Browsing by Author "Tseng, Wen-Jie"

Now showing 1 - 20 of 21

An analysis for strength improvement of an MCTS-based program playing Chinese dark chess
(2016-09-06)
Hsueh, Chu-Hsuan
;
Wu, I-Chen
;
Tseng, Wen-Jie
;
Yen, Shi-Jim
;
Chen, Jr-Chang
Monte Carlo tree search (MCTS) has been successfully applied to many games recently. Since then, many techniques are used to improve the strength of MCTS-based programs. This paper investigates four recent techniques: early playout terminations, implicit minimax backups, quality-based rewards and progressive bias. The strength improvements are analyzed by incorporating the techniques into an MCTS-based program, named DARKKNIGHT, for Chinese Dark Chess. Experimental results showed that the win rates against the original DARKKNIGHT were 60.75%, 71.85%, 59.00%, and 82.10%, respectively for incorporating the four techniques. The results indicated that the improvement by progressive bias was most significant. By incorporating all together, a better win rate of 84.75% was obtained. (C) 2016 Elsevier B.V. All rights reserved.
Chimo wins Chinese chess tournament
(2018-01-01)
Tseng, Wen-Jie
;
Chen, Jr-Chang
;
Wu, I-Chen
;
資訊工程學系
;
Department of Computer Science
Comparison Training for Computer Chinese Chess
(2020-06-01)
Chen, Jr-Chang
;
Tseng, Wen-Jie
;
Wu, I-Chen
;
Wei, Ting-Han
;
資訊工程學系
This paper describes the application of modified comparison training for automatic feature weight tuning. The final objective is to improve the evaluation functions used in Chinese chess programs. First, we apply n-tuple networks to extract features. N-tuple networks require very little expert knowledge through its large numbers of features, while simultaneously allowing easy access. Second, we propose a modified comparison training into which tapered eval is incorporated. Experiments show that with the same features and the same Chinese chess program, the automatically tuned feature weights achieved a win rate of 86.58% against the hand-tuned features. The above trained version was then improved by adding additional features, most importantly n-tuple features. This improved version achieved a win rate of 81.65% against the trained version without additional features.
Connect6 Programs on Mobile Devices
(2012)
Zheng, Ji-Hong
;
Hu, Chia-Yun
;
Wu, I-Chen
;
Tseng, Wen-Jie
;
Wei, Ching-Hsuan
In this demonstration, we port our current Connect6 program, NCTU6, to mobile devices, such as iOS devices, Android devices. However, computing powers of mobile devices are also relatively weaker when compared with high-end machines. For example, the CPU speeds of mobile devices are about 1/5 to 1/8 of those of desktops; and the memory of mobile devices is usually limited to several megabytes. In this demonstration, we tune the program NCTU6 to fit in mobile devices, while maintaining reasonable strength. The program also won the silver medal in TCGA 2012. In addition, we also support Connect6 puzzles in the program.
DARKKNIGHT WINS CHINESE DARK CHESS TOURNAMENT
(2013-09-01)
Yen, Shi-Jim
;
Chen, Jr-Chang
;
Chen, Bo-Nian
;
Tseng, Wen-Jie
;
資訊工程學系
DarkKnight Wins Chinese Dark Chess Tournament
(2017-01-01)
Tseng, Wen-Jie
;
Chen, Jr-Chang
;
Wu, I-Chen
;
資訊工程學系
;
Department of Computer Science
FacePush: Experiencing Pressure Forces on Face with HMDs
(2018-01-01)
Chang, Hong-Yu
;
Tseng, Wen-Jie
;
Tsai, Chia-En
;
Chen, Hsin-Yu
;
Peiris, Roshan Lalintha
FacePush: Introducing Normal Force on Face with Head-Mounted Displays
(2018-01-01)
Chang, Hong-Yu
;
Tseng, Wen-Jie
;
Tsai, Chia-En
;
Chen, Hsin-Yu
;
Peiris, Roshan Lalintha
This paper presents FacePush, a Head-Mounted Display (HMD) integrated with a pulley system to generate normal forces on a user's face in virtual reality (VR). The mechanism of FacePush is obtained by shifting torques provided by two motors that press upon a user's face via utilization of a pulley system. FacePush can generate normal forces of varying strengths and apply those to the surface of the face. To inform our design of FacePush for noticeable and discernible normal forces in VR applications, we conducted two studies to identify the absolute detection threshold and the discrimination threshold for users' perception. After further consideration in regard to user comfort, we determined that two levels of force, 2.7 kPa and 3.375 kPa, are ideal for the development of the FacePush experience via implementation with three applications which demonstrate use of discrete and continuous normal force for the actions of boxing, diving, and 360 guidance in virtual reality. In addition, with regards to a virtual boxing application, we conducted a user study evaluating the user experience in terms of enjoyment and realism and collected the user's feedback.
FaceWidgets: Exploring Tangible Interaction on Face with Head-Mounted Displays
(2019-01-01)
Tseng, Wen-Jie
;
Wang, Li-Yang
;
Chan, Liwei
;
資訊工程學系
;
Department of Computer Science
We present FaceWidgets, a device integrated with the back-side of a head-mounted display (HMD) that enables tangible interactions using physical controls. To allow for near range-to-eye interactions, our first study suggested displaying the virtual widgets at 20 cm from the eye positions, which is 9 cm from the HMD backside. We propose two novel interactions, widget canvas and palm-facing gesture, that can help users avoid double vision and allow them to access the interface as needed. Our second study showed that displaying a hand reference improved performance of face widgets interactions. We developed two applications of FaceWidgets, a fixed-layout 360 video player and a contextual input for smart home control. Finally, we compared four hand visualizations against the two applications in an exploratory study. Participants considered the transparent hand as the most suitable and responded positively to our system.
HAPPYNOGO WINS NOGO TOURNAMENT
(2013-09-01)
Tseng, Wen-Jie
;
Chen, Jr-Chang
;
She, Pohsuan
;
Wu, I-Chen
;
資訊工程學系
JOB-LEVEL ALGORITHMS FOR CONNECT6 OPENING BOOK CONSTRUCTION
(2015-09-01)
Wei, Ting Han
;
Wu, I-Chen
;
Liang, Chao-Chin
;
Chiang, Bing-Tsung
;
Tseng, Wen-Jie
This article investigates the use of job-level (JL) algorithms for Connect6 opening book construction. In the past, JL proof-number search (JL-PNS) was successfully used to solve Connect6 positions. Using JL-PNS, many opening plays that lead to losses can be eliminated from consideration during the opening game. However, it is unclear how the information of unsolved positions can be exploited for opening book construction. For this issue, the current article first proposes four heuristic metrics when using JL-PNS to estimate move quality. The article then proposes a JL upper confidence tree (JL-UCT) algorithm and three heuristic metrics that work with JL-UCT. Of the three, the best way to estimate move quality for JL-UCT is the number of nodes in each candidate move's subtree. In order to compare the heuristic metrics among the two algorithms objectively, we proposed two kinds of measurement methods to analyze the suitability of these metrics when choosing best moves for a set of benchmark positions. Experimental results show that the node count heuristic metric for JL-UCT outperforms all other heuristic metrics, including the four for JL-PNS. We then verify the results by constructing three separate opening books using the top three performing heuristic metrics. Competitive play also shows that the node count heuristic metric for JL-UCT is most suitable among the currently proposed heuristic metrics for Connect6 opening book construction.
Job-Level Algorithms for Connect6 Opening Position Analysis
(2014-01-01)
Wei, Ting-Han
;
Wu, I-Chen
;
Liang, Chao-Chin
;
Chiang, Bing-Tsung
;
Tseng, Wen-Jie
This paper investigates job-level (JL) algorithms to analyze opening positions for Connect6. The opening position analysis is intended for opening book construction, which is not covered by this paper. In the past, JL proof-number search (JL-PNS) was successfully used to solve Connect6 positions. Using JL-PNS, many opening plays that lead to losses can be eliminated from consideration during the opening game. However, it is unclear how the information of unsolved positions can be exploited for opening book construction. For this issue, this paper first proposes four heuristic metrics when using JL-PNS to estimate move quality. This paper then proposes a JL upper confidence tree (JL-UCT) algorithm and some heuristic metrics, one of which is the number of nodes in each candidate move\'s subtree. In order to compare these metrics objectively, we proposed two kinds of measurement methods to analyze the suitability of these metrics when choosing best moves for a set of benchmark positions. The results show that for both metrics this node count heuristic metric for JL-UCT outperforms all the others, including the four for JL-PNS.
Job-Level Alpha-Beta Search
(2015-03-01)
Chen, Jr-Chang
;
Wu, I-Chen
;
Tseng, Wen-Jie
;
Lin, Bo-Han
;
Chang, Chia-Hui
An approach called generic job-level (JL) search was proposed to solve computer game applications by dispatching jobs to remote workers for parallel processing. This paper applies JL search to alpha-beta search, and proposes a JL alpha-beta search (JL-ABS) algorithm based on a best-first search version of MTD(f). The JL-ABS algorithm is demonstrated by using it in an opening book analysis for Chinese chess. The experimental results demonstrated that JL-ABS reached a speed-up of 10.69 when using 16 workers in the JL system.
LONGCAT WINS MAHJONG TOURNAMENT IN ICGA 2013
(2013-09-01)
Tseng, Wen-Jie
;
Chuang, Li-Kai
;
Wu, I-Chen
;
Lin, Shun-Shii
;
Yen, Shi-Jim
Merging Metrics of Special Rules in Chinese Chess Endgame Databases
(2019-01-01)
Tseng, Wen-Jie
;
Chen, Jr-Chang
;
Wu, I-Chen
;
交大名義發表
;
資訊工程學系
Perpetual check and perpetual chase are two special rules in the endgame phase of Chinese chess. Two metrics, distance-to-check and distance-to-chase, respectively record the distances measured in plies from the current position to the positions where the two rules apply in endgame databases. In the past, the two metrics are separately recorded at different levels in the databases. However, the endgame strategy is not efficient if the distances of the two metrics are not compared. Moreover, the more the number of levels is, the bigger the database is. In this paper, we modify retrograde analysis, an algorithm to build endgame databases, to merge the two metrics for special rules. The experiments show that the endgame strategy becomes efficient and the sizes of endgame databases are reduced as well.
MOBILE6 WINS CONNECT6 TOURNAMENT
(2013-09-01)
Wei, Ting-Han
;
Tseng, Wen-Jie
;
Wu, I-Chen
;
Yen, Shi-Jim
;
資訊工程學系
NONOGRAM TOURNAMENT IN OLYMPIAD YOKOHAMA 2013
(2014-03-01)
Chen, Lung-Pin
;
Sun, Der-Johng
;
Tseng, Wen-Jie
;
資訊工程學系
;
Department of Computer Science
On the Efficiency of Executing Diverse Game Tree Search Applications in a Volunteer Computing Federation
(2013-01-01)
Chen, Lung-Ping
;
Wu, I-Chen
;
Chang, Yuan-Yao
;
Tseng, Wen-Jie
;
資訊工程學系
A game tree search application can be implemented as the malleable parallel jobs that are adapted to various processor allocations. We establish an efficient desktop grid federation to enable the small to mid-sized organizations to perform large-scale game tree search tasks via resource sharing. Due to the uneven task scales of the organizations as well as the dynamic generation/pruning of game tree search tasks, the user credits of the desktop grids may fluctuate dramatically, leading an unstable resource allocation to the hosted applications. This paper shows that stable processor allocation leads to higher efficiency for the parallel tasks. A new brokering algorithm is developed that ensures both fairness and stable resource allocation.
Only-One-Victor Pattern Learning in Computer Go
(2017-03-01)
Wang, Jiao
;
Xiao, Chenjun
;
Zhu, Tan
;
Hsueh, Chu-Husan
;
Tseng, Wen-Jie
Automatically acquiring domain knowledge from professional game records, a kind of pattern learning, is an attractive and challenging issue in computer Go. This paper proposes a supervised learning method, by introducing a new generalized Bradley-Terry model, named Only-One-Victor, to learn patterns from game records. Basically, our algorithm applies the same idea with Elo rating algorithm, which considers each move in game records as a group of move patterns, and the selected move as the winner of a kind of competition among all groups on current board. However, being different from the generalized Bradley-Terry model for group competition used in Elo rating algorithm, Only-One-Victor model in our work simulates the process of making selection from a set of possible candidates by considering such process as a group of independent pairwise comparisons. We use a graph theory model to prove the correctness of Only-One-Victor model. In addition, we also apply the Minorization-Maximization (MM) to solve the optimization task. Therefore, our algorithm still enjoys many computational advantages of Elo rating algorithm, such as the scalability with high dimensional feature space. With the training set containing 115,832 moves and the same feature setting, the results of our experiments show that Only-One-Victor outperforms Elo rating, a well-known best supervised pattern learning method.
Strength Improvement and Analysis for an MCTS-Based Chinese Dark Chess Program
(2015)
Hsueh, Chu-Hsuan
;
Wu, I-Chen
;
Tseng, Wen-Jie
;
Yen, Shi-Jim
;
Chen, Jr-Chang
Monte-Carlo tree search (MCTS) has been successfully applied to Chinese dark chess (CDC). In this paper, we study how to improve and analyze the playing strength of an MCTS-based CDC program, named DARKKNIGHT, which won the CDC tournament in the 17th Computer Olympiad. We incorporate the three recent techniques, early playout terminations, implicit minimax backups, and quality-based rewards, into the program. For early playout terminations, playouts end when reaching states with likely outcomes. Implicit minimax backups use heuristic evaluations to help guide selections of MCTS. Quality-based rewards adjust rewards based on online collected information. Our experiments showed that the win rates against the original DARKKNIGHT were 60.75 %, 70.90 % and 59.00 %, respectively for incorporating the three techniques. By incorporating all together, we obtained a win rate of 76.70 %.