, 3(2):220–234, June 2015.
[J7] J. Xu, C. Tekin, S. Zhang and M. van der Schaar, "Distributed multi-agent online learning based on global feedback", IEEE Transactions on Signal Processing, 63(9):2225–2238, May 2015.
[J6] C. Tekin and M. van der Schaar, "Contextual online learning for multimedia content aggregation", IEEE Transactions on Multimedia, 17(4):549-561, April 2015.
[J5] O. Atan, Y. Andreopoulos, C. Tekin, and M. van der Schaar, "Bandit framework for systematic learning in wireless video-based face recognition", IEEE Journal of Selected Topics in Signal Processing (JSTSP) , 9(1):180–194, February 2015.
[J4] L. Song, C. Tekin and M. van der Schaar, "Online learning in large-scale contextual recommender systems", IEEE Transactions on Services Computing, 9(3): 433-445, 2016.
[J3] C. Tekin, S. Zhang, and M. van der Schaar, "Distributed online learning in social recommender systems", IEEE Journal of Special Topics in Signal Processing (JSTSP), 8(4): 638-652, August 2014.
[J2] C. Tekin, M. Liu, R. Southwell, J. Huang, S. H. A. Ahmad, "Atomic congestion games on graphs and their applications in networking", IEEE/ACM Transactions on Networking (ToN), 20(5): 1541-1552, October 2012.
[J1] C. Tekin and M. Liu, "Online learning of rested and restless bandits", IEEE Transactions on Information Theory (IT), 58(8): 5558-5611, August 2012.

Papers in conference proceedings

[C35] C. Ararat and C. Tekin, "Vector optimization with stochastic bandit feedback", to appear in AISTATS 2023.
[C34] I. Demirel, A. A. Celik and C. Tekin "ESCADA: Efficient safety and context aware dose allocation for precision medicine", to appear in NeurIPS 2022.
[C33] A. Huyuk, D. Jarret, C. Tekin, M. van der Schaar, "Explaining by imitating: Understanding decisions by interpretable policy learning", in Proc. 9th International Conference on Learning Representations (ICLR), May 2021.
[C32] I. Demirel and C. Tekin, "Combinatorial Gaussian process bandits with probabilistically triggered arms", in Proc. 24th International Conference on Artificial Intelligence and Statistics (AISTATS), April 2021.
[C31] A. Nika, S. Elahi and C. Tekin, "Contextual combinatorial volatile multi-armed bandit with adaptive discretization", in Proc. 23rd International Conference on Artificial Intelligence and Statistics (AISTATS), August 2020.
[C30] M. A. Qureshi and C. Tekin, "Online Bayesian learning for rate selection in millimeter wave cognitive radio networks", Proc. 2020 IEEE Conference on Computer Communications (INFOCOM), July 2020.
[C29] X. Zhang, M. M. Khalili, C. Tekin and M. Liu, "Group retention when using machine learning in sequential decision making: The interplay between user dynamics and fairness", in Proc. 33rd Conference on Neural Information Processing Systems (NeurIPS), December 2019.
[C28] A. Huyuk and C. Tekin, "Analysis of Thompson sampling for combinatorial multi-armed bandit with probabilistically triggered arms", in Proc. 22nd International Conference on Artificial Intelligence and Statistics (AISTATS), April 2019.
[C27] E. Turgay, D. Oner and C. Tekin, "Multi-objective contextual bandit problem with similarity information", in Proc. 21st International Conference on Artificial Intelligence and Statistics (AISTATS), April 2018.
[C26] A. O. Saritac, C. Tekin, "Combinatorial multi-armed bandit problem with probabilistically triggered arms: A case with bounded regret", in Proc. IEEE GlobalSIP, November 2017.
[C25] N. Akbarzadeh, C. Tekin and M. van der Schaar, "Online learning in limit order book trade execution", in Proc. IEEE GlobalSIP, November 2017.
[C24] C. Tekin and E. Turgay, "Multi-objective contextual bandits with a dominant objective", in Proc. IEEE MLSP, September 2017.
[C23] H. Lee, C. Tekin, M. van der Schaar and J. Lee, "Contextual learning for unit commitment with renewable energy sources", in Proc. IEEE GlobalSIP , December 2016.
[C22] A. O. Saritac, A. Karakurt and C. Tekin, "Online contextual influence maximization in social networks", in Proc. 54th Allerton Conference, September 2016, Monticello, Illinois.
[C21] N. Akbarzadeh and C. Tekin, "Gambler's ruin bandit problem", in Proc. 54th Allerton Conference, September 2016, Monticello, Illinois.
[C20] C. Tekin, J. Yoon and M. van der Schaar, "Adaptive ensemble learning with confidence bounds for personalized diagnosis", in AAAI Workshop on Expanding the Boundaries of Health Informatics using AI (HIAI'16): Making Proactive, Personalized, and Participatory Medicine A Reality, February, 2016.
[C19] O. Atan, C. Tekin and M. van der Schaar, " Global Multi-armed Bandits with Hölder Continuity", in Proc. 18th International Conference on Artificial Intelligence and Statistics (AISTATS), April 2015.
[C18] C. Tekin, Jonas Braun and M. van der Schaar, "eTutor: Online learning for personalized education", in Proc. 40th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), April 2015.
[C17] Onur Atan, William Hsu, C. Tekin and M. van der Schaar, "A data-driven approach for matching clinical expertise to individual cases", in Proc. 40th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), April 2015.
[C16] S.D. Amuru, C. Tekin, M. van der Schaar and R. Buehrer, "A systematic learning method for optimal jamming", in Proc. IEEE International Conference on Communications (ICC), May 2015.
[C15] C. Tekin and M. van der Schaar, " Discovering, learning and exploiting relevance", in Proc. 28th Annual Conference on Neural Information Processing Systems (NIPS), December 2014, Montreal, Canada.
[C14] C. Tekin and M. van der Schaar, "An experts learning approach to mobile service offloading", in Proc. 52nd Allerton Conference, October 2014, Monticello, Illinois.
[C13] C. Tekin , L. Canzian, and M. van der Schaar, "Context adaptive Big Data stream mining", in Proc. 52nd Allerton Conference, October 2014, Monticello, Illinois.
[C12] O. Atan, Y. Andreopoulos, C. Tekin, and M. van der Schaar, "Bandit framework for systematic learning in wireless video-based face recognition", in Proc. 39th IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2014.
[C11] L. Song, C. Tekin, and M. van der Schaar, " Clustering based online learning in recommender systems: a bandit approach", in Proc. 39th IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2014.
[C10] J. Xu, C. Tekin and M. van der Schaar, " Learning optimal classifier chains for real-time big data mining", in Proc. 51st Allerton Conference, October 2013, Monticello, Illinois.
[C9] C. Tekin and M. van der Schaar, " Distributed online Big Data classification using context information", in Proc. 51st Allerton Conference, October 2013, Monticello, Illinois.
[C8] C. Tekin and M. Liu, " Online learning in decentralized multi-user spectrum access with synchronized explorations", in Proc. IEEE Military Communications Conference (MILCOM), October 2012.
[C7] C. Tekin and M. Liu, " Approximately optimal adaptive learning in opportunistic spectrum access", in Proc. 31st IEEE International Conference on Computer Communications (INFOCOM), March 2012, Orlando, Florida.
[C6] C. Tekin and M. Liu, " Adaptive learning of uncontrolled restless bandits with logarithmic regret", in Proc. 49th Allerton Conference, September 2011, Monticello, Illinois.
[C5] C. Tekin and M. Liu, " Performance and convergence of multi-user online learning", in Proc. 2nd International ICST Conference on Game Theory for Networks (GAMENETS), April 2011, Shanghai, China.
[C4] C. Tekin and M. Liu, " Online learning in opportunistic spectrum access: a restless bandit approach", in Proc. 30th IEEE International Conference on Computer Communications (INFOCOM), April 2011, Shanghai, China.
[C3] C. Tekin and M. Liu, " Online algorithms for the multi-armed bandit problem with Markovian rewards", in Proc. 48th Allerton Conference, September 2010, Monticello, Illinois.
[C2] S. Hong, E. Like, Z. Wu and C. Tekin, " Multi-user signal classification via spectral correlation", in Proc. 7th IEEE Consumer Communications and Networking Conference (CCNC) , January 2010, Las Vegas, Nevada.
[C1] C. Tekin, S. Hong and W. Stark " Enhancing cognitive radio dynamic spectrum sensing through adaptive learning", in Proc. IEEE Military Communications Conference (MILCOM), October 2009, Boston, Massachusetts.

Permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the copyright holder.

Funding

  • TÜBİTAK 1001 Program: Active Learning in Large Scale Vector Optimization Problems.
  • Distinguished Young Scientist (BAGEP) Award, Science Academy Association of Turkey.
  • TÜBİTAK 3501 Career Development Program: Optimal Decentralized Cross-Layer Learning in Cognitive Radio Networks.
  • TÜBİTAK 1003 Program: Improving Quality and Efficiency of Clinical Processes by Learning from Big Health Data.

Projects

Active Learning in Large Scale Vector Optimization Problems

Program: TÜBİTAK 1001 Program
Project Summary:

Many complex scientific problems, including vaccine development for pandemic diseases and deep neural architecture search, require identifying optimal designs that satisfy certain performance criteria from a design set that includes many alternative designs via sequential experimentation. This problem has been mainly investigated in the literature when there is a single performance objective to be optimized. It has been studied under the names active learning, Bayesian experimental design, Bayesian optimization, etc. Due to their complex nature, many real-world applications of active learning require the optimization of multiple objectives. However, finding a design that simultaneously maximizes each objective under this setting may not be possible. Therefore, for the experts to determine which designs to focus their research on, it is necessary to identify Pareto optimal designs that are not dominated by other designs.

It is very difficult for domain experts to design optimal sequential experiments for problems that involve large design sets, multiple performance objectives intricately tied with each other, and restrictions on observations of design performances such as partial and noisy observations. The goal of this project is to design active learning algorithms that are mathematically rigorously founded, provably efficient and scalable, for complex learning problems in which identification of designs that optimize conflicting objectives by sequential experimentation via using the minimum amount of resources is required, given that the performances of the designs under different objectives were unknown initially. The algorithms developed in this project will enable experts in various disciplines, ranging from biology to data science, to quickly identify optimal designs in a small number of evaluations and use the minimum amount of resources. The results of this project will pave the path for having efficient and fast solutions for many scientific problems in different fields that require many trials and a lot of resources.

Related Publications:
C. Ararat and C. Tekin "Vector optimization with stochastic bandit feedback".

Optimal Decentralized Cross-Layer Learning in Cognitive Radio Networks

Program: TÜBİTAK 3501 Career Development Program
Project Summary:

Cognitive Radio Network (CRN) is defined as a wireless communication system that is aware of the changes in its surroundings and that adapts its transmission parameters according to these changes. The goal of a CRN is to use the radio spectrum as efficiently as possible while operating resiliently under a diverse set of environmental conditions. The goal of this project is to design new learning and adaptation based methods for CRNs. The learning methods that will be developed in this project will enable each CRN user and each CRN layer to learn the transmission parameters that will maximize the system’s performance over time with limited inter-layer and inter-user communication.

Related Publications:
C. Tekin and E. Turgay " Multi-objective contextual multi-armed bandit problem with a dominant objective", IEEE Transactions on Signal Processing, 66(14): 3799-3813, July, 2018.
S. Muller, C. Tekin, M. van der Schaar and A. Klein " Context-aware hierarchical online learning for performance maximization in mobile crowdsourcing", IEEE/ACM Transactions on Networking, 26(3): 1334-1347, June, 2018.
E. Turgay, D. Oner and C. Tekin "Multi-objective contextual bandit problem with similarity information", in Proc. 21st International Conference on Artificial Intelligence and Statistics (AISTATS), April 2018.
M. A. Qureshi and C. Tekin "Online cross-layer learning in heterogeneous cognitive radio networks without CSI", in Proc. SIU, 2018.
C. Shen, R. Zhou, C. Tekin and M. van der Schaar " Generalized global bandit and its application in cellular coverage optimization", IEEE Journal of Selected Topics in Signal Processing, 12(1): 218-232, January, 2018.
C. Tekin and E. Turgay "Multi-objective contextual bandits with a dominant objective", in Proc. IEEE MLSP, September 2017.

Improving Quality and Efficiency of Clinical Processes by Learning from Big Health Data

Program: TÜBİTAK 1003 Program
Project webpage is here.
Project Summary: Big Health Data (BHD) has a great potential to support clinicians in diagnosis, treatment administration and patient monitoring. However, BHD includes a vast amount of documents, test results and medical images related to the past treatments of the patient. Leveraging this information requires development of sophisticated statistical learning techniques for diagnosis and treatment recommendation and risk management. The goal of this project is to design trustable machine learning algorithms that will learn to make optimal decisions using the BHD, and to integrate them into new decision support systems that will be used by clinicians and patients.

Courses

  • EEE 485/585 Statistical Learning and Data Analytics, Fall 2016-2017-2018-2019, Spring 2017-2018-2021, Bilkent University
  • EEE 443/543 Neural Networks, Fall 2015, Bilkent University
  • CS 421 Computer Networks, Fall 2019, Spring 2015-2016-2017, Bilkent University
  • EEE 102 Introduction to Digital Circuit Design, Fall 2016-2017-2018-2020, Spring 2016-2017-2018-2020 Bilkent University
  • EEE 591-592 Seminar Series, Fall 2015-2016-2017-2018, Spring 2016-2017-2018, Bilkent University

Talks

  • "Exploration and exploitation in complex domains: Learning with multiple objectives and contexts", University of Cambridge, July 2019.
  • "Exploration and exploitation in complex domains: Learning with multiple objectives and combinatorial actions", Sabanci University, November 2018.
  • "Online learning in complex domains", Ohio State University, July 2018.
  • "Online learning in complex domains", University of Michigan, July 2018.
  • "Combinatorial and multi-objective online learning", McGill University, November 2017.
  • "Online learning in limit order book trade execution", Oxford-Man Institute of Quantitative Finance, August 2017.
  • "Combinatorial and multi-objective online learning", The Alan Turing Institute, August 2017.
  • "Online learning in big data", ODTU, April 2017.
  • "Online learning in big data", Sabanci University, November 2016.
  • "Online learning in big data", Koc University, April 2016.
  • "Exploiting relevance and structure in multi-armed bandit problems", UCLA Center for Engineering Economics, Learning, and Networks, August 2015.
  • "Decentralized online big data classification – a cooperative contextual bandit framework", Southern California Symposium on Network Economics and Game Theory, November 2013.
  • "Online learning in single-agent and multi-agent bandit problems", Information Theory and Applications (ITA) Workshop, February 2013.
  • "Performance and convergence of multi-user online learning", INFORMS Midwestern Conference, August 2011.

Service

Technical Program Committe Memberships

  • (2021) Thirty fifth AAAI Conference on Artificial Intelligence (AAAI-21)
  • (2018) Thirty second AAAI Conference on Artificial Intelligence (AAAI-18)
  • (2017) Eighteenth International Symposium on Mobile Ad Hoc Networking and Computing (ACM MobiHoc 2017)
  • (2017) Thirty first AAAI Conference on Artificial Intelligence (AAAI-17)
  • (2016) IEEE International Symposium of Multimedia (ISM-16)
  • (2016) The Biennial European Conference on Artificial Intelligence (ECAI-16)
  • (2016) Thirtieth AAAI Conference on Artificial Intelligence (AAAI-16)
  • (2014) IEEE JSAC – Special Issue on Recent Advances in Heterogeneous Cellular Networks
  • (2015) IEEE International Workshop on Machine Learning for Signal Processing (MLSP-15)
  • (2015) IEEE Global Conference on Signal and Information Processing (GlobalSIP-15)

Reviewer

  • 10th International Conference on Learning Representations (ICLR-22)
  • 25th International Conference on Artificial Intelligence and Statistics (AISTATS-22)
  • 38th International Conference on Machine Learning (ICML-21)
  • 35th Conference on Neural Information Processing Systems (NeurIPS-21)
  • 24th International Conference on Artificial Intelligence and Statistics (AISTATS-21)
  • 34th Conference on Neural Information Processing Systems (NeurIPS-20)
  • 37th International Conference on Machine Learning (ICML-20)
  • IEEE Transactions on Information Theory
  • IEEE Transactions on Automatic Control
  • IEEE/ACM Transactions on Networking
  • IEEE Transactions on Signal Processing
  • IEEE Journal of Selected Topics in Signal Processing (JSTSP)
  • IEEE Transactions on Image Processing
  • IEEE Transactions on Neural Networks and Learning Systems
  • IEEE Transactions on Mobile Computing
  • IEEE Transactions on Wireless Communications
  • IEEE Journal on Selected Areas in Communications (JSAC)
  • IEEE Communications Letters
  • IEEE Transactions on Services Computing
  • IEEE Transactions on Control of Networked Systems
  • Journal of Theoretical Computer Science
  • Pattern Recognition
  • Performance Evaluation
  • Digital Signal Processing
  • Computer Communications