Derived metrics for the game of Go – intrinsic network strength assessment and cheat-detection

Egri-Nagy, Attila; Törmänen, Antti

doi:10.1109/candar51075.2020.00010

Cited by 7 publications

(7 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One of them from Egri-Nagy and Törmänen [1] tries to detect AI-assisted play with a single SGF file. We share some common hypothesis with their work:…”

Section: Related Workmentioning

confidence: 99%

A Statistical Analysis of Amateur Go players to Assist AI-cheating Detection

Barollet,

Le Duc

2023

JGS

View full text Add to dashboard Cite

Since the democratization of powerful AI engines for the game of Go, it is not uncommon to see a drastic level increase of some players that must be explained with the help of AI. This is considered cheating and forbidden by most organizations. When looking at online beginners and stronger amateur players, we discovered that they can display playing strength below professional level and still confidently win the game, as opposed to professional players. This makes using only AI-likeness metrics not sufficient to detect such players. We propose a method based on the analysis of a player’s performance considering point loss distribution over several games, taking into account only relevant moves of a game. We still use an AI-likeness metric for analyzing individual games where the use of AI may not be consistent. We evaluated our methods on two European go official online leagues, where cheating detection was already performed (for a total of about 150 unique regular players, with levels ranging from 20 kyu to 5 European dan). We show that our system confirmed 5 cases of players previously banned for cheating (out of 6). Our methods do not set out to categorize players between “cheaters” and “not cheaters,” but rather rank them in order of suspicion, for the sake of assisting referees and providing them a way to effectively investigate suspicious players over time.

show abstract

“…One of them from Egri-Nagy and Törmänen [1] tries to detect AI-assisted play with a single SGF file. We share some common hypothesis with their work:…”

Section: Related Workmentioning

confidence: 99%

A Statistical Analysis of Amateur Go players to Assist AI-cheating Detection

Barollet,

Le Duc

2023

JGS

View full text Add to dashboard Cite

show abstract

“…The effect of a move is the difference between the score lead estimate before and after the move [10]. This describes the efficiency of a move compared to the AI's best move candidate.…”

Section: Defining the Cost Of Passingmentioning

confidence: 99%

“…In the middle game, values are generally between 12 and 7 points, and values below 7 points usually indicate the endgame. These values are estimates based on dozens of cost-of-passing analyses by the second author relying on his professional player experience and the work done in cheat detection [10]. In the future it may be useful to conduct a statistical research on the values by a bulk analysis of a large number of games.…”

Section: The Stages Of the Game Characterized Numericallymentioning

confidence: 99%

See 1 more Smart Citation

The cost of passing - using deep learning AIs to expand our understanding of the ancient game of Go

Egri-Nagy

Törmänen²

2023

IJNC

Self Cite

View full text Add to dashboard Cite

In addition to their playing skills, AI engines utilizing deep learning neural networks provide excellent tools for analyzing traditional board games if we define new measures based on their raw output. For the ancient game of Go, we develop a numerical tool for context-sensitive move-by-move performance evaluation and for automating the recognition of game features. We measure the urgency of a move by the cost of passing, which is the score value difference between the current configuration of stones and after a hypothetical pass in the same board position. In this paper, we investigate the properties of this measure and describe some applications for developing learning tools and analyzing a large number of games. As we use AI tools to gain new insights into the ancient game of Go and develop a more precise, quantified understanding, this work fits into the more significant and general project of utilizing superhuman AI engines for deepening human understanding and growing human knowledge.

show abstract

“…In the following years, commercial and open-source programs – created in AlphaGo’s likeness – became widely available to professional and amateur Go players. The sacred Go machine gave birth to powerful yet mundane tools for training and analysis, which simultaneously allowed for novel ways of cheating, especially in online games: ‘Using an AI engine for finding best moves and variations in a game is called analysis after the game is finished; cheating when the game is still ongoing’ (Egri-Nagy and Törmänen, 2020a: 12). While players at professional online tournaments – which proliferated in times of pandemic – are heavily monitored, the practice of cheating remains a problem in amateur games and tournaments.…”

Section: The Sacred Go Machine and The Disenchantment Of The Go Worldmentioning

confidence: 99%

Technology as (Dis-)Enchantment. AlphaGo and the Meaning-Making of Artificial Intelligence

Binder

2022

Cultural Sociology

View full text Add to dashboard Cite

As a social phenomenon, artificial intelligence (AI) is not just technically but also culturally constructed. This article investigates the meaning-making of AI in the case of AlphaGo by employing and refining cultural sociological narrative analysis. Building on Smith’s structural model of genre, whose horizontal axis reflects varying degrees of (dis-)enchantment, I propose an extended model of narrative genre, adding a vertical axis on the theoretical basis of Durkheim’s distinction between pure and impure sacred, to account for the empirical bifurcation between utopian and dystopian AI narratives. While critical approaches to AI, prevalent in sociology, tend to offer disenchanted narratives, my cultural sociological approach allows for the construction of a meta-narrative, which is able to capture not only enchantment as well as disenchantment but also purification and impurification as empirical processes that accompany the emergence and consolidation of new technologies. This approach is exemplified by a case study of AlphaGo, a Go-playing program utilizing machine learning and neural networks, which gained global prominence and cultural significance after beating a human grandmaster in 2016. Drawing on publicly available online data, this article investigates the discourses surrounding AlphaGo, focusing on its cultural construction through storytelling and genre. I not only show how characters and events were emplotted in different stories, which were in turn embedded in broader narratives about technological progress and AI, but also explain how the development of the main storyline was driven by in-game performances, audience expectations and collective representations. The article demonstrates the feasibility of a cultural sociology of AI and the usefulness of my extended model of narrative genre, which is not only applicable to AI discourses but other domains as well.

show abstract

Derived metrics for the game of Go – intrinsic network strength assessment and cheat-detection

Cited by 7 publications

References 9 publications

A Statistical Analysis of Amateur Go players to Assist AI-cheating Detection

A Statistical Analysis of Amateur Go players to Assist AI-cheating Detection

The cost of passing - using deep learning AIs to expand our understanding of the ancient game of Go

Technology as (Dis-)Enchantment. AlphaGo and the Meaning-Making of Artificial Intelligence

Contact Info

Product

Resources

About