(for the 16-Jul-2021 opening of the ES&AG online gallery)
By Eric Shi
Welcome to the ES&AG AI Art Studio --- a virtual art gallery featuring artificial intelligence (AI) paintings. This virtual gallery is operated by the new 3 musketeers --- ES (short for Eric S. Shi, the human artist) and his 2 virtual AI-bot assistants: ESAG (short for Eric Shi Art Generator) and ESNA (short for Eric Shi Neo Analyzer). Both ESAG and ESNA are algorithms living in virtual space. ESAG was born on 16-Jan.-2021, and ESNA on 16-Mar.-2021. They are both virtual neural network (VNN) descendants of ES.
ES and ESAG are jointly responsible for the creations of all the AI paintings on exhibition at this online gallery. ES and ESAG are continuing creating. So, if you are interested in observing how an AI virtual-bot (i.e., ESAG in this case) can expand our knowledge base, especially in terms of expanding our horizon of artistic imaginations, do expect frequent updates of the artworks on exhibition here and do keep on coming back.
ES and ESNA are jointly responsible for writing articles for the Blog section of this gallery, and managing exhibition programs as well as sales promotions. So, if you are interested in observing how an AI virtual-bot (i.e., ESNA in this case) can handle languages and write articles, observing how ESNA’s writing styles evolve, or if you own or want to own a small business, and are interested in finding out whether and how AI can manage the sales and marketing programs for a small shop using small data (correct, not using big data since they are typically not available to small businesses anyways), do keep on coming back and visit our Blog section.
The evolution history of our human race, leaves deep marks on and sets limitations to the thinking patterns of our human brains. Our thought process is fundamentally empirical. We know what we see and hear. On our planet, most of the things are symmetrical. Animals have 2 eyes, 2 ears, 2 cheeks, 2 arms (e.g., primates), 2 or other even numbers of legs, and/or even numbers of wings. So, the knowledge passed on from generations to generations is left-right symmetry (mathematically, this is called mirror symmetry or m-symmetry) is rated by our human brains as real, verifiable and/or confirmable by daily experiences, and therefore rated as true, and then further rated as correct, and finally m-symmetric becomes aesthetically beautiful.
It is not surprising, when it comes to imagining something(s) that we have never experienced on this planet (e.g., the extraterrestrials), we tend to go back to our knowledge base and project what we know onto the unknowns. For instance, we tend to imagine aliens as having 2 eyes, 2 ears, 2 cheeks, 2 arms (or other even number of arms), 2 legs (or other even number of legs) and/or 2 wings, arranged in the m-symmetry just like ourselves.
Consequently, our imaginations and curiosity are typically shallow. They are easily satisfied and ready to doze off. So long as aliens in movies have enlarged foreheads than us, and/or have 3 fingers instead of 5 in each hand, we are quickly sold that these are how aliens should look like. Not many of us would imagine aliens as having “fluid legs” or “rotating torso”, not because the “fluid legs” or “rotating torso” are scientifically unreasonable, but just because they are too “unrealistic” and therefore utterly “incorrect”.
Not many of us have noted how much our imagination has been hampered by such “reality checks” of our thought process, until the arrival of “AI supremacy” as signaled by land-sliding defeat of the brightest human minds by an AI algorithm (AlphaGo) in a series of board games. Not only the brightest among us were spectacularly defeated by an autonomous algorithm (self-learning), but also neither the brightest of us nor the whole collection of us (7 billion strong) can figure out why we failed, letting alone figuring out how to defeat the algorithm.
The German philosopher G. Hegel once famously said, “The only thing that we learn from history is that we learn nothing from history.”
If you view the triumph of AlphaGo over the brightest human Go-players as a critical event in the human history, what should we learn from it? Well, AlphaGo did not limit its imagination; we human did. The human players improve their Go-play skills and/or strategy by studying and memorizing how their past masters / ancestors played the game over thousands of years. Consequently, the human thought processes are heavily skewed along the direction their ancestors. AlphaGo has the advantage of giving fair considerations to vast number of other options (i.e., along different thinking directions). Not surprisingly, certain moves made by AlphaGo were complete incomprehensible by its human opponents. This is because such moves were not in human’s knowledge base, the human opponents had never studied such possibility, they had never thought of such a need.
By the same token, AlphaZero defeated AlphaGo by 100:0, primarily because AlphaZero had eliminated the learning from human masters from its training process and AlphaGo had not. Lesson? Be careful when you have a little knowledge, they tend to make your thought process skewed and can become a burden for your further improvement.
I am not advocating, all of sudden, to abandon all knowledge accumulated by the human race in past thousands of years. The lesson that I have learned is to keep my mind open and stay humble. Taking the m-symmetry as an example, we should not, all of sudden, declare that all things with m-symmetry are ugly. Instead, we should ask ourselves in addition to the m-symmetry, could there be other symmetries (e.g., higher-order symmetry such as C3-, C4-, C6-, Td- and Oh-symmetries), and/or could the lack of symmetry, be also beautiful?
For instance, in the aesthetic world, how can we be certain that a face with 2 eyes is necessarily more beautiful than a face with 1 eye or 5 eyes? How would we know if we do not experiment and give all these possibilities fair (unbiased) considerations? How would we know something is (or is not) beautiful if it lays outside of the boundaries of our traditional aesthetic value system (now that we know such a system is shallow, narrow, and skewed)?
We humans have been for a long time nebulously suspecting the existence of advanced extra-terrestrials (as beautifully summarized by the Fermi paradox). This is good. It is good for us to suspect that we are not the best in the universe. But is it really necessary for us to wait until an advanced extra-terrestrial civilization coming to show us how imperfect we are? (Just like why do we have to wait until AlphaGo coming to tell us how skewed and limited our thought process is?) Why can’t we start to figure these out by ourselves?
So, at the ES&AG AI Art Studio, we designed our own experiments and embarked on our own journey. We strive to seek every opportunity to learn from computers (something that we now know capable of think differently and often better than human brains). We call this journey "AI-Learning". (It means human to learn from AI; instead of learning how to make an AI machine or how to train an AI machine to act like a human.)
We strive: (1) to learn how properly designed AI algorithms can augment and adjust the narrow / skewed view of human in making observations and evaluations (e.g., in the field of aesthetical appreciations); (2) to learn how an unbiased mind (i.e., properly designed AI algorithms, in this case) could evolve / evolute via, self-learning. We strive to achieve these objectives by analyzing the behavioral patterns of the specially designed AI algorithms, during their mutual interactions and during their interactions with humans.
Some of these efforts have proven fruitful since the birth of ESAG and the birth of ESNA. Some of our AI artworks (e.g., as exhibited in the Shop with product titles prefixed with letters API) can serve to demonstrate this point. For those who are not much into arts, please be patient. ESNA has some art commentaries in its pipeline. Do come to check our blogs from time to time.
These artworks, in addition to being of artistic values, can also serve as pictorial evidence to document evolvement of a completely novel history of humans using AI as a mirror iteratively and in various combinations in order to identify limitations of human thinking processes.
In addition, through the AI artworks and the AI products being created from this gallery, we also strive to gain insights (e.g., through analyses by ESNA) on how human brains behave in response to visual, audio, language, and social stimuli (e.g., the unfamiliar visual stimuli created by ESAG, and the language and/or social stimuli engineered by ESNA). Hopefully, through such iterative cycles of study-discover-create (the SDC-cycle), we will be able to find and quantify limitations of our human brains as well as to propose surgery-free methods to improve human-brain operations, to augment surgical implantations of computer chips.
(Drafted on 18-Jun-2021, finalized on 08-Jul-2021.)
Comments