![\"Logo](\"https:\/\/humanas.blog.scielo.org\/wp-content\/uploads\/2021\/10\/trans_glogo.gif\" "\\"Logo")
<\/a><\/p>\nAr\u00e3o in his article titled <\/span>Behind Artificial Intelligence: An Analysis of the Epistemological Foundations of Machine Learning<\/span><\/a> (2024), states that “[…] it will be <\/span>demonstrated <\/span><\/i>how <\/span>induction and mathematization <\/span><\/i>work as the epistemological basis of artificial intelligence and how some of the <\/span>limitations <\/span><\/i>of this technology can be explained through the weaknesses of the <\/span>methods <\/span><\/i>that support it” (Ar\u00e3o, 2024, p. 01, italics mine).\u00a0<\/span><\/p>\n Nevertheless, the author is <\/span>not<\/span><\/i>, <\/span>strictly speaking<\/span><\/i>, presenting a <\/span>demonstration <\/span><\/i>in the said article, as we should, for this purpose, consider both a <\/span>logical-mathematical <\/span><\/i>analysis and a methodological <\/span>refutation <\/span><\/i>from the perspective of computer science, as emphasized in Sarmento (2004). In principle, the noted “weaknesses” are subject to correction if we <\/span>a priori <\/span><\/i>admit the construction of a maximally consistent system as the base of knowledge for the <\/span>learning machines <\/span><\/i>model.\u00a0<\/span><\/p>\n These observations are presented in the <\/span>Commentary on “Behind Artificial Intelligence: An Analysis of the Epistemological Foundations of Machine Learning”<\/span><\/a>, recently published in the journal <\/span>Trans\/Form\/A\u00e7\u00e3o<\/span><\/i>, whose reflections will be summarized here.<\/span><\/p>\n In an article titled “Artificial Intelligence, Undecidability, and Trans-computability” (see Sarmento, 2004) I proposed a critical approach to the <\/span>logical-epistemic<\/span><\/i> foundations of Artificial Intelligence, in which some formal and technical aspects \u2013 that instantiate and delineate the field of investigations encompassing Artificial Intelligence \u2013 were addressed.<\/span><\/p>\n Without proposing a technical and formal discussion in this essay, I present, in general terms, some points directed towards the issue of using an inductive-probabilistic logic for the “Large Language Models” (LLMs) built based on “neural networks,” which, in turn, enable the machine learning process.<\/span><\/p>\n Firstly, some of the areas of computer science directly related to the development of LLMs are <\/span>computational linguistics<\/span><\/i>, <\/span>computational neuroscience<\/span><\/i> and <\/span>quantum computing<\/span><\/i> and, especially, the <\/span>algorithmic theory<\/span><\/i> of information. In all the highlighted areas, algorithmic methods of statistical inference are widely used.\u00a0<\/span><\/p>\n One of the subfields of the algorithmic theory of information is called Kolmogorov <\/span>complexity<\/span><\/i>, which deals with essential concepts for statistical-probabilistic methods of inference. This field of study involves a rigorous treatment of the concept of randomness, utilizing algorithmic methods such as, for example, the “Solomonoff probability” algorithm.<\/span><\/p>\n Another specific method for machine learning systems, via neural networks, is the use of the so-called Bayesian Logic. In general, there are various algorithmic methodological techniques that allow for a rigorous (and subject to correction) computational treatment of probabilistic-statistical approaches applicable to the development of AI systems.<\/span><\/p>\n Imagem: Freepik.<\/p><\/div>\n From a philosophical perspective, the theory of knowledge compatible with the development of these AI systems is the <\/span>reductionist epistemic<\/span><\/i> approach proposed by Lehrer (1990), in which the concept of knowledge is strictly reduced to the concept of information. Lehrer admits that the explanation of knowledge, in the sense of <\/span>information<\/span><\/i>, assumes that:<\/span><\/p>\n Here, “P” and “s” are variables, “P” designates any individual, and “s” denotes any information that can be obtained. Thus, “P knows that s” expresses the “recognition of the correctness of the information.” The justification (or corroboration) of the information is the central aspect of the epistemic analysis, given that this procedure should, in principle, be understood as a “measure” between “strong reasonableness and not complete certainty,” as Lehrer argues.<\/span><\/p>\n This means that justification, from the perspective of inductive-probabilistic logic, is bounded between reasonableness and complete certainty (the concept of “certainty” here means strong evidence of a probative nature, given the sufficient conditions at a certain time \u201ct\u201d, for the confirmation of the correctness of the information). In probabilistic terms, the closer the correctness of the information is to 1, the higher the degree of certainty of its confirmation.<\/span><\/p>\n The algorithmic theory of information (a subfield of information theory), which lies at the intersection of Turing’s computability theory and Shannon’s classical information theory, could serve as a <\/span>computational foundation<\/span><\/i> for the approach advocated by Lehrer, when it comes to the concept of information and, consequently, justify that LLMs process (and generate) knowledge about the “physical world” in a manner analogous to humans.<\/span><\/p>\n The main philosophical issue, in this case, shifts to the level of autonomy that AI cybernetic systems could have to act independently of human intervention. This is a question that undoubtedly requires deeper reflection.<\/span><\/p>\n Up to the present moment, however, these cybernetic systems do not effectively exhibit a capacity for self-correction and decision-making in an interactive and autonomous manner.<\/span><\/p>\n Those who advocate for such an approach, i.e., that AI cybernetic systems can generate knowledge, cannot ignore that AI will always be limited \u2014 with respect to its cognitive capacity \u2014 due to <\/span>Chaitin’s incompleteness<\/span><\/i> theorem (a version of G\u00f6del’s incompleteness metatheorem). And, consequently, AI cannot surpass the cognitive capacity of humans.<\/span><\/p>\n He has a doctorate in philosophy from Unicamp, SP, and is an Associate Professor in the Department of Philosophy at the Federal University of Para\u00edba (UFPB), Jo\u00e3o Pessoa, PB \u2013 Brazil. Research Areas: Mathematical Logic and Philosophy of Mathematics.<\/span><\/p>\n AR\u00c3O, C. Behind artificial intelligence: an analysis of the epistemological bases of machine learning. <\/span>Trans\/Form\/A\u00e7\u00e3o<\/span><\/i> [online]. 2024, vol. 47, n. 3, e02400163 [viewed 15 August 2024]. <\/span>https:\/\/doi.org\/10.1590\/0101-3173.2024.v47.n3.e02400163<\/span><\/a>. Available from: <\/span>https:\/\/www.scielo.br\/j\/trans\/a\/wKP3thTz35fmhG9pnmXNdMj\/<\/span><\/a><\/p>\n LEHRER, K. <\/span>Theory of Knowledge<\/span><\/i>. New York. Westview, 1990.<\/span><\/p>\n SARMENTO, G. Intelig\u00eancia artificial, indecidibilidade e transcomputabilidade. <\/span>Perspectiva Filos\u00f3fica<\/span><\/i>. 2004, vol. I, no. 21, pp. 179-209.<\/span><\/p>\n SARMENTO, G. Coment\u00e1rio a \u201cPor tr\u00e1s da intelig\u00eancia artificial: uma an\u00e1lise das bases epistemol\u00f3gicas do aprendizado de m\u00e1quina\u201d. <\/span>Trans\/Form\/A\u00e7\u00e3o<\/span><\/i> [online]. 2024, vol. 47, no. 3, e02400221 [viewed 15 August 2024]. <\/span>https:\/\/doi.org\/10.1590\/0101-3173.2024.v47.n3.e02400221<\/span><\/a>. Available from: <\/span>https:\/\/www.scielo.br\/j\/trans\/a\/Zk3GNmcnCHVmCkDRdWvXcRj\/<\/span><\/a>\u00a0<\/span><\/p>\n Trans\/Form\/A\u00e7\u00e3o \u2013 TRANS: https:\/\/www.scielo.br\/trans\/<\/a><\/p>\n Trans\/Form\/A\u00e7\u00e3o \u2013 Journal: Instagram<\/a> | Facebook<\/a> | Academia.edu<\/a><\/p>\n Garibaldi Sarmento \u2013 ORCID: <\/span>https:\/\/orcid.org\/0000-0002-9242-3945<\/span><\/a><\/p>\n","protected":false},"excerpt":{"rendered":" Outlined here is a critical review of the logical-epistemic foundations of machine learning, focusing on the limitation of AI systems’ autonomy in generating knowledge. It contrasts this possibility with the theoretical constraints posed by Chaitin’s incompleteness theorem, which argues that AI cannot surpass human cognitive capacity. …<\/span> Read More →<\/span><\/a><\/span><\/p>\n","protected":false},"author":664,"featured_media":1428,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"footnotes":"","_links_to":"","_links_to_target":""},"categories":[7,110],"tags":[146,111],"class_list":["post-1427","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-press-release","category-trans","tag-philosophy","tag-trans-form-acao"],"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/posts\/1427","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/users\/664"}],"replies":[{"embeddable":true,"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/comments?post=1427"}],"version-history":[{"count":4,"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/posts\/1427\/revisions"}],"predecessor-version":[{"id":1434,"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/posts\/1427\/revisions\/1434"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/media\/1428"}],"wp:attachment":[{"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/media?parent=1427"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/categories?post=1427"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/humanas.blog.scielo.org\/en\/wp-json\/wp\/v2\/tags?post=1427"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}<\/a>\n
About Garibaldi Sarmento<\/h3>\n
References<\/h3>\n
To read the article, access<\/h3>\n
External links<\/h3>\n