501 Hits in 1.0 sec

Using Multi-Agent Reinforcement Learning in Auction Simulations [article]

Medet Kanmaz, Elif Surer
2020 arXiv   pre-print
Game theory has been developed by scientists as a theory of strategic interaction among players who are supposed to be perfectly rational. These strategic interactions might have been presented in an auction, a business negotiation, a chess game, or even in a political conflict aroused between different agents. In this study, the strategic (rational) agents created by reinforcement learning algorithms are supposed to be bidder agents in various types of auction mechanisms such as British
more » ... , Sealed Bid Auction, and Vickrey Auction designs. Next, the equilibrium points determined by the agents are compared with the outcomes of the Nash equilibrium points for these environments. The bidding strategy of the agents is analyzed in terms of individual rationality, truthfulness (strategy-proof), and computational efficiency. The results show that using a multi-agent reinforcement learning strategy improves the outcomes of the auction simulations.
arXiv:2004.02764v1 fatcat:haxdupddxbg4dazwrv5qbq4xxe

Playtesting: What is Beyond Personas [article]

Sinan Ariyurek, Elif Surer, Aysu Betin-Can
2022 arXiv   pre-print
Playtesting is an essential step in the game design process. Game designers use the feedback from playtests to refine their designs. Game designers may employ procedural personas to automate the playtesting process. In this paper, we present two approaches to improve automated playtesting. First, we propose developing persona, which allows a persona to progress to different goals. In contrast, the procedural persona is fixed to a single goal. Second, a human playtester knows which paths she has
more » ... tested before, and during the consequent tests, she may test different paths. However, Reinforcement Learning (RL) agents disregard these previous paths. We propose a novel methodology that we refer to as Alternative Path Finder (APF). We train APF with previous paths and employ APF during the training of an RL agent. APF modulates the reward structure of the environment while preserving the agent's goal. When evaluated, the agent generates a different trajectory that achieves the same goal. We use the General Video Game Artificial Intelligence (GVG-AI) and VizDoom frameworks to test our proposed methodologies. We use Proximal Policy Optimization (PPO) RL agent during experiments. First, we compare the playtest data generated by developing and procedural persona. Our experiments show that developing persona provides better insight into the game and how different players would play. Second, we present the alternative paths found using APF and argue why traditional RL agents cannot learn those paths.
arXiv:2107.11965v2 fatcat:o4vaa6wunjcydeqsattjkkadeq

Automated Video Game Testing Using Synthetic and Human-Like Agents [article]

Sinan Ariyurek, Aysu Betin-Can, Elif Surer
2019 arXiv   pre-print
Automated Video Game Testing Using Synthetic and Human-Like Agents Sinan Ariyurek, Aysu Betin-Can, Elif Surer Graduate School of Informatics Middle East Technical University 06800, Ankara, Turkey {sinan.ariyurek  ...  , betincan, elifs} Abstract-In this paper, we present a new methodology that employs tester agents to automate video game testing.  ... 
arXiv:1906.00317v1 fatcat:7n2lynwyivbqxaea34ejckfsxa

Boosted Multiple Kernel Learning For First-Person Activity Recognition

Mehmet Arabaci, Fatih Ozkan, Elif Surer, Alptekin Temizel
2018 Zenodo  
Publication in the conference proceedings of EUSIPCO, Kos island, Greece, 2017
doi:10.5281/zenodo.1159451 fatcat:ksr2frb7xfas3gkcqhosxlewbu

Enhancing the Monte Carlo Tree Search Algorithm for Video Game Testing [article]

Sinan Ariyurek, Aysu Betin-Can, Elif Surer
2020 arXiv   pre-print
In this paper, we study the effects of several Monte Carlo Tree Search (MCTS) modifications for video game testing. Although MCTS modifications are highly studied in game playing, their impacts on finding bugs are blank. We focused on bug finding in our previous study where we introduced synthetic and human-like test goals and we used these test goals in Sarsa and MCTS agents to find bugs. In this study, we extend the MCTS agent with several modifications for game testing purposes. Furthermore,
more » ... we present a novel tree reuse strategy. We experiment with these modifications by testing them on three testbed games, four levels each, that contain 45 bugs in total. We use the General Video Game Artificial Intelligence (GVG-AI) framework to create the testbed games and collect 427 human tester trajectories using the GVG-AI framework. We analyze the proposed modifications in three parts: we evaluate their effects on bug finding performances of agents, we measure their success under two different computational budgets, and we assess their effects on human-likeness of the human-like agent. Our results show that MCTS modifications improve the bug finding performance of the agents.
arXiv:2003.07813v1 fatcat:zm3n52lrcfh5xpsrgn7uasqbwq

Using Generative Adversarial Nets on Atari Games for Feature Extraction in Deep Reinforcement Learning [article]

Ayberk Aydın, Elif Surer
2020 arXiv   pre-print
Deep Reinforcement Learning (DRL) has been successfully applied in several research domains such as robot navigation and automated video game playing. However, these methods require excessive computation and interaction with the environment, so enhancements on sample efficiency are required. The main reason for this requirement is that sparse and delayed rewards do not provide an effective supervision for representation learning of deep neural networks. In this study, Proximal Policy
more » ... n (PPO) algorithm is augmented with Generative Adversarial Networks (GANs) to increase the sample efficiency by enforcing the network to learn efficient representations without depending on sparse and delayed rewards as supervision. The results show that an increased performance can be obtained by jointly training a DRL agent with a GAN discriminator. ---- Derin Pekistirmeli Ogrenme, robot navigasyonu ve otomatiklestirilmis video oyunu oynama gibi arastirma alanlarinda basariyla uygulanmaktadir. Ancak, kullanilan yontemler ortam ile fazla miktarda etkilesim ve hesaplama gerektirmekte ve bu nedenle de ornek verimliligi yonunden iyilestirmelere ihtiyac duyulmaktadir. Bu gereksinimin en onemli nedeni, gecikmeli ve seyrek odul sinyallerinin derin yapay sinir aglarinin etkili betimlemeler ogrenebilmesi icin yeterli bir denetim saglayamamasidir. Bu calismada, Proksimal Politika Optimizasyonu algoritmasi Uretici Cekismeli Aglar (UCA) ile desteklenerek derin yapay sinir aglarinin seyrek ve gecikmeli odul sinyallerine bagimli olmaksizin etkili betimlemeler ogrenmesi tesvik edilmektedir. Elde edilen sonuclar onerilen algoritmanin ornek verimliliginde artis elde ettigini gostermektedir.
arXiv:2004.02762v1 fatcat:cah4426ubvf4rdclgtqglom4hy

Relational-Grid-World: A Novel Relational Reasoning Environment and An Agent Model for Relational Information Extraction [article]

Faruk Kucuksubasi, Elif Surer
2020 arXiv   pre-print
Reinforcement learning (RL) agents are often designed specifically for a particular problem and they generally have uninterpretable working processes. Statistical methods-based agent algorithms can be improved in terms of generalizability and interpretability using symbolic Artificial Intelligence (AI) tools such as logic programming. In this study, we present a model-free RL architecture that is supported with explicit relational representations of the environmental objects. For the first
more » ... we use the PrediNet network architecture in a dynamic decision-making problem rather than image-based tasks, and Multi-Head Dot-Product Attention Network (MHDPA) as a baseline for performance comparisons. We tested two networks in two environments ---i.e., the baseline Box-World environment and our novel environment, Relational-Grid-World (RGW). With the procedurally generated RGW environment, which is complex in terms of visual perceptions and combinatorial selections, it is easy to measure the relational representation performance of the RL agents. The experiments were carried out using different configurations of the environment so that the presented module and the environment were compared with the baselines. We reached similar policy optimization performance results with the PrediNet architecture and MHDPA; additionally, we achieved to extract the propositional representation explicitly ---which makes the agent's statistical policy logic more interpretable and tractable. This flexibility in the agent's policy provides convenience for designing non-task-specific agent architectures. The main contributions of this study are two-fold ---an RL agent that can explicitly perform relational reasoning, and a new environment that measures the relational reasoning capabilities of RL agents.
arXiv:2007.05961v1 fatcat:rfzftkcmu5cxdcnk5fzuizimti

Boosted Multiple Kernel Learning for First-Person Activity Recognition [article]

Fatih Ozkan, Mehmet Ali Arabaci, Elif Surer, Alptekin Temizel
2017 arXiv   pre-print
Activity recognition from first-person (ego-centric) videos has recently gained attention due to the increasing ubiquity of the wearable cameras. There has been a surge of efforts adapting existing feature descriptors and designing new descriptors for the first-person videos. An effective activity recognition system requires selection and use of complementary features and appropriate kernels for each feature. In this study, we propose a data-driven framework for first-person activity
more » ... which effectively selects and combines features and their respective kernels during the training. Our experimental results show that use of Multiple Kernel Learning (MKL) and Boosted MKL in first-person activity recognition problem exhibits improved results in comparison to the state-of-the-art. In addition, these techniques enable the expansion of the framework with new features in an efficient and convenient way.
arXiv:1702.06799v2 fatcat:4kdn7ymdgrfu5krnhywlkom4zy

Behavior and usability analysis for multimodal user interfaces

Hamdi Dibeklioğlu, Elif Surer, Albert Ali Salah, Thierry Dutoit
2021 Journal on Multimodal User Interfaces  
B Hamdi Dibeklioglu Elif Surer Albert Ali Salah Thierry Dutoit Department of Computer Engineering  ...  In another game study, Surer et al. focus on the usability of a scenario-based game generator framework created by game developers for the Chemical, Biological, Radiological, Nuclear, and Explosives (CBRNe  ... 
doi:10.1007/s12193-021-00372-0 fatcat:pappj7oc7jfsxcy5mrqg3mnj2e


Emin Alp BIYIK, Elif SÜRER
2020 Mugla Journal of Science and Technology  
Procedural content generation (PCG) has been an essential catalyzer in the last decade with its efficiency in creating game elements such as textures, game levels, and maps. Despite being successfully applied in various studies, new reliable evaluation tools are still needed to assess the quality of the generated game content. One example limitation of procedurally generated game worlds is the lacking spatial configuration. To address this issue, in this study, an assessment method was
more » ... to evaluate the spatial quality of procedurally generated game worlds. For this purpose, Space Syntax was used, which incorporates a set of methods to analyze spatial configurations and movement. The analyses were applied to a new game developed by the authors -the Haunted House-and the performance was evaluated in terms of integration, connectivity, and depth distance. Results show that changing the room dimensions (i.e., 15x15, 25x25, and 35x35 units) modifies the performance measures as well as game design parameters -number of the spawning points (ranging from 1 to 4), critical axes (1 to 5), to name a few. The proposed approach is a first attempt to create various improved spatial configurations and provide an evaluation tool to analyze the PCG algorithms in level design. PROSEDÜREL OLARAK ÜRETİLMİŞ OYUN SEVİYELERİ İÇİN MEKAN DİZİLİMİ BAZLI DEĞERLENDİRME YÖNTEMİ GELİŞTİRİLMESİ Özet Prosedürel içerik üretimi, dokular, oyun seviyeleri ve haritalar gibi oyun unsurlarını yaratmadaki verimliliği ile son on yılda önemli bir katalizör olmuştur. Çeşitli çalışmalarda başarılı bir şekilde uygulanmasına rağmen, oluşturulan oyun içeriğinin kalitesini değerlendirmek için yeni ve güvenilir değerlendirme araçlarına hala ihtiyaç vardır. Prosedürel olarak oluşturulmuş oyun dünyalarının bir örnek sınırlaması, eksik mekansal konfigürasyondur. Bu sorunu ele almak için, bu çalışmada, prosedürel olarak oluşturulan oyun dünyalarının mekansal kalitesini değerlendirmek için bir değerlendirme yöntemi geliştirilmiştir. Bu amaçla, mekansal konfigürasyonları ve hareketi analiz etmek için bir dizi yöntem içeren Mekan Dizilimi (Space Syntax) kullanılmıştır. Analizler, yazarlar tarafından geliştirilmiş yeni bir oyun -Perili Ev-üzerinde uygulanmış ve yöntemin performansı entegrasyon, bağlantı ve derinlik mesafesi açısından değerlendirilmiştir. Sonuçlar, oda boyutlarının değiştirilmesinin (15x15, 25x25 ve 35x35 boyutlarında) performans ölçümlerinin yanı sıra oyun elementlerinin ortaya çıkma noktalarının sayısı (1 ila 4 arasında), kritik eksenler (1 ila 5) başta olmak üzere oyun tasarım parametrelerini de değiştirdiğini göstermektedir. Önerilen yaklaşım, çeşitli gelişmiş mekansal konfigürasyonlar yaratmaya yönelik bir ilk girişimdir ve seviye tasarımında prosedürel içerik algoritmalarını analiz etmek için bir değerlendirme aracı sağlamaktadır.
doi:10.22531/muglajsci.706895 fatcat:uwsiifivcfc5fchjh2lwohwukq

Developing a Scenario-Based Video Game Generation Framework: Preliminary Results [article]

Elif Surer, Mustafa Erkayaoğlu, Zeynep Nur Öztürk, Furkan Yücel, Emin Alp Bıyık, Burak Altan, Büşra Şenderin, Zeliha Oğuz, Servet Gürer, H. Şebnem Düzgün
2019 arXiv   pre-print
Surer Elif Surer received her Ph.D in Bioengineering in 2011 from the University of Bologna.  ...  Elif Zeynep NurÖztürk is studying on her Bachelor's degree in Computer Science at Bilkent University.  ... 
arXiv:1911.07380v1 fatcat:h2sjzvv5ozccfoosnergq4xwee

Automated Game Mechanics and Aesthetics Generation Using Neural Style Transfer in 2D Video Games

Deniz ŞEN, Hasan Tahsin KÜÇÜKKAYKI, Elif SÜRER
2021 Bilişim Teknolojileri Dergisi  
Video oyunu araştırması, karmaşık yöntemlerin ve algoritmaların geliştirildiği, sürekli değişmekte olan, dinamik bir alandır. Prosedürel içerik üretimi, kullanıcı tarafından oluşturulan parçaları video oyunu içeriğini otomatikleştirmek ve geliştirmek için algoritmalarla birleştirmeyi amaçlamakta ve bu yöntemlerin temelini oluşturmaktadır. Bununla birlikte, sonuçlar oyun mekaniğine ve oyunun oynanış biçimine değil, çoğunlukla oyun estetiğine yansımaktadır. Bu çalışmada, "tuval olarak oyun
more » ... " konsepti ile kullanıma hazır çarpıştırıcılar ve oyun estetiğini geliştiren, sanatsal açıdan farklı stiller kullanarak iki boyutlu oyun seviyesindeki bir görüntüyü basit bir prototip oyun geliştirme ortamına dönüştürebilen yöntem ve süreç sunulmaktadır. Bu amaçla, giriş oyun seviyesi görüntüsünün kenar ve renk bazlı özellikleri Canny kenar belirleme, basit doğrusal yinelemeli kümeleme ve Felzenszwalb segmentasyonu kullanılarak çıkarılmaktadır. Daha sonra, Unity oyun motoru, mekansal kontrol ile oyun seviyesinin stilinin aktarıldığı kenar ve renk özelliklerine göre çarpıştırıcılar oluşturmak için kullanılmaktadır. Farklı sinir stil transfer algoritmalarının sonuçları, Super Mario, Lode Runner ve Kid Icarus gibi oyunlar üzerinde karşılaştırılmakta ve tartışılmaktadır. Sonuçlar, bu çalışmanın oyun mekaniği ve oyun estetiğine odaklanarak iki boyutlu video oyunu geliştirmeyi kolaylaştırma potansiyeline sahip bir araç olduğunu göstermektedir.
doi:10.17671/gazibtd.706884 fatcat:r4iqnr3pnvelve5bqzhdib4jjy

Human inbreeding has decreased in time through the Holocene [article]

Francisco C Ceballos, Kanat Gurun, N. Ezgi Altinisik, Hasan Can Gemici, Cansu Karamurat, Dilek Koptekin, Kivilcim Basak Vural, Elif Surer, Yilmaz Selim Erdal, Anders Gotherstrom, Fusun Ozer, Cigdem Atakuman (+1 others)
2020 bioRxiv   pre-print
The history of human inbreeding is controversial. The development of sedentary agricultural societies may have had opposite influences on inbreeding levels. On the one hand, agriculture and food surplus may have diminished inbreeding by increasing population sizes and lowering endogamy, i.e. inbreeding due to population isolation. On the other hand, increased sedentism, as well as the advent of private property may have promoted inbreeding through the emergence of consanguineous marriage
more » ... or via ethnic and caste endogamy. The net impact is unknown, and to date, no systematic study on the temporal frequency of inbreeding in human societies has been conducted. Here we present a new approach for reliable estimation of runs of homozygosity (ROH) in genomes with ≥3x mean coverage across >1 million SNPs, and apply this to 440 ancient Eurasian genomes from the last 15,000 years. We show that the frequency of inbreeding, as measured by ROH, has decreased over time. The strongest effect is associated with the Neolithic transition, but the trend has since continued, indicating a population size effect on inbreeding prevalence. We further show that most inbreeding in our historical sample can be attributed to endogamy, although singular cases of high consanguinity can also be found in the archaeogenomic record.
doi:10.1101/2020.09.24.311597 fatcat:wwdbopamvjfyzefmfcnhi3l7hu

Developing Adaptive Serious Games for Children With Specific Learning Difficulties: A Two-phase Usability and Technology Acceptance Study

Oguzcan Yildirim, Elif Surer
2021 JMIR Serious Games  
Specific learning difficulties (SpLD) include several disorders such as dyslexia, dyscalculia, and dysgraphia, and the children with these SpLD receive special education. However, the studies and the educational material so far focus mainly on one specific disorder. Objective This study's primary goal is to develop comprehensive training material for different types of SpLD, with five serious games addressing different aspects of the SpLD. The second focus is measuring the impact of adaptive
more » ... ficulty level adjustment in the children's and their educators' usability and technology acceptance perception. Receiving feedback from the children and their educators, and refining the games according to their suggestions have also been essential in this two-phase study. Methods A total of 10 SpLD educators and 23 children with different types of SpLD tested the prototypes of the five serious games (ie, Word game, Memory game, Category game, Space game, and Math game), gave detailed feedback, answered the System Usability Scale and Technology Acceptance Model (TAM) questionnaires, and applied think-aloud protocols during game play. Results The games' standard and adaptive versions were analyzed in terms of average playtime and the number of false answers. Detailed analyses of the interviews, with word clouds and player performances, were also provided. The TAM questionnaires' average and mean values and box plots of each data acquisition session for the children and the educators were also reported via System Usability Scale and TAM questionnaires. The TAM results of the educators had an average of 8.41 (SD 0.87) out of 10 in the first interview and an average of 8.71 (SD 0.64) out of 10 in the second interview. The children had an average of 9.07 (SD 0.56) out of 10 in the first interview. Conclusions Both the educators and the children with SpLD enjoyed playing the games, gave positive feedback, and suggested new ways for improvement. The results showed that these games provide thorough training material for different types of SpLD with personalized and tailored difficulty systems. The final version of the proposed games will become a part of the special education centers' supplementary curriculum and training materials, making new enhancements and improvements possible in the future.
doi:10.2196/25997 pmid:34057415 fatcat:mgui7tko6rd23hsazrakgru4wm

A Canine Gait Analysis Protocol for Back Movement Assessment in German Shepherd Dogs

Elif Surer, Andrea Cereatti, Maria Antonietta Evangelisti, Gabriele Paolini, Ugo Della Croce, Maria Lucia Manunta
2020 Veterinary Sciences  
Objective—To design and test a motion analysis protocol for the gait analysis of adult German Shepherd (GS) dogs with a focus in the analyses of their back movements. Animals—Eight clinically healthy adult large-sized GS dogs (age, 4 ± 1.3 years; weight, 38.8 ± 4.2 kg). Procedures—A six-camera stereo-photogrammetric system and two force platforms were used for data acquisition. Experimental acquisition sessions consisted of static and gait trials. During gait trials, each dog walked along a 6 m
more » ... long walkway at self-selected speed and a total of six gait cycles were recorded. Results—Grand mean and standard deviation of ground reaction forces of fore and hind limbs are reported. Spatial-temporal parameters averaged over gait cycles and subjects, their mean, standard deviation and coefficient of variance are analyzed. Joint kinematics for the hip, stifle and tarsal joints and their average range of motion (ROM) values, and their 95% Confidence Interval (CI) values of kinematics curves are reported. Conclusions and Clinical Relevance—This study provides normative data of healthy GS dogs to form a preliminary basis in the analysis of the spatial-temporal parameters, kinematics and kinetics during quadrupedal stance posture and gait. Also, a new back movement protocol enabling a multi-segment back model is provided. Results show that the proposed gait analysis protocol may become a useful and objective tool for the evaluation of canine treatment with special focus on the back movement.
doi:10.3390/vetsci7010026 pmid:32092869 fatcat:x4jftshq4jdotkrx4igzfnciii
« Previous Showing results 1 — 15 out of 501 results