11/5: Alison Gopnik

[ehuppert]

Comment below with questions or thoughts about the reading for this week's workshop.

Please make your comments by Wednesday 11:59 PM, and upvote at least five of your peers' comments on Thursday prior to the workshop. You need to use 'thumbs-up' for your reactions to count towards 'top comments,' but you can use other emojis on top of the thumbs up.

[YanjieZhou]

Thanks very much for your presentation. I think child learning process is very charming and may represents an ideal form of our people learning processes. Although it may be hard to replicate this process in adults, it is still very conducive to learning the mechanisms of learning and may help build a deep learning model.

[NikkiTing]

Thank you for sharing your work with us! I am not familiar with this subject but I learned a lot from your paper. What do you think this implies for children who are forced to mature during their childhood years (for example, the eldest siblings of underprivileged families who have to work at an early age to provide for their family)?

[ddlee19]

Thanks for the presentation! My question is what possible weaknesses do you see in your process of quantifying the amount of learning that the children exhibit?

[Yutong0828]

Thanks very much for sharing your exciting view on childhood and exploration-exploitation tradeoff! I have two questions for you.

1. As a psychology major, I am not very familiar with evolutionary psychology, but it seems to be a very important basis for understanding the meaning and future study of exploration-exploitation tradeoff. So, I would be grateful if you could spare your time explaining some unique characteristics of this specialty, like its fundamental hypothesis or logic, and also what can we further explore if childhood could be a solution for exploration-exploitation tradeoff? Thanks !
2. Much like @wanitchayap 's question, I am also wondering about the individual differences of life history strategies or exploration-exploitation tendency. What contributes to the level of using these strategies in each human being? For example, will people with longer or higher quality childhood be more inclined to use exploratory rather than exploiting strategy? Thanks again for your wonderful introduction and analysis on this topic~

[ghost]

How do you expect the duration of developmental period to change as life expectancy increases?

[Jasmine97Huang]

Thank you for sharing this really interesting paper. I haven't really thought about the shifts in the ways we learn as we age. After reading your paper, I can't stop but thinking about the differences in which female and male children are being given different opportunities to explore (girls are encouraged to learn about care-taking knowledge/roles, boys are motivated to be interested in science and engineering, etc. Additionally, in certain culture and countries, girls are forced to mature early in order to help support the family or to start their own family. How do these differences in the exploration stage influence their cognitive abilities later?

[caibengbu]

This idea of comparing the learning pattern of Children and AI is amazing. Though the paper does not mention it, I do hope you can introduce to us your methodology on how to link children’s cognitions to reinforcement learning algorithms. Looking forward to your presentation!

[shenyc16]

Thank you for sharing this interesting research. I find your comparsion bewteen learning pattern of Children and AI very inspiring. From my perspective, AI always learns things for certain purpose, but children do not. I am wondering how you can measure the quality of exploration and exploitation in childhood.

[RuoyunTan]

Thank you for sharing this insightful work with us. Do you think the explore-exploit mechanism would work differently at home and at a school environment? Also, in the current pandemic, children are studying in a different environment than before. I wonder what impact this may have.

[ginxzheng]

Thank you for sharing! It is very interesting to find out a similar problem of economic decisions also in psychology. The explore-exploit dilemma is also related to risk-averse or risk-welcome types. May I wonder, can this trade-off finding also apply to the variation of people's economic decisions as one grows up?

[xxicheng]

Fascinating sharing, and I am looking forward to tomorrow's presentation! Just like several posts mentioned before: what is your expectation for applying AIs and reinforcement learning with the concept of learning, life history, and maybe evolutionary behavioral science in general? How do you see the field moving toward? What should be the general goal or approach the researchers should take?

Thank you.

[anqi-hu]

Thank you for sharing your work with us. Do you find significant differences in the explore-exploit paradigm in various cultural contexts? That is, do children of different backgrounds exhibit different ways of learning through the model?

[NaiyuJ]

Thanks for sharing! The finding is super interesting. I notice that you're exploring these learning methods through experiments putting children and AIs in the same virtual environments. I'm wondering how the experiment setting would be to test this hypothesis.

[JuneZzj]

Thank you for presenting. It is interesting to know that children are better at inferring the hypothesis that is an unusual abstract. Do we really need to make decisions on historical experience and learning? As we are accumulating our knowledge, do you really have to give up the ability to abstractly thinking? Thank you.

[luckycindyyx]

Thank you for sharing such an intersting topic to us! I find some of my own image on the relationship between childhood and explore-exploit tension. My question is that, how do you precisely train the AI to simulate children's learning behavior? What's more, what are the algorithms that are commonly used in this area?

[goldengua]

This is a fansinating paper. I was particularly interested in one kind of common sense: the understanding of object typicality for blind children and machines. Counterintuitively, we don't talk about typical aspects of the object. If the apple is red, we usually say apple instead of 'red apple'. When the apple is atypical with a green color, we are more likely to mention that the color of the apple. For blind children and machines without graphical inputs, I was wondering how they could understand the typicality of objects under the exploring-exploit paradigm?

[jinfei1125]

Hi! This is a very interesting article! In fact, I think I encounter the explore-exploit dilemma every day--when I encounter a new problem, I can convert it to processes that I can solve with old methods, or I can explore new methods to solve this problem more efficiently and therefore learning new things. However, the second way is usually more time-consuming and I have to take the risk that I can't find a more efficient method or I can't understand it. I really appreciate the suggestion you give: the learner should begin by exploring and gradually converge on exploiting. I am also interested in the theme issue you mentioned 'Life history and learning: how childhood, caregiving and old age shape cognition, and culture in humans and other animals'. Could you talk more about this series? Thanks! Great article! (It also clarifies one of my questions before--if there is only one author of an article, will she/he use the first person pronoun 'I'? Because single-author article is really rare I haven't found answers. From your article, I know the answer is 'yes' :)

[weijiexu-charlie]

Thanks for your presentation. The mechanism of the exploration-exploitation tradeoff sounds quite interesting. But I agree with many other posts that the mechanism of learning for children is quite different from those machine learning algorithms. In this sense, I'm wondering how the learning mechanism of children could provide insights into ML?

[lulululugagaga]

Thanks for sharing this research with us. This is very impressive and valuable. Do you think that there will some cultural differences in child learning process and what might be the features we should include if we're going to train different models for different cultures?

[Rui-echo-Pan]

Thank you for sharing! I wonder how the explore-exploit mechanism evolves in people's life long in different cultures? Is there evidence or is it possible to explore, how culture and environment would influence the development of people's cognitive ability? Thanks!

[mingtao-gao]

Thank you for your presentation in advance! I have a more general question about this paper. I do see the similarity between human learning and artificial intelligence, however, can we compare child learning, which has not been fully understood by humans, with machine learning, which is created by human?

[heathercchen]

Thank you in advance for your presentation! This is a brand new field for me. What I am wondering is some details of "explore-exploit" behavior you mentioned a lot in your paper. At the beginning of section 2, you mentioned that "With random exploration heuristics, the agent might just randomly try new actions". Do people's risk preferences influence the probability that they take on new actions? Thanks!

[MegicLF]

Thank you for your presentation! I wonder if you could elaborate on how you connect the reinforcement learning algorithms with childhood cognition and your motivation for it. Thanks!

[luxin-tian]

Thank you for sharing this work. This is super innovative. I am interested in @yierrr's question, how can this be generalized to general biological brain?

[FrederickZhengHe]

Thanks very much for this inspiring paper! Just an irrelevant question. Nowadays more and more children have begun learning programming by using mainstream platforms such as Python and R. Do you agree that these children's learning and cognitive capabilities after they become adults are generally better than those learning programming at a later age, say, 18 years old?

[yongfeilu]

Thank you very much for the presentation! In comparison with children's learning process, how do you think machine learning is different? If they are basically very similar, what do you think is the thing that makes a human human and a machine machine?

[YuxinNg]

Thank you for your paper. My question would be some common senses we learned are our human nature or in our gene. Do you think if there a barrier for machine to learn these things. Thank you!

[adarshmathew]

I'll just be happy if you can answer @wanitchayap's detailed question. Thank you!

[Panyw97]

Thanks for sharing! It's an interesting topic that you used reinforcement learning in the study of childhood cognition. Could you elaborate on the probable application of this research in reality?

[Tanzi11]

I am looking forward to your presentation! You mention exploratory play as a fundamental aspect of a child’s growth: how does it differ across cultures and to some extent, could exploratory play be seen as a privilege in some societies?

[WMhYang]

Thank you very much for your presentation. As a student major in economics, the first thing that comes to my mind is risk preference when I see the explore-exploit tradeoff. Though it might be quite diverse from the main idea, I was wondering whether there is any heterogeneity in the length of explore period of children in terms of the risk preference and the volatility of the living environment of their parents. Thanks again.

[aolajide]

Are there people for whom this child-found nature still exists? Perhaps a psychological population? And is there a sub-population of children in which the more adult-found nature of exploitation exists?

[kthomas14]

Thank you for sharing your research! My question is can cultural differences influence the amount of time a child stays in the exploratory phase before growing into using exploitive methods?

[97seshu]

Thank you for sharing! I wonder for socially isolated kids like Genie, how would you expect the explore-exploit pattern to change?

[XinSu6]

Thank you so much for the shared work. And I am just wondering how big is the schooling effect here? And what other fields of research studies do you think your methodology can be applied to? Thank you!

[lihanhuisherry]

Thanks for sharing with us your work! I really look forward to learning more about how childhood cognition is similar to machine learning.

[siruizhou]

Thanks for sharing this work. I have a similar question with @afchao. Are these exploratory and exploit/executive behaviors caused by different kinds of training received during childhood and adulthood periods? It seems children are taught to be exploratory and young adults are often advised to be more goal-oriented. Besides, is this explore/exploit tradeoff related to fluid and crystallized intelligence?

[YijingZhang-98]

Thanks for sharing. I am an econ girl with little background in psychology, but I enjoy reading your paper, which should be an invoke for computational investigations of the links between childhood, learning, exploration, and play. Look forward to your presentation!

[qishenfu1]

Hi Alison, thank you for sharing your work! Although I am majoring in economics, I am personally very interested in psychology. Your work is a great combination of computational methods and psychology research. The topic is also interesting.

[ttsujikawa]

Thank you for the super interesting ideas. I am originally very interested in how we can apply brain neural networks to the development of artificial intelligence especially in the field of machine learning. having this background, I really enjoyed your reading. I was just wondering if we could go inversely. In other words, how can we deepen our understanding of our brain by studying neural networks on a computer? Would it be feasible that we construct a neural network that reveals the detailed structure of human/animal brains? Anyway, I am looking forward to listening to your presentation!

[fyzh-git]

Thank you for sharing this paper! The work on how the statistical methods i.e. machine learning techniques simulate the human being's construction of common sense is really intriguing; and comparing the unsupervised/supervised learning with child's explore/exploit picking up process re-shows the classical principles and development of machine learning. Look forward to hear more about the demonstrating details from the presentation!

[Yiqing-Zh]

Thank you for sharing the work. It is quite an interesting topic of linking children learning and machine learning. I am wondering that changes in the educational environment, such as transiting from learning on-site to learning remotely, would influence children's learning patterns.

[Qlei23]

Thank you for sharing your work. The explore-exploit tradeoff is an interesting idea. I'm curious about the cross-culture heterogeneous effects and does the result have some implications in the field of biological psychology? Look forward to the workshop presentation!

[YileC928]

Thank you so much for sharing! Just out of curiosity, how might interdisciplinary learning affect explore-exploit behavior during a person's life history? Also, I am interested in learning more details about possible experiment designs for further research. Will be glad if @wanxii 's question is answered.

[YaoYao121]

Thanks for sharing with us your work!I'm also curious about how machine learning could furture contribute to research on the effect of school education on children's development. Thanks!

[chuqingzhao]

Thank you for sharing your work! This is very interesting to learn about the exploratory-exploit trade-off and its computational implication. I am very interested into two following questions:1. As you have mentioned "sensitive periods", the animal in the sensitive period is more likely to learn new things. I am wondering is there any periodic change in human's life-time learning? 2. Like many other questions concerning different cultural contexts, I am curious in different cultures is there any difference in exploratory-exploit learning? Thank you.

[chun-hu]

Thanks for your work! There are many unknowns of applying AI to cognitive development. How do you see the future of this research?

[luyingjiang]

Thank you for sharing. I am just curious whether you think there are some differences in the explore-exploit paradigm in different environments? (i.e. different schooling environment or cultural background, etc)

[chentian418]

Thanks for sharing the paper in the interesting realm of cognition! I also resonate with the conclusion a lot! During my childhood, I was both exploring things with purpose, like the schoolworks at my kindergarten, and being exploratory unintentionally, like doing things I find interesting naturally. And indeed I find my growth has some developing patterns from the exploration in childhood.

One concern is that, since the time window to study may be from a person's childhood to adulthood, how do you keep track of and collect this data in such a great time span?

Thanks!

[j2401]

Thanks for presenting your work! Like some others have mentioned, I feel it might also be interesting we could make a comparison between different educational strategies in the childhood. It seems that in Asian culture education tends to be more focused on development and recently tends to encourage more exploration. While a reverse trend appears in some western countries. If these data points are available, would them be useful to find more implications like those in your work?