A small Go board Study of metric and dimensional (2002)
The difficulty to write successful 19x19 go programs lies not only in the combinatorial complexity of go but also in the complexity of designing a good evaluation function containing a lot of knowledge. |
Mathematical morphology applied to computer go
This paper shows how mathematical morphological operators can be applied to computer go. On one hand, mathematical morphology is a very powerful tool within image processing community. On the other hand, the Zobrist's model is well-known within the computer go community for its "influence" recognition. We present a model, derived from the closing operator of mathematical morphology and from the Zobrist's model. |
Global and Local Game Tree Search (2000)
Minimax search has been used with great success to solve a number of games including Gomoku and Nine Men's Morris, and to reach a performance approaching or surpassing the best human players in other well-known games such as checkers, Othello and chess. All these highperformance game-playing programs use global search methods, which evaluate complete game positions. Local search is an alternative approach. |
The INDIGO program (1995)
This paper is now to give a ten pages description of this program. In the first part we describe INDIGO. In particular, we show how an interaction model between groups is useful to static evaluation of life and death of not fully cercled groups. Also, we show how fuzzy mathematical morphology is useful to static evaluation of territories. In the second part, before conclusion, we show the results and we discuss about strong and weak points of our approach. |
Strategic Evaluation in Complex Domains (1998)
In some complex domains, like the game of Go, evaluating a position is not simple. In other games, like Chess for example, material balance gives good and fast to compute insight on the value of a position. In Go all the stones have the same value, so material balance is not a good heuristic. To evaluate a Go position, a computer needs a lot of knowledge and much more time. |
Development and Evaluation of Strategic Plans (1997)
At the strategic level, a Go program has to manage uncertainty because of the difficulty to correctly evaluate middle game positions (strength of groups, battles). It has to be cautious not to rely on too many uncertain assumptions, otherwise its opponent will find a weakness in the plan. When faced with multiple choices for achieving a given strategic goal, we provide a method for assessing the least hazardous plan (a plan is a subtree of goals that leads to the success of the root goal). |
Applying Adversarial Planning Techniques to Go (2001)
Approaches to computer game playing based on alpha-beta search of the tree of possible move sequences combined with a position evaluation function have been successful for many games, notably Chess. Such approaches are less successful for games with large search spaces and complex positions, such as Go, and we are led to seek alternatives. |
The Move Decision Strategy of Indigo
This paper describes the move decision strategy of Indigo. By using the example of Indigo, the paper shows that the move decision process of a Go program can be very different from the processes used in other games with lower complexity than the complexity of Go, even if the basic modules are conventional (move generator, evaluation function and tree search). Indigo uses them in a specific way, adapted to computer Go. |
Machine Learning, Game Play, and Go (1991)
The game of go is an ideal problem domain for exploring machine learning: it is easy to define and there are many human experts, yet existing programs have failed to emulate their level of play to date. |
Metaprogramming domain specific metaprograms (1999)
When a metaprogram automatically creates rules, some created rules are useless because they can never apply. Some metarules, that we call impossibility metarules, are used to remove useless rules. |
Incremental Updating of Objects in INDIGO (1997)
This paper shows the incremental updating that is used in the Go playing program Indigo. Due to the size of the board and time constraints, incremental mechanisms are relevant to update data. The evaluation of a position includes the construction of a taxonomy of objects which are linked by a lot dependencies. |
On Some Combinatorial Games Connected with Go (1993)
The two-person strategy game of Go has the feature that, with a simple set of rules, situations called kos allowing infinitely long play often arise. |
Partial Order Bounding: A new Approach to Evaluation in Game Tree Search
In computer game-playing, the established method for constructing an evaluation function uses a scalar value computed as a weighted sum of features. This paper advocates the use of partial order evaluation, and describes an ecient new search method called partial order bounding (POB). |
Multiobjective Heuristic State-Space Planning (2001)
Modern domain-independent heuristic planners evaluate their plans on the basis of their length only. However, in real-world problems there are other criteria that also play an important role, such as resource consumption, profit, safety, etc. This paper extends the GRT planner, an efficient domain-independent heuristic state-space planner, with the ability to consider multiple criteria. |
Pursuing abstract goals in the game of Go (2001)
Reasoning and planning at dierent levels of abstraction is an important skill in the game of Go, for both human and computer players. |
Studies in Human and Computer Go: Assessing the Game of Go as a Research Domain for Cognitive Scienc
This thesis assesses the game of Go as a research domain for Cognitive Science by investigating some of the research issues within the domain of Go, and in particular, by assessing Go as a research domain for both Artificial Intelligence and Cognitive Psychology. |
Automatic Acquisition of Tactical Go Rules (1996)
Gogol is a rule-based computer Go program. It uses a lot of reliable tactical rules. Tactical rules are rules about simple goals such as connecting and making an eye. Gogol uses a simplified game theory to represent the degree of achievement of the goals |
The Zen Way to Go
The more I read from the Zen masters, the more connections I see between its legacy and the goban. Ergo, here is a collection of quotations from the Zen tradition which strike me as being directly relevant to Go. |
Admissible Moves in Two-player Games (2002)
Some games have abstract properties that can be used to design admissible heuristics on moves. These admissible heuristics are useful to speed up search. |
Metarules to Improve Tactical Go Knowledge (2002)
Three main problems arise with automatically generated rules databases. They are too large to fit in memory, they can take a lot of time to generate, and it takes time to match many rules on a board. |
Evolving Go Playing Strategy in Neural Networks (1994)
The game of Go is an ideal problem domain for exploring machine learning: it has simple rules yet requires more and more complex strategies to play well as the board size is increased. Despite much effort, existing Go programs, which have largely employed knowledge based and symbolic AI techniques, have failed to achieve a standard much above an average human amateur |
Solving Ponnuki-Go on Small Boards
This paper presents a search-based approach for the game of PonnukiGo. |
Not Like Other Games - Why Tree Search in Go is Different (2000)
Large-scale minimax search has been used with great success in many games, but not in Go. We investigate the reasons for the diculty of applying minimax search to Go, using late stage endgames as a test case. |
Visual learning in Go
Learning to play Go from the rules alone is extremely hard for computers. |
Machine Introspection for Machine Learning
Playing systems rely on a knowledge intensive approach. A Go expert uses a large number of rules. Go programmers usually try to enter these rules by hand in a Go program. Creating this large number of rules requires a high level of expertise, a lot of time and a long process of trial and error. |
Experiments in Computer Go Endgames (1996)
Recently, the mathematical theory of games has been applied to late-stage Go endgames. Based upon this theory, we developed a tool to solve local Go endgames. |
Generation of Patterns With External Conditions for the Game of Go (1999)
Patterns databases are used to improve search in games. We have generated pattern databases for the game of Go. |
Shared concepts between complex systems and the game of Go
We developed complex systems playing the game of Go using specific concepts and methods. Due to the inherent complexity and to the riches of the game of Go, some concepts can be shared with other complex domains. |
Self Fuzzy Learning
This paper explains a method to learn from fuzzy explanations. This method has been applied to learn strategic rules in the Game of Go. Explanation Based Learning uses a representation of knowledge mainly based on the predicate logic. My goal is to extend this method of learning to systems using fuzzy logic. |
Applying Retrograde Analysis to Nine Men's Morris (1990)
This paper presents results gained through retrograde analysis of Nine Men's Morris. |
Go patterns generated by retrograde analysis
Introduction Retrograde analysis is a qualification and generation technique of positions in two-player, complete information games. |
Speedup Mechanisms for Large Learning Systems (1998)
Eliminating combinatorics from the match in production systems is important for expert systems, real-time performance, machine learning, parallel implementation and cognitive modeling. We describe a way of managing the tradeoff between generality and efficiency in knowledge representation for large learning systems. We propose an architecture that enables to combine efficiency in problem solving to generality in learning. |
Eyespace Values in Go (1996)
Most of the application of combinatorial game theory to Go has been focussed on late endgame situations and scoring. However, it is also possible to apply it to any other aspect of the game that involves counting. |
On Meta-Game Approaches to Computer Go
Meta-game approaches are studied through the design of the Go playing program INDIGO. |
Generalized Thermography: Algorithms, Implementation, and Application to Go Endgames (1996)
Thermography [1] is a powerful method for analyzing combinatorial games. It has been extended to games that contain loops in their game graph by Berlekamp [2]. We survey the main ideas of this method and discuss how it applies to Go endgames. |
Proof-Number Search and Transpositions (1993)
|
Management of Uncertainty in Combinatorial Game Theory (1996)
We describe a method to handle uncertainty in combinatorial game theory. This method has been implemented in Gogol, a general game playing system mainly applied to the game of Go. In complex games such as Go, a search intensive approach is intractable. Gogol uses a lot of knowledge and little search. |
Game Theories and Computer Go (1993)
Two kinds of theories have traditionally influenced computer game playing: Classical game theory, with its minimax principle, and specialized theories developed for a particular game. In addition to these, combinatorial game theory promises to be useful for computer Go. We propose a model for Go based on this theory and discuss our preliminary implementation. |
A Generalized Threats Search Algorithm (2002)
A new algorithm based on threat analysis is proposed. It can model existing related algorithms such as Lambda Search and Abstract Proof Search. It solves 6x6 AtariGo much faster than previous algorithms. It can be used in other games. |
Program Generation for Firms Simulations in Competitive Environments (1998)
In a competitive environment, the results of a firm strongly depend on the behaviors of the other firms. Therefore, the valuation of a firm must take into account its competitors. |
Review: Computer Go 1984 2000 (2000)
Computer Go is maybe the biggest challenge faced by game programmers. |
There Are No Winning Moves Except the Last
This paper, we want to shed light on links between Computer Go and Uncertainty Management with fuzzy methods. The nature of this paper is prospective. It suggests an Uncertainty Management in INDIGO with numeric methods and fuzzy functions. |
Computer Go (2000)
Computer Go is maybe the biggest challenge faced by game programmers. |
Developments On Monte Carlo Go
We have developed two go programs, Olga and Oleg, using a Monte Carlo approach. |
Automatic Acquisition of Go Knowledge from Game Records: Deductive and Evolutionary Approaches (1998
A deductive approach and an evolutionary approach. While the deductive approach acquires strict knowledge, the evolutionary approach acquires heuristic knowledge. The deductive approach acquires patterns, which can be compiled into sequences of moves. The evolutionary approach acquires patterns and sequences of moves. |
A positional Judgment System for Computer Go
Computer Go offers researchers a new challenge and opens up a very wide scope of possibilities for artificial intelligence. In a computer Go program, the most important element is a positional judgment system. |
Automatic Ordering of Predicates by Metarules (1996)
I describe metarules which order predicates contained in first order logic rules. These metarules are applied to rules created by a learning Go program. After the ordering of the predicates, the rules are matched orders of magnitude faster than the original learned rules. |
Toward Spatial Reasoning About Natural Objects
This paper addresses Spatial Reasoning in the game of Go that is a domain where the objects look like "natural". |
The Economist's View of Combinatorial Games (1996)
We introduce two equivalent methodologies for defining and computing a position's mean (value of playing Black rather than White) and temperature (value of next move). Both methodologies apply in more generality than the classical one. |
Mate in 38: applying proof-number search to chess
Proof-number search (pn-search) has shown its merit in contributing to the solution of Connect-Four, Qubic and Go-Moku. In this contribution we show that pn-search is a highly capable searcher for mates in chess. |
Integration of Different Reasoning Modes in a Go Playing and Learning System (1998)
Integrating multiple reasoning mode is useful in complex domains like the game of Go. Go players use various forms of reasoning during a game. Reasoning at the tactical level is completely different from reasoning at the strategic level. Choosing a plan requires a different form of reasoning than knowing how to execute a plan. |
Why did TD-Gammon Work? (1997)
Although TD-Gammon is one of the major successes in machine learning, it has not led to similar impressive breakthroughs in temporal difference learning for other applications or even other games. |
Loopy Games and Go (1996)
Berlekamp, Conway and Guy have developed a theory of partizan loopy combinatorial games---that is, partizan combinatorial games that allow infinite play---under disjunctive composition. We review this theory of loopy games and show how it can be adapted to the two-person strategy game of Go, which also has the feature that situations involving infinitely long play often arise. |
A Learning Architecture For The Game Of Go (2001)
In this paper, a three-component architecture of a learning environment for Go is sketched, which can be applied to any two-player, deterministic, full information, partizan, combinatorial game. |
Learning to Forecast by Explaining the Consequences of Actions (1996)
I explain a method to learn to achieve goals in games. In very complex games, such as the game of Go, a search intensive approach is intractable. |
Generating Search Knowledge in a Class of Games (2000)
We present the Introspect system that generates search knowledge for different games. |
Computer Go: an AI Oriented Survey (2001)
Since the beginning of AI, mind games have been studied as relevant application fields. Nowadays, some programs are better than human players in most classical games. Their results highlight the efficiency of AI methods that are now quite standard. Such methods are very useful to Go programs, but they do not enable a strong Go program to be built. The problems related to Computer Go require new AI problem solving methods. |
Complex Games in Practice (1999)
This paper highlights an important issue linked to the global move decision process in a Go playing program, in other terms, how to describe, determine and use group strength in a simple way so as to select the move to play. Our present study is based on our experience in developing Indigo, a current Go program that ranked tenth at the last Ing Cup in London in November 1998. |
A New Computational Approach to The Game of Go
This paper investigates the application of neural network techniques to the creation of a program that can play the game of Go with some degree of success. |
Spatial Reasoning in the game of Go
This paper addresses Spatial Reasoning in the game of Go. Combinatorial complexity of the game of Go obliges the Computer Go community to study spatial representations of human players that are complex. These representations contain the concepts of grouping and fractioning. |
Retrograde Analysis Of Patterns Versus Metaprogramming (2001)
|
Local move prediction in Go
The paper presents a system that learns to predict local strong expert moves in the game of Go at a level comparable to that of strong human kyu players. |
Searches, tree pruning and tree ordering in Go
|
Metaprogramming Forced Moves (1998)
Knowledge about forced moves enables to select a small number of moves from the set of possible moves. It is very important in complex domains where search trees have a large branching factor. Knowing forced moves drastically cuts the search trees. |
Proof Planning Methods as Schemas (1999)
A major problem in automated theorem proving is search control. |
GRT: A Domain Independent Heuristic for STRIPS Worlds based on Greedy Regression Tables (1999)
This paper presents Greedy Regression Tables (GRT), a new domain independent heuristic for STRIPS worlds. |
Why Computer Go Is Hard (1992)
|
Benson Algorithm
Benson's algorithm is a rigorous and static (i.e. no search is needed) method for recognizing stones that are uncapturable even if the attacker is allowed to play an infinite number of times in a row (i.e. the defender always passes). |
Knowledge Representation in The Many Faces of Go
Interesante nota de David Fotland |
