Open ehuppert opened 3 years ago
Thanks for sharing your work with us, Prof James! My question is about replicating patterns of decontextualized or incomplete knowledge through combinatory science. Science happens in a historical and cultural context, and can unintentionally replicate inequities in society (i.e. facial recognition bias). Is it possible for advanced alien models to experience similar bias implications? Additionally, do you see any downsides to innovating faster? Finally, if AI drives innovation that is beyond the scope of human imagination, if humans are to be the recipients of that science, do you think these alien models will be abopted and embraced?
Thank you so much for presenting this amazing work, professor! I'm curious about your general vision about AI. Do you think AI would eventually surpass human, or be more like human because we have all this biases embedded in human-generated data.
Thanks for sharing professor! Can you explain more about the concept of alien ai?
Thank you for sharig this interesting research with us. The definition of "Alien" AI is inspiring and novel. Since the timespan of your dataset is very long, I wonder whether this will generate some time bias and distort the result. Also, could you describe more on purely self-supervised models? Are they similar to uninformed AI agents without disciplinary knowledge?
Thanks for your sharing! A really interesting topic! I see many interesting questions made by the peers and I am looking forward to your answers to those questions. And as @Jasmine97Huang says, I am also curious about why explicit inclusion of authorship has such an effect.
Thank you Professor Evans for sharing this interesting research. I look forward to tomorrow's presentation.
Thank you for your presentation. Looking forward to your presentation tomorrow!
Looking forward to your presentation!
Thank you so much for sharing your work Dr Evans! I look forward to your presentation.
Thank you for this interesting research, Professor Evans! In terms of accounting for human experts, would you mind elaborating a bit on how you model the distribution of inferences cognitively available to scientists by constructing a hypergraph over research publications, as I was confused by the model and the hypergraph. Thank you and look forward to your presentation!
Thank you for sharing your research, Prof. Evans. I wanted to ask how your methods can be applied in public policy. I'm also looking forward to your presentation tomorrow.
Thank you so much for sharing your work Professor Evans! I look forward to your presentation.
Thank you so much for sharing your work. My question is how do you think your methods can be applied to other fields? Thank you and looking forward to your presentation
Welcome back Prof. Evans! My question is how do you distinguish human bias and evaluate their scale? And what do you think about the robustness of precision increase of AI models? Thanks!
Thank you for sharing your interest paper with us professor Evans! I want to hear about your opinions on what types of human intervention we need in AI learning process and how to deal with bias embeded in it like what we had discussed in the fall workshop. Thank you and looking forward to your presentation!
Hi Professor Evans, this is a very inspiring study! I am very optimistic about the further application of the model developed in your study. Since it can be used to predict future scientific discoveries through linking relevant literatures together, do you think it can be utilized to aid individual researcher, especially those who is new to being a professional, to quickly grasp the core topics and other possibilities of their research areas? I think that would be helpful for researchers to figure out what people are currently studying and who are studying the topics. Looking forward to your presentation!
Thank you for sharing such an interesting and brave paper with us! This topic is so new to me that I'm actually curious about what inspired you to generate this idea on alien AI philosophy. What's more, just as a lot of guys have asked, how can we test the validity of the alien hypotheses? In this case, is the hypothesis - testing method the same as the traditional science proceduce?
It's a great honor and pleasure to have James with us this week!! The most exciting speaker ever! XD I really cannot wait to hear your sharing!
I can see the great application of this AI prediction method. So I wonder if there are some areas that are specifically suitable for AI predictions and others are not very suitable (AI makes not very accurate predictions). Thank you!
Thank you for presenting Prof. Evans. When you elaborated on the two components of the framework, you mentioned candidate materials that are scientifically irrelevant. How do you measure those materials and how to rule out them from the framework? Thank you!
Thank you so much for sharing your work Prof. Evans! I look forward hearing more about it.
Thanks James! I am curious about if the codes and graphs for constructing the hypergraphs could be shared. I am very interested in constructing hypergraphs for political ideologies.
Thank you so much for sharing your work Prof. Evans! I look forward hearing more about it.
Thank you Prof. Evans! This is a really timely and interesting topic. Looking forward to your discussion!
Thank you for your paper ! Im not quite familiar with this topic but it looks quite exciting ! Look forward to your presentation tomorrow !
Thanks for sharing the work Prof. Evans, and I am very looking forward to seeing more about its applications in other fields. Can you kindly tell us what other fields it might be applied to?
Thank you for sharing. I look forward to hearing more about it tomorrow. Could you talk about how we can apply this to our study?
Thank you for presenting this week Dr. Evans! My first question is more of a clarification, but I'm a bit unsure about how the models described in this manuscript may be applied. For example, when predicting future discoveries (let's use the example of thermoelectric compounds), is a prediction being made about the thermoelectric properties of a specific candidate compound, an unexplored chain of inferences in available literature that is likely to lead to some breakthrough compound, the research group who are likely to discover a new compound, or some other thing entirely (perhaps a combination of these factors)?
That clarification aside, I'm also curious about how these models address the nonlinear nature of scientific discovery (can serendipity be predicted?) as well as the related assumption that the inferences available to scientists are accessible through the publications in their fields. It may be asking too much of an as-of-yet not omniscient model, but my understanding is that scientific breakthroughs can often be achieved by conceptualizing problems in terms of topics far removed from the actual problem. At the same time, the "alien modeling augmentation" approach described in this manuscript seems to promise a thorough exploration of the possibility space - perhaps even the corners otherwise only accessible through lateral thought and good luck?
Thank you so much for your exciting paper. I was really enjoying how you have creatively integrated AI methodologies and scientific phenomena to create real-world insight. I would like to hear more about how you think that this insight could be applied to educational situations. Thank you!
Hi Prof. Evans, thanks for sharing this fascinating work! I am wondering could innovation be predicted by alien artificial intelligence? Besides, could alien artificial intelligence complement humans in innovation?
Thanks for sharing your paper professor Evans! Also thanks Mike for sharing the podcast. I don't have any specific questions to ask. I look forward to the workshop and hope to see @Jasmine97Huang and @JadeBenson's questions answered.
What kind of impact can a social scientist make on the artificial intelligence research? It seems that we are too often ignored by the computer science community and this actually does hurt our opportunities.
Hi Professor Evans, thank you for sharing some of your research with us! I believe I am echoing some previous questions, but I am curious to know about any applications of this research beyond the scope discussed in the paper.
Thank you Prof. Evans! The evidence for that incorporating human expertise could contribute to the AI prediction of future human discoveries and inventions is very important to the long-lasted controversy towards artificial intelligence and human's work.
Whether being substitutes or complements serves as a key determinant in the public's attitudes towards the scientific advance, and thus, could largely affect the social movement in supporting or defeating the development of AI related work.
I'm looking forward to your presentation on the more intelligible interpretations and explanations for the 'Methods' part, especially on the visualization results. Thank you!
Thanks for sharing this new work! I would very much like to read it more carefully later. I am interested in the concept of alien versus humans proposed in the paper and would like to hear more of your explanations in the presentation tomorrow!
By the way, I would like to know how those beautiful visualizations are generated? Are there any specific tools to draw such graphs?
Really interesting research! Broadly speaking, to what degree is this approach to predicting future research inherently limited by the methods/paradigm of the reductionist approach taken by AI?
Thanks for sharing. I think it is very interesting that AI and human can work together in scientific research and I am looking forward for the presentation.
Thanks for sharing such interesting work, Prof. Evans! I have some similar questions as the top-voted ones, and I am interested in how and why you constructed the "human expertise" in the way you presented.
It's a really intriguing topic and I'm looing forward to your presentation!
Hi Professor Evans, thank you for sharing your work! I am also curious the complement of human expert into innovation and looking forward to your speech.
Thank you so much for sharing the work! I am looking forward to your presentation.
Thank you for sharing your research! I really look forward to the presentation.
Thanks very much for your sharing! I think AI and human cooperation is a very brilliant idea and denote a possible way of future research.
Hello Professor Evans, Thank you for sharing with us! As @MengChenC mentioned, I am also curious about the choice of the mixing parameter β. Are there going to be a good rule - of - thumb choice of this parameter across different research fields? If not, how can we test and justify the our choice? Thank you!
Thank you very much for sharing Prof. Evans. I believe this is a substantial field of the future. Look forward to your speech!
Thanks very much for this wonderful study! Look forward to your presentation tomorrow!
Thank you very much for sharing,Prof. Evans. I'm interested in the application in other fields.
Thank you so much for coming to our workshop Dr. Evans! My main question now: who will be our moderator if you are the panelist?
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.