The Computer Science of Human Decision Making | Tom Griffiths | TEDxSydney

Summary

This talk explores the application of computer science to everyday decision-making, focusing on the '37% rule' for optimal stopping in scenarios like house hunting. It delves into the Explorer-Exploit trade-off, illustrating how to balance trying new options with sticking to known quantities. The speaker, a computational cognitive scientist, uses examples like choosing a restaurant and organizing a wardrobe to demonstrate how computer science principles can simplify complex human problems, ultimately encouraging a more forgiving approach to our own decision-making processes.

Takeaways

• 🏠 The '37 percent rule' is a strategy for making decisions when faced with a limited number of options, suggesting that one should look at 37% of the options before making a decision on the next best one.
• 🤔 The script discusses the challenges of decision-making, particularly in the context of finding a home in Sydney, and how to balance the risk of missing out on better options.
• 🧠 The presenter is a computational cognitive scientist who studies the computational structure of everyday problems and compares ideal solutions with actual human behavior.
• 💡 Applying computer science to human decision-making can simplify complex problems and provide strategies for making better-informed choices.
• 🍴 The 'Explorer-Exploit trade-off' is a concept that arises in various scenarios, such as choosing a restaurant, where one must decide between trying something new or sticking with a known preference.
• 👶 The value of information increases with the number of opportunities to use it, which is exemplified by babies exploring their environment to learn.
• 👴 Similarly, older individuals who stick to familiar routines are seen as making optimal decisions based on a lifetime of experiences.
• 👗 Computer memory principles, such as the Least Recently Used (LRU) strategy, can be applied to decluttering a wardrobe, emphasizing the importance of keeping frequently used items accessible.
• 📚 Yukio Noguchi's filing system and the concept of organizing documents by recent use can be applied to personal and office organization for efficiency.
• 🔑 The script highlights that even when using the best processes, outcomes are not guaranteed, which is a concept that can lead to a more forgiving approach to decision-making.
• 🤖 Computer science offers insights into problem-solving strategies that can help us understand and accept the limitations of human decision-making processes.

Q & A

What is the 37% rule mentioned in the script?

The 37% rule is a strategy to maximize the probability of finding the best option when faced with a series of choices. It suggests that you should look at 37% of the available options without making a decision, then choose the first option that is better than all the ones you've seen so far.

What is an optimal stopping problem?

An optimal stopping problem is a type of problem in decision theory where the decision-maker must choose the best time to stop information gathering and make a decision based on the information already obtained.

How does the script relate the concept of the Explorer-Exploit trade-off to everyday life?

The script relates the Explorer-Exploit trade-off to everyday life by illustrating it with the example of choosing a restaurant. It's the dilemma of whether to try something new (explore) or to stick with something known to be good (exploit).

What is the personal motivation behind the speaker's interest in computational cognitive science?

The speaker's personal motivation stems from growing up as an overly cerebral kid in Perth, always trying to act rationally and reason through every decision, which led to exhaustion and failure in some aspects of life, including personal relationships.

How does the script suggest we approach decision-making in life?

The script suggests that we should consult experts like computer scientists when faced with computational problems that are too hard to solve by applying sheer effort. It also encourages taking chances, exploring new options, and being forgiving of our own limitations.

What is the Explorer-Exploit trade-off in the context of technology companies?

For technology companies, the Explorer-Exploit trade-off is the decision between showing a new ad to learn about its effectiveness or showing an ad that is already known to perform well, thus exploiting the existing knowledge.

How does the value of information increase according to the script?

The value of information increases the more opportunities you have to use it. For example, if you are going to be in town for a longer time, it's worth exploring new options because the information gained can improve future choices.

What principle does the script suggest for tidying up a wardrobe?

The script suggests applying the least recently used principle, which is a strategy often used in computer memory systems, to decide what items to keep or give away in a wardrobe.

How does Yukio Noguchi's filing system work?

Yukio Noguchi's filing system works by placing documents in a box from the left-hand side and moving existing documents along each time a new document is added. When a document is accessed, it is put back on the left side, resulting in an order from left to right by how recently they've been used.

What is the significance of the 37% rule in the context of finding a home?

The 37% rule in the context of finding a home suggests that after viewing 37% of the available properties, you should make an offer on the next property that is better than any of the ones you've seen so far, with the aim of maximizing the probability of finding the best home.

How can computer science help in making decisions in difficult problems?

Computer science can help by breaking down difficult problems into simpler ones, making use of randomness, removing constraints, or allowing approximations. Solving these simpler problems can provide insight into the harder problems and sometimes produce good solutions.