• Trust me! I am a robot: an affective computational account of scaffolding in robot-robot interaction 

  Kırtay, M.; Öztop, Erhan; Asada, M.; Hafner, V. V: (IEEE, 2021-08-08)

  Forming trust in a biological or artificial interaction partner that provides reliable strategies and employing the learned strategies to scaffold another agent are critical problems that are often addressed separately in ...

• Misclassification risk and uncertainty quantification in deep classifiers 

  Şensoy, Murat; Saleki, Maryam; Julier, S.; Aydoğan, Reyhan; Reid, J. (IEEE, 2021)

  In this paper, we propose risk-calibrated evidential deep classifiers to reduce the costs associated with classification errors. We use two main approaches. The first is to develop methods to quantify the uncertainty of a ...

• Learning system dynamics via deep recurrent and conditional neural systems 

  Pekmezci, Mehmet; Uğur, E.; Öztop, Erhan (IEEE, 2021)

  Although there are various mathematical methods for modeling system dynamics, more general solutions can be achieved using deep learning based on data. Alternative deep learning methods are presented in parallel with the ...

• Road to salvation: Streaming clients and content delivery networks working together 

  Beğen, Ali Cengiz; Bentaleb, A.; Silhavy, D.; Pham, S.; Zimmermann, R.; Law, W. (IEEE, 2021-11-01)

  Streaming media has truly become one of the most popular applications on the Internet. Viewers are spoiled for choice, and content providers compete to provide the best viewer experience. Traditionally, it has been challenging ...

• Hierarchical mixtures of generators for adversarial learning 

  Ahmetoğlu, A.; Alpaydın, Ahmet İbrahim Ethem

  Generative adversarial networks (GANs) are deep neural networks that allow us to sample from an arbitrary probability distribution without explicitly estimating the distribution. There is a generator that takes a latent ...

• Data model extension impact analysis 

  Ersoy, E.; Sözer, Hasan (IEEE, 2021)

  Relational database schemas are subject to change. For instance, columns of a table can be modified, deleted or extended. These changes have an impact on the source code that utilizes the corresponding table. They also ...

• Dynamic filtering and prioritization of static code analysis alerts 

  Yüksel, U.; Sözer, Hasan (IEEE, 2021)

  We propose an approach for filtering and prioritizing static code analysis alerts while these alerts are being reviewed by the developer. We construct a Prolog knowledge base that captures the data flow information in the ...

• Time-correlated sparsification for communication-efficient federated learning 

  Özfatura, E.; Özfatura, Ahmet Kerem; Gündüz, D. (IEEE, 2021)

  Federated learning (FL) enables multiple clients to collaboratively train a shared model, with the help of a parameter server (PS), without disclosing their local datasets. However, due to the increasing size of the trained ...

• Content-aware playback speed control for low-latency live streaming of sports 

  Aladağ, Ö. F.; Uğur, Deniz; Akçay, Mehmet Necmettin; Beğen, Ali Cengiz (The ACM Digital Library, 2021)

  There are two main factors that determine the viewer experience during the live streaming of sports content: latency and stalls. Latency should be low and stalls should not occur. Yet, these two factors work against each ...

• COSMOS on steroids: a Cheap detector for cheapfakes 

  Akgül, T.; Civelek, Tuğçe Erkılıç; Uğur, Deniz; Beğen, Ali Cengiz (The ACM Digital Library, 2021)

  The growing prevalence of visual disinformation has become an important problem to solve nowadays. Cheapfake is a new term used for the altered media generated by non-AI techniques. In their recent COSMOS work, the authors ...

• Automated defect prioritization based on defects resolved at various project periods 

  Gökçeoğlu, M.; Sözer, Hasan (Elsevier, 2021-09)

  Defect prioritization is mainly a manual and error-prone task in the current state-of-the-practice. We evaluated the effectiveness of an automated approach that employs supervised machine learning. We used two alternative ...

• Emotion as an emergent phenomenon of the neurocomputational energy regulation mechanism of a cognitive agent in a decision-making task 

  Kırtay, M.; Vannucci, L.; Albanese, U.; Laschi, C.; Öztop, Erhan; Falotico, E. (Sage, 2021-02)

  Biological agents need to complete perception-action cycles to perform various cognitive and biological tasks such as maximizing their wellbeing and their chances of genetic continuation. However, the processes performed ...

• Would you imagine yourself negotiating with a robot, Jennifer? Why not? 

  Aydoğan, Reyhan; Keskin, Mehmet Onur; Çakan, Umut (IEEE, 2022-02)

  With the improvement of intelligent systems and robotics, social robots are becoming part of our society. To accomplish complex tasks, robots and humans may need to collaborate, and when necessary, they need to negotiate ...

• FedADC: Accelerated federated learning with drift control 

  Özfatura, E.; Özfatura, Ahmet Kerem; Gündüz, D. (IEEE, 2021)

  Federated learning (FL) has become de facto framework for collaborative learning among edge devices with privacy concern. The core of the FL strategy is the use of stochastic gradient descent (SGD) in a distributed manner. ...

• Common media client data (CMCD): Initial findings 

  Bentaleb, A.; Lim, M.; Akçay, Mehmet Necmettin; Beğen, Ali Cengiz; Zimmermann, R. (Association for Computing Machinery, Inc, 2021-07-16)

  In September 2020, the Consumer Technology Association (CTA) published the CTA-5004: Common Media Client Data (CMCD) specification. Using this specification, a media client can convey certain information to the content ...

• Campaign participation prediction with deep learning 

  Ayvaz, Demet; Aydoğan, Reyhan; Akçura, Munir Tolga; Şensoy, Murat (Elsevier, 2021-08)

  Increasingly, on-demand nature of customer interactions put pressure on companies to build real-time campaign management systems. Instead of having managers to decide on the campaign rules, such as, when, how and whom to ...

• Dropout regularization in hierarchical mixture of experts 

  Alpaydın, Ahmet İbrahim Ethem (Elsevier, 2021-01-02)

  Dropout is a very effective method in preventing overfitting and has become the go-to regularizer for multi-layer neural networks in recent years. Hierarchical mixture of experts is a hierarchically gated model that defines ...

• MaLeFICE: Machine learning support for continuous performance improvement in computational engineering 

  Sönmezer, Hasan Berk; Muhtaroğlu, Nitel; Arı, İsmail; Gökçin, Deniz (Wiley, 2022-04-25)

  Computer aided engineering (CAE) practices improved drastically within the last decade due to ease of access to computing resources and open-source software. However, increasing complexity of hardware and software settings ...

• On the use of evolutionary coupling for software architecture recovery 

  Saydemir, Abdullah; Simitcioğlu, Muhammed Esad; Sözer, Hasan (IEEE, 2021)

  Software architecture documentation can be partially obtained automatically by means of software architecture recovery tools. These tools mainly cluster software modules to provide a high level structural organization of ...

• Improving server and client-side algorithms for adaptive streaming of non-immersive and immersive media 

  Akçay, Mehmet Necmettin (The ACM Digital Library,, 2021)

  HTTP adaptive streaming is a technique widely used in the internet today to stream live and on-demand content. Server and client-side algorithms play an important role in achieving a better user experience in terms of ...

View more