Alexander Khomiakov
Philosopher, artificial intelligence researcher,
author of books on the philosophy of mind and epistemology, specialist in AI implementation.
author of books on the philosophy of mind and epistemology, specialist in AI implementation.
Invited specialist at the National Research University HSE in the philosophy of AI (2021-2023). Member of the AGI Russia community. Regular participant in the seminar on the philosophy of mathematics at Moscow State University (2020-2025).
Born in Russia in 1968. Studied physics (year of admission 1985) and philosophy (1998) at St. Petersburg State University. Studied philosophy, neurophysiology and developmental psychology since the early 2000s (with gratitude to V. Karpinskaya and N. Shemyakina). Published books on the philosophy of consciousness and epistemology. Completed an MBA course at the Open School of Business (St. Petersburg, 2004).
Kantian - developed the concept of interactive constructivism as a development of Kant's teachings, in which he resolves the issue of the connection between the "phenomenon-for-me" and the "thing-in-itself" (see "Prolegomena to All Knowledge", Chapter 2). Also developed a theory of the origin of consciousness based on the teachings of Russian neurophysiologists and psychologists. Influenced by: A. Luria, L. Vygotsky, P. Galperin.
Currently, I am the director of innovations for a large company, creating intelligent agents. I am a visionary in the use of AI in business with reports at specialized conferences.
Born in Russia in 1968. Studied physics (year of admission 1985) and philosophy (1998) at St. Petersburg State University. Studied philosophy, neurophysiology and developmental psychology since the early 2000s (with gratitude to V. Karpinskaya and N. Shemyakina). Published books on the philosophy of consciousness and epistemology. Completed an MBA course at the Open School of Business (St. Petersburg, 2004).
Kantian - developed the concept of interactive constructivism as a development of Kant's teachings, in which he resolves the issue of the connection between the "phenomenon-for-me" and the "thing-in-itself" (see "Prolegomena to All Knowledge", Chapter 2). Also developed a theory of the origin of consciousness based on the teachings of Russian neurophysiologists and psychologists. Influenced by: A. Luria, L. Vygotsky, P. Galperin.
Currently, I am the director of innovations for a large company, creating intelligent agents. I am a visionary in the use of AI in business with reports at specialized conferences.
Concepts under development
Three concepts have been developed that are closely related to each other
The author is an opponent of any realism in all its manifestations (objectivism). This epistemological position claims that one can know something about the implied reality objectively, that is, how it is beyond our ideas and without their mediation, which is a contradictory, logical position. In the philosophy of science, the author adheres to the position of constructivism, while the concept of interactive constructivism has been developed, which solves the problem of how our constructive ideas are consistent with the reality in which we act. This approach is illustrated by the thought experiment "Blind in the Room" presented in the author's latest book (Chapter 2).
A theory of functional consciousness, including the origin of consciousness, has been developed, which has a broad justification and experimental confirmation. Consciousness is interpreted as a separate process of recognizing one's actions during the implementation of directed activity (external and internal, including emotions). And the hypothesis of the emergence of subjectivity as a functional property as a result of the closure of the cycle - awareness of action, reaction in the form of action and new awareness. See the book "On the nature of consciousness and the problem of consciousness-body".
A new theory of intelligence has been developed, based on the theory of predictive recognition of changes in the environment as a basic function of the organism in the environment. An analogy of thinking and schemes as a combination of predicates is built from it. And the analogy and schemes serve as the basis for the forms of thinking, learning, generalization, etc. known to us, which is shown in the author's articles. The author experimentally confirmed the basic function of intelligence - analogy through general predicates, dynamic multivariate schemes.
Thoughts
And yet, LLMs don't think like that, as people at the Max Planck Institute have discovered
A very important discovery, in my opinion. Scientists have discovered a difference in the "third wheel" task between MLM and humans. And this difference points to a difference in the "intelligence" of humans and MLM. The multimodal model looks at superficial differences, the similarity of which it knows (not necessarily by appearance), and humans separate by other features that are predicates to the concepts of objects - by actions with them, their purpose (also actions), by hidden properties, everything that we call the meaning of an object, which do not clearly follow from their image. It is interesting that there is an article with seemingly the same experience, but opposite conclusions - this article also studied triplets for the third wheel in words and looked at their distribution in humans by semantic proximity and in the language model. And they found a similarity. But this is a similarity in the distribution of the surface structure. The experiment of the first and second articles is about different things - about the difference in understanding relationships and the distribution of words and objects. This paper supports my theory that intelligence is based on predicative relations rather than a simple probabilistic distribution of tokens.
My books
А. Хомяков, Пролегомены ко всякому знанию, Ridero 2025, Флинта, 2025
Prolegomena to all knowledge
Translation: Anton A. Borisov.
The book criticizes the paradigm of realism and presents as an alternative the concept of interactive constructivism. It answers the questions of what we can actually know about the world outside our ideas, how knowledge is created and constructed, how perception, thinking, and intellect work in general. The presentation goes from the basic functions of intellect such as predicativity and analogy to how entire models of situations and scientific theories are built. The book also presents the concept of the emergence of consciousness as the highest function of intellect, the role of consciousness in the evolution of society and reason, which is given a new definition. The presented concept of intellect allows us to take a new look at such problems as the consciousness-body problem, subjectivism versus objectivism, etc.
Download the book from Yandex-disk: English, Russian
Translation: Anton A. Borisov.
The book criticizes the paradigm of realism and presents as an alternative the concept of interactive constructivism. It answers the questions of what we can actually know about the world outside our ideas, how knowledge is created and constructed, how perception, thinking, and intellect work in general. The presentation goes from the basic functions of intellect such as predicativity and analogy to how entire models of situations and scientific theories are built. The book also presents the concept of the emergence of consciousness as the highest function of intellect, the role of consciousness in the evolution of society and reason, which is given a new definition. The presented concept of intellect allows us to take a new look at such problems as the consciousness-body problem, subjectivism versus objectivism, etc.
Download the book from Yandex-disk: English, Russian
А. Хомяков, О природе сознания и проблеме сознания тела, Флинта, 2020
On the nature of consciousness and the problem of body consciousness.
The monograph presents a hypothesis of the origin and function of consciousness, developed by the author based on research in cognitive psychology and phenomenology. Consciousness is distinguished as a separate phenomenon with functions different from the functions of other mental phenomena. The argumentation of the proposed hypothesis is presented and a definition of consciousness is given based on it, which allows us to formulate the development of the Turing test for identifying consciousness in artificial agents. Based on the presented hypothesis of consciousness, an explanation is given of the reason for the existence of the “consciousness – body” problem, as well as the insolubility of the problem of free will.
Download the book from Yandex-disk: Russian
The monograph presents a hypothesis of the origin and function of consciousness, developed by the author based on research in cognitive psychology and phenomenology. Consciousness is distinguished as a separate phenomenon with functions different from the functions of other mental phenomena. The argumentation of the proposed hypothesis is presented and a definition of consciousness is given based on it, which allows us to formulate the development of the Turing test for identifying consciousness in artificial agents. Based on the presented hypothesis of consciousness, an explanation is given of the reason for the existence of the “consciousness – body” problem, as well as the insolubility of the problem of free will.
Download the book from Yandex-disk: Russian
А. Хомяков, Высший разум, Ridero 2018
Higher intelligence.
Can there be intelligent beings on earth smarter than humans? Why can't we meet aliens? How can a person be transferred to a computer? What is consciousness? Why is our thinking so limited? The hero asks himself these and other questions, being one of those who managed to survive the terrible pandemic. The story was written before the pandemic (2020). In it, the author explores the issues of self-identification of the "I", the possibility of another mind, the social structure of humanity in the era of transition to artificial intelligence. The author does not claim the artistic value of the book, these are philosophical reflections in the form of a story.
Download the book from Yandex-disk:: Russian
Can there be intelligent beings on earth smarter than humans? Why can't we meet aliens? How can a person be transferred to a computer? What is consciousness? Why is our thinking so limited? The hero asks himself these and other questions, being one of those who managed to survive the terrible pandemic. The story was written before the pandemic (2020). In it, the author explores the issues of self-identification of the "I", the possibility of another mind, the social structure of humanity in the era of transition to artificial intelligence. The author does not claim the artistic value of the book, these are philosophical reflections in the form of a story.
Download the book from Yandex-disk:: Russian
Articles and research
Large language models in the form of chatbots very realistically imitate a dialogue as an omniscient interlocutor and therefore have become widespread. But even in the Google Gemini chatbot, they do not advise to trust what the chatbot will write and ask to check its answers. In this review, various types of LLM errors, such as the curse of inversion, number processing, etc., will be analyzed to identify their causes. Such an analysis led to the conclusion about the common causes of errors, which consist in the fact that transformers do not have deep analogy, abstraction and selectivity of the content taken into account in calculating the answer (inference). Show more.
The author conducts an analytical study, which puts forward a hypothesis about the role of analogy in thinking. The article presents criticism of the hypothesis of structural mapping in analogy as an explanation of the analogical argument and will offer an alternative explanation - the functional hypothesis of the analogical argument. It will also be shown how such functions of thinking as recognition and memory, metaphor and syllogism, generalization and ontology, deduction and induction can be realized with the help of the developed method of analogy. Show more
The article presents a new method for obtaining analogs of words, characterized by simplicity and the absence of the need for preliminary training on big data, as in existing methods. In the proposed method, analogs are determined by their syntactic predicates using the methods of distributive semantics. In the study, analogs of adjectives, nouns and verbs were obtained and analyzed. This made it possible to come to a result that is not inferior to the results obtained using the most popular neural network approach word2vec, in a qualitative comparison of analogs. The demonstrated method shows that obtaining analogs is possible using the methods of distributive semantics using a more interpretable method, which opens up the possibility of studying semantic analogy. Show more
Presentations at seminars and conferences
Participation in various seminars and conferences, which for me is the best way to present and discuss my ideas
Community AGI Russia
Discussion by A. Khomyakov, A. Kuznetsov and A. Yashin. Seminar on Philosophy of Mathematics, Moscow State University
Seminar on Philosophy of Mathematics at Moscow State University
Community AGI Russia. .
