Machine Intelligence

Machine Intelligence

first of all, i don’t like the term Artificial intelligence (AI), i think machine intelligence (MI) is better as intelligence can’t be artificial; in fact, i have a problem with MI as well; so, my choice would be machine networks (MN), but then no one would know what am i talking about.

second, lately people have been asking me this question .. do i use some form of Artificial intelligence (or machine intelligence (MI)) for my trading ..?

& my reply:  why would i use AI/MI/MN when i have an option to use real intelligence .. it is called human mind, and it’s so cheap that all it needs is spicy Indian food & meditation to run …  hiking to rejuvenate … & walking & driving to solve problems ..!  

like other buzzwords in technology used in past, it’s normal for people to associate everything with artificial intelligence now-a-days, so the consumers / corporate give attention to them and venture capitalists open their checkbooks.

in my opinion, most (not all) of the machine learning as it exists today (2018-2019) is simply smart programming (level 0, see below)

so, let me first differentiate between machine intelligence & machine learning: 

machine learning (or smart programming as i like to call) includes smart usage of algorithms, access to huge set of storage / data & generous processing power – likely in a cloud setup; e.g. a self-driving car uses a program for auto-drive feature; however, this feature is simply a program, whose code is developed by another program, which is developed based on the data captured over driving behavior of a human ..  & the auto-drive program uses GPS data, sensors data around the car and driver’s habits (recorded while driving) etc. and make real-time decisions to control the movements of a car  .. all of these are provided to a computer in some shape and form to which it can make sense.

Alexa helping you with your mood or Netflix making a movie suggestion is smart programming that uses huge amount of data, coupled with processing power of cloud, and NOT intelligence.

from here on, i define 4 additional levels of MI, as:

2) intelligence of a 16 year old human: it would be something like this, a machine is given a graduate level book of physics or economics (both my favorite subjects) .. the machine reads the book(s), solves all the questions and passes an exams like a graduate level student does .. that would be level 2

3) intelligence of a 25 years old genius: e.g. both Bill gates & Steve Jobs started their companies in mid-twenties, which became billion dollar enterprises in few years (& lately trillion dollars!) … so the day a machine can generate an idea which turns into a billion dollar enterprise would be a level 3.

4)  duplicate itself: one of the key function of humans is reproduction; so, the day, a machine can reproduce itself, that would be level 4 of MI .. this stage would also require dealing with emotions; as to how do the machine deals in the context of family, siblings etc. .. funny thing is that i already have the mathematical equations for them, which can be used in reverse to generate human emotions.

5) intelligence of a 50 year old smart human being (also human cognition):  back to Apple & Microsoft; both Tim Cook & Satya Nadela have crossed 50 and turned their companies into trillion dollar enterprises .. so that would be the ultimate level 5  .. an equivalent of this in political world would be to strategize and fight a war with another country with no loss of life on either side ..! 

i believe in MI and am neither discouraged nor afraid of it, because i hope humans will become way more smarter in future .. just like kids now-a-days learn how to play a video on Youtube or talk on phone in infancy & it took me more than 25 years to learn similar functions … but it will take a while for machines to go to stage 2 & beyond  ..!

Jan 2023: Open AI & Chat GPT

 Open  AI & Chat GPT being talk of the town and with fantastic reviews all  around, i decided to put it to test to see what does Chat GPT has to say  on my work; really i was scared to type my questions below as to what  if this tool already knows the solution that i am working on for last 10  years, and to my satisfaction and joy, it only knows what everyone else  does .. so my work is still sacred ..!

Thank God!!!

VIP: What’s Behavior finance ?

chat  GPT: Behavioral finance is a field of study that combines psychology  and economics to better understand why people make certain financial  decisions. It examines the psychological and emotional factors that  influence investors’ behavior, and how those factors can lead to  irrational decision making. Some key concepts in behavioral finance  include cognitive biases, such as the tendency to overestimate one’s own  abilities or to overestimate the predictability of financial markets,  as well as emotional biases, such as the tendency to make decisions  based on fear or greed. The goal of behavioral finance is to help  investors make more rational and effective decisions by understanding  and overcoming these biases. VIP: does behavior finance have any mathematical model?chat GPT: Behavioral  finance does not rely on a single mathematical model, but rather draws  from a variety of different models and theories from economics,  psychology, and other fields. Some of the mathematical models used in  behavioral finance include:

• Prospect  theory: A model that describes how people make decisions under  uncertainty, and how their preferences for gains and losses are affected  by the way the options are framed.
• Heuristics and  biases: A model that describes how people use simple rules of thumb, or  heuristics, to make decisions quickly and efficiently, but that these  heuristics can lead to systematic errors or biases in judgment.
• Mental  accounting: A model that describes how people categorize and evaluate  financial outcomes based on their source, rather than their absolute  value.
• Behavioral asset pricing models: A model that  describes how investors’ irrational behavior can affect stock prices and  returns, such as overreaction, under reaction and momentum.
• Neuro – finance: A model that uses neuro -imaging techniques to study the neural processes underlying financial decision making

Overall  Behavioral finance uses a combination of mathematical models and  empirical studies to understand the psychological and emotional factors  that influence financial decisions, to provide a more comprehensive  understanding of how markets operate.

VIP: is financial market behavior predictable? 

chat GPT: The  predictability of financial market behavior is a topic of ongoing  debate in the field of finance. Some theorists argue that markets are  efficient, meaning that all available information is already reflected  in current prices, and that it is impossible to consistently predict  future prices or returns. Others argue that markets are not fully  efficient and that there are patterns or anomalies that can be used to  make profitable predictions.

Behavioral  finance argues that investors’ irrational behavior can lead to  predictable patterns in market prices, such as overreaction to news  events or momentum in stock prices. However, it’s important to note that  even if certain patterns or anomalies can be identified, it does not  necessarily mean that they can be consistently exploited to generate  abnormal returns. The financial markets are complex systems with many  participants and factors influencing prices, making it difficult to  predict future market movements with high accuracy and reliability.

In  summary, while some patterns or anomalies may exist in financial  markets, it is unlikely that they can be consistently used to predict  future prices or returns with high accuracy.

VIP: do humans have predictive powers?

chat GPT: Humans  have the ability to make predictions based on patterns and information  they have learned from past experiences. However, the accuracy of these  predictions can vary depending on the complexity of the situation and  the amount and quality of information available. Humans are prone to  cognitive biases that can affect their ability to make accurate  predictions. These biases include overconfidence, confirmation bias, and  the availability heuristic, among others.

In the field of  finance, professional investors and analysts often use complex  mathematical models and vast amounts of data to make predictions about  future market movements. However, even with this information, it is  still difficult to make accurate predictions about the financial markets  due to the complexity and uncertainty of the system.

In summary,  humans have the ability to make predictions, but the accuracy of these  predictions can be affected by cognitive biases and the complexity of  the situation. In finance, accurate predictions are challenging due to  the complexity and uncertainty of the financial markets.

Feb 2023: more on Chat GPT

 as you can see, i am loving chatGPT, not because its great, but its honest … here is what i asked  the software … does chatGPT have cognitive ability ?

 “As an AI language model, I do not have cognitive abilities in the same  sense as humans. I am a sophisticated computer program that can process  and generate human-like language based on the vast amounts of data that I  have been trained on. I can understand and respond to a wide range of  questions and prompts, but I do not have subjective experiences,  emotions, or consciousness like humans do. I am designed to simulate  conversation and provide helpful information, but I do not have my own  desires or intentions.” 

conclusion

i can’t tell you how long will it take for machine Intelligence to catch up to the 5 levels described above, if ever; as the growth in technology capabilities further help building human (brain) capabilities; e.g. only in 21st century that we have trillion dollar companies and still managed by a single person (CEO) at the top, which in my view is incredible capability.

further, tools like chatGPT only use the information available on net and link them creatively (nuronic ways) but they can’t create new stuff … maybe now-a-days most people have given up on developing new stuff; instead rely on what’s already available on-the-web  ..!

however, what i can tell you is that human mind is a seriously complex thing .. so much that it can conceive and execute a trillion dollar plan … so complex that even humans like me fail to understand ..!