Barkochba Or 20Q (Twenty Questions) – the Game Against the Magic of Words Laszlo Pitlik, Laszlo Pitlik (Jun), Matyas Pitlik, Marcell Pitlik (MY-X Team)
Total Page:16
File Type:pdf, Size:1020Kb
Barkochba or 20Q (Twenty Questions) – The game against the magic of words Laszlo Pitlik, Laszlo Pitlik (jun), Matyas Pitlik, Marcell Pitlik (MY-X team) Abstract: This article demonstrates a similarity-based rule system for evaluating 20Q-games. 20Q- games and taxonomies (like plant and/or animal taxonomies) are useful supporters in order to be capable of explaining what is an expert system, why a single abstraction (like human words) can not be defined with an arbitrary exactness, why an appropriate definition has to have two layers: one for a direct description and an other one about exclusions of parallel terms/objects (like in the taxonomies). The KNUTH’s principle expects that each real knowledge-like phenomenon should be transformed into source code. Expert systems are the lowest level of structured descriptions being prepared for deriving words like in a 20Q-game. Keywords: taxonomy, expert system, similarity, evaluation rule set Introduction This paper is the newest part of the series about experiences of the QuILT-based education processes. Previous articles can be downloaded here: • https://miau.my-x.hu/miau/quilt/Definitions_of_knowledge.docx + annexes like: o https://miau.my-x.hu/miau/quilt/demo_questions_to_important_messages.docx o https://miau.my-x.hu/mediawiki/index.php/QuILT-IK045-Diary o https://miau.my-x.hu/mediawiki/index.php/Vita:QuILT-IK045-Diary o https://miau.my-x.hu/mediawiki/index.php/QuILT-IK059-Diary o https://miau.my-x.hu/mediawiki/index.php/Vita:QuILT-IK059-Diary • https://miau.my-x.hu/miau/quilt/reality_driven_education.docx + annexes like: o https://miau.my-x.hu/miau/quilt/chained-translations-legal-slang.docx o https://miau.my-x.hu/miau/quilt/demo_chained_translations.docx o https://miau.my-x.hu/miau/quilt/demos_chained_translations.docx o https://miau.my-x.hu/miau/quilt/forum_details.docx o https://miau.my-x.hu/mediawiki/index.php/QuILT-IK057-Diary o https://miau.my-x.hu/mediawiki/index.php/Vita:QuILT-IK057-Diary • https://miau.my-x.hu/miau/quilt/Exercises_for_critical_thinking_and_doing.docx • https://miau.my-x.hu/miau/quilt/st1_all.docx • (https://miau.my-x.hu/miau/quilt/20Q.docx) The running experimental education (where the experimental characteristics belongs to the international frames and not for the methodology as such) tries to support the shifting of paradigms from the classic teaching (e.g. based on the magic of words) towards to AI-based approaches where the KNUTH’s principle can be realized in an objective way. The first attempts about the definition of the word of knowledge made possible to derive that neither a good definition could be created, nor the huge volume of parallel definition could be ranked – because no rule system for evaluation of the available definitions could be created. Parallel, there are a lot of log-data about the Moodle-activities of the Students and there is a valid question too: Who is the best Student? Yet, there are no valid rule systems to evaluate Student’s activities. In ideal case, a valid rule system would be created by Students so, that each Student would suggest at least one variable being objective measurable about the activities of Students. This question about the best Student should always have a solution – in the most cases a subjective one (c.f. general scoring systems in the schools – without the using of the anti-discriminative principle: Each Student can have the same evaluation value based on a lot of performance variables (like subjects/competences). As it can be seen in a clear way: the basic problem is (in both cases like log-data and/or set of potential definitions) the lack of the objective/valid rule systems because this makes impossible to handle problems in a target-oriented/objectivized way. About the 20Q-games in general In case of the question “What is the number between 1-100 being chosen in a randomized way?”, the mathematical interpretation delivers a clear answer: https://math.stackexchange.com/questions/512994/guessing-a-number-between-1-and-100-in-7- guesses-or-less - The maximal number of the questions is 7. If somebody derive the numbers of the necessary questions based on the binary approximation for 7 digits with 0 or 1 for each guessable number between 1-100 and these numbers will be added, then we have a lowest sum of questions needed for each guessing from 1 to 100 (step=1). Assumed, that this is a real optimum (=minimum of the sum of the needed questions), then we have a chance to define the mathematics compared e.g. to the social sciences and/or natural/physical sciences. We have already a definition about science/knowledge from KNUTH. A further definition for parts of the science as such can be created as follows: • A problem is a mathematical problem, if the best solution(s) can be derived based on one single evaluation scale (like sum of each necessary questions). • If we could define more than one ranking scale for evaluation of the alternative solutions, then we might not speak about mathematics. In case of the guessing game above-described we could search for the best solution for human beings where not only the total sum of the needed question is a variable being capable of describing expected situation, but it would also be possible to say: the more a strategy leads to wins the better is the strategy. The above- outlined binary strategy is a kind of symmetrical strategy. However, we could create asymmetrical strategies too: e.g. 1. question: Has the searched number just 1 digit? And if we have an answer “YES”, then we will win compared to the binary strategy in case of the lowest values. Similar questions delivering useful asymmetries are e.g. Is the value a prime number? or Contains the value a zero? or each kind of the partibility can lead to the expected effects in the first (or further) level(s). It is important each not pure binary strategy should be defined in advance. Strategies can have mixed characteristics – it means they can combine binary and not binary (symmetrical and asymmetrical) questions. Parallel to the number/ratio of win- positions (compared to the binary strategy), it is also possible to derive the differences between the number of questions for binary and not-binary strategies. The partial rule would be the more is the difference the better is the strategy. We can probably define a lot of further evaluation variables. So, we have a multilayered evaluation problem needing an aggregated evaluation scale for deriving the best strategies. The natural and social sciences can be classified just here and now: o If the multilayered variable come from human beings (it means the variables/words are described in a high abstraction level like knowledge, competence, etc.), then we should speak about social sciences. o If the variables can be measured (like temperature, pressure, etc.), then we should speak about natural sciences. o (If a variable set has both types of the variables, then we can speak about interdisciplinarity.) o (If the measurable variables use terms belonging to human beings like price, income, etc., then we could speak about economics what means economics could belong rather to the natural sciences, than to the social sciences – or even it is a kind of multidisciplinary set of problems.) Relevant question: Are the Wikipedia “definitions” (especially compared to each other) more exact/precise than the above-derived definition based on one single characteristic concerning the used terminology (the type of variables/words/phenomena) or the magic of words can be observed in the single definitions in Wikipedia compared to the above-outlined definition-system where each definition has two layers: a direct one (describing the word as such) and an indirect one (excluding each other parallel terms): • https://en.wikipedia.org/wiki/Social_science (incl. economics) • https://en.wikipedia.org/wiki/Natural_science The 20Q-game can be seen as a kind of generalized version of the guessing game for numbers where the number of the potential solutions can be unlimited. In a limited case of words (see partial taxonomies) it is relevant to ensure, that each question makes the number of the potential solutions step by step fewer. If the number of the potential options in case of a taxonomy in each level is the same value (c.f. Kingdom / Division / Class / Subclass / Series (Order) / Family / Genus / Species / Varieties), and this value would be the 7, then we could speak about a numeral system of the base of 7. The taxonomies can not ensure the above-mentioned sameness. The taxonomies are a kind of asymmetric question systems compared to the above-outlined binary question system as mathematical optimum for searching numbers. Relevant remark: The 20Q-game needs question being answerable with yes or no, but this expectation can not enforce to handle in a binary way: e.g. Is the searched animal a vertebrate animal? (yes/no). BUT: we do not know about relationships to the other options in the DIVISION/PHYLUM level (c.f. https://www.toppr.com/guides/biology/diversity-in-living- organisms/animal-kingdom/). Conclusion: The taxonomy experts could not create a binary (question) system for centuries – not a single asymmetrical one. But a kind of approximately symmetric approach could be created in case of the taxonomies tending to be a numeral system on the base of higher than 2: If somebody knows how many objects are totally available, then on each level (digit), an estimation can be made e.g. what kind of divisions (phylum) together covers ca. the one half of the objects? This aggregated way can be used level by level in the taxonomy.